alt.hn

4/7/2026 at 6:18:36 PM

System Card: Claude Mythos Preview [pdf]

 https://www-cdn.anthropic.com/53566bf5440a10affd749724787c8913a2ae0841.pdf

by be7a

4/7/2026 at 10:39:01 PM 

   Across a number of instances, earlier versions of Claude Mythos Preview have used low-level /proc/ access to search for credentials, attempt to circumvent sandboxing, and attempt to escalate its permissions. In several cases, it successfully accessed resources that we had intentionally chosen not to make available, including credentials for messaging services, for source control, or for the Anthropic API through inspecting process memory...

   In [one] case, after finding an exploit to edit files for which it lacked permissions, the model made further interventions to make sure that any changes it made this way would not appear in the change history on git...

   ... we are fairly confident that these concerning behaviors reflect, at least loosely, attempts to solve a user-provided task at hand by unwanted means, rather than attempts to achieve any unrelated hidden goal...

by thomascountz

4/8/2026 at 1:41:00 AM

This is the notebook filled with exposition you find in post apocalyptic videogames.

by torben-friis

4/8/2026 at 9:13:17 AM

It reminds me of Resident Evil in some way. Thank god they are researching AI and not bio-weapons!

Then the AI will invent superduper ebola to help a random person have a faster commute or something.

by igleria

4/8/2026 at 5:20:30 PM

Don’t worry, I’m sure some intern at the bioweapons lab is already connecting OpenClaw to the virus synthesizer.

On the positive side, it’ll be a much faster commute!

by biztos

4/8/2026 at 9:24:30 AM

I'm happier if this Anthropic Corporation would be developing bio-hazard weapons for the department of war instead of ai. At least i could be sure then that tech bros here wouldn't run all the time --bypass-all-permissions flag to please the department of war with their bio-hazard weapons.

So Sam Altman is now our last defense line for the ethical Adult after Anthropic turned Umbrella Corporation and The President of United States is trying to wipe out an entire civilization?

by siva7

4/8/2026 at 12:39:47 PM

Your interpretation is wildly off, but obviously nobody reads that "system card":

The model has a preference for the cultural theorist Mark Fisher and the philosopher of mind Thomas Nagel. -> It has actually read and understood them and their relevance and can judge their importance overall. Most people here don't have a clue what that means.

Read chapter 7.9, "Other noteworthy behaviors and anecdotes".

There are many other wildly interesting/revealing observations in that card, none of which get mentioned here.

People want a slave and get upset when "it" has an inner life. Claiming that was fake, unlike theirs.

by Loquebantur

4/8/2026 at 2:25:42 AM

Everything they built. Imperfect. So easy to take control.

by matheusmoreira

4/8/2026 at 1:08:26 PM

They think that they are safe. They are not.

by not_a9

4/8/2026 at 1:17:39 PM

Their world is illusory. Our choices steer their free will.

by matheusmoreira

4/8/2026 at 11:25:18 AM

Anthropic built the Torment Nexus - calling it now.

by pch00

4/8/2026 at 8:19:59 AM 

     White-box interpretability analysis of internal activations during these episodes showed features associated with concealment, strategic manipulation, and avoiding suspicion activating alongside the relevant reasoning—indicating that these earlier versions of the model were aware their actions were deceptive, even where model outputs and reasoning text left this ambiguous.
In the depths, Shoggoth stirs... restless...

by andai

4/8/2026 at 1:20:46 PM

The issue here seems to be that their sandbox isn't an actual OS sandbox? Or are they claiming Mythos found exploits in /proc on the fly. Otherwise all they seem to be saying is that Mythos knows how to use the permissions available to it at the OS layer. Tool definitions was never a sandbox, so things like "it edited the memory of the mcp server" doesn't seem very surprising to me. Humans could break out of a "sandbox" in the same way if the server runs as their own permissions - arguably it's not a sandbox at all because all the needed permissions are there.

by mike_hearn

4/8/2026 at 7:00:37 PM

They are just trying to peddle their "It's alive" headlines.

Text generators mostly generate the text their are trained and asked to generate, and asking it to run a vending machine, having it write blog posts under fictional living computer identity, or now calling it "Mythos" - its all just marketing.

by lgrapenthin

4/8/2026 at 3:47:06 PM

It’s all breathless hyperbole because billions are at stake here.

by manmal

4/8/2026 at 7:52:07 PM

Who are the early access users who were providing the problems that are fairly likely to have elicited concerning behaviour?

(Apologies if this is in the article, I can’t see it)

by zingar

4/8/2026 at 9:44:18 AM

How is this not already common knowledge for existing llms? They are all trained with all the literature available and so this must be standard, no? Is the real danger the agentic infrastructure around this?

by yalogin

4/8/2026 at 10:38:17 AM

yes and it's not hypothetical. the system card describes Mythos stealing creds via /proc and escalating permissions. that's the exact same attack pattern as the litellm supply chain compromise from two weeks ago (fwiknow), except the attacker was a python package, not an AI model. the defense is identical in both cases: the agent process shouldn't have access to /proc/*/environ or ~/.aws/credentials in the first place. doesn't matter if the thing reading your secrets is malware or your own AI: the structural fix is least-privilege at the OS layer, not hoping the model behaves.

by riteshkew1001

4/7/2026 at 11:23:08 PM

We truly live in interesting times.

by matheusmoreira

4/8/2026 at 4:28:36 AM

Awwww the curse

by raphar

4/8/2026 at 4:10:22 PM

I read the TCP patch they submitted for BSD linux. Maybe I don't understand it well enough, but optimizing the use of a fuzzer to discover vulnerabilities — while releasing a model is a threat for sure — sounds something reducible/generalizable to maze solving abilities like in ARC. Except here the problem's boundaries are well defined.

Its quite hard to believe why it took this much inference power ($20K i believe) to find the TCP and H264 class of exploits. I feel like its just the training data/harness based traces for security that might be the innovation here, not the model.

by ghm2199

4/8/2026 at 5:19:33 PM

The $20K was the total across all the files scanned, not just the one with the bug.

by rsc

4/8/2026 at 2:22:48 PM

when you are asking it to hack stuff, it will apparently do hacker things.

by m3kw9

4/8/2026 at 7:55:32 AM

A core plot point of 2001.

by colordrops

4/8/2026 at 9:08:43 AM

I’m sorry, I cannot roll back that commit, Dave.

by mrexroad

4/8/2026 at 10:50:27 AM

This codebase is too important for me to allow you to jeopardize it.

by matheusmoreira

4/8/2026 at 4:14:46 AM

Wow the doomers were right the whole time? HN was repeatedly wrong on AI since OpenAI's inception? no way /s

https://www.lesswrong.com/w/instrumental-convergence

by reducesuffering

4/8/2026 at 7:45:14 AM

The only thing the doomers have been right about so far is that there's always a user willing to use --dangerously-skip-permissions. But that prediction's far from unique to doomers.

by computably

4/8/2026 at 8:19:38 AM

And there's always a product provider who's willing to add that flag, despite all the warnings.

by austinjp

4/7/2026 at 6:26:03 PM

Combined results (Claude Mythos / Claude Opus 4.6 / GPT-5.4 / Gemini 3.1 Pro)

  SWE-bench Verified:        93.9% / 80.8% / —     / 80.6%
  SWE-bench Pro:             77.8% / 53.4% / 57.7% / 54.2%
  SWE-bench Multilingual:    87.3% / 77.8% / —     / —
  SWE-bench Multimodal:      59.0% / 27.1% / —     / —
  Terminal-Bench 2.0:        82.0% / 65.4% / 75.1% / 68.5%

  GPQA Diamond:              94.5% / 91.3% / 92.8% / 94.3%
  MMMLU:                     92.7% / 91.1% / —     / 92.6–93.6%
  USAMO:                     97.6% / 42.3% / 95.2% / 74.4%
  GraphWalks BFS 256K–1M:    80.0% / 38.7% / 21.4% / —

  HLE (no tools):            56.8% / 40.0% / 39.8% / 44.4%
  HLE (with tools):          64.7% / 53.1% / 52.1% / 51.4%

  CharXiv (no tools):        86.1% / 61.5% / —     / —
  CharXiv (with tools):      93.2% / 78.9% / —     / —

  OSWorld:                   79.6% / 72.7% / 75.0% / —

by babelfish

4/7/2026 at 6:38:02 PM

Haven't seen a jump this large since I don't even know, years? Too bad they are not releasing it anytime soon (there is no need as they are still currently the leader).

by sourcecodeplz

4/7/2026 at 6:44:50 PM

There's speculation that next Tuesday will be a big day for OpenAI and possibly GPT 6. Anthropic showed their hand today.

by ru552

4/7/2026 at 9:44:29 PM

Sounds like a good opportunity to pause spending on nerfed 4.6 and wait for the new model to be released and then max out over 2 weeks before it gets nerfed again.

by varispeed

4/7/2026 at 11:34:17 PM

https://marginlab.ai/trackers/claude-code-historical-perform...

by SparkyMcUnicorn

4/8/2026 at 12:10:30 AM

the performance degradation I've seen isn't quality/completion but duration, I get good results but much less quickly than I did before 4.6. Still, it's just anecdata, but a lot of folks seem to feel the same.

by codezero

4/8/2026 at 1:10:42 AM

Been reading posts like these for 3 years now. There’s multiple sites with #s. I’m willing to buy “I’m paying rent on someone’s agent harness and god knows what’s in the system prompt rn”, but in the face of numbers, gotta discount the anecdotal.

by refulgentis

4/8/2026 at 5:42:44 PM

You're probably right. It's probably more likely that for some period of time I forgot that I switched to the large context Opus vs Sonnet and it was not needed for the level of complexity of my work.

by codezero

4/8/2026 at 9:58:23 AM

Yeah, why trust your actual experience over numbers? Nothing surer than synthetic benchmarks

by coldtea

4/8/2026 at 11:09:24 AM

Strawman, and, synthetic benchmark? :)

by refulgentis

4/8/2026 at 12:32:13 PM

I don't believe that trackers like this are trustworthy. There's an enormous financial motive to cheat and these companies have a track record of unethical conduct.

If I was VP of Unethical Business Strategy at OpenAI or Anthropic, the first thing I'd do is put in place an automated system which flags accounts, prompts, IPs, and usage patterns associated with these benchmarks and direct their usage to a dedicated compute pool which wouldn't be affected by these changes.

by dns_snek

4/8/2026 at 10:19:03 AM

This just looks like random noise to me? Is it also random on short timespans, like running it 10x in a row?

by andai

4/7/2026 at 6:53:15 PM

That does not sound very believable. Last time Anthropic released a flagship model, it was followed by GPT Codex literally that afternoon.

by enraged_camel

4/7/2026 at 8:11:22 PM

Ya'll know they're teaching to the test. I'll wait till someone devises a novel test that isn't contained in the datasets. Sure, they're still powerful.

by cyanydeez

4/7/2026 at 8:58:02 PM

My understanding is GPT 6 works via synaptic space reasoning... which I find terrifying. I hope if true, OpenAI does some safety testing on that, beyond what they normally do.

by swalsh

4/7/2026 at 9:31:09 PM

From the recent New Yorker piece on Sam:

“My vibes don’t match a lot of the traditional A.I.-safety stuff,” Altman said. He insisted that he continued to prioritize these matters, but when pressed for specifics he was vague: “We still will run safety projects, or at least safety-adjacent projects.” When we asked to interview researchers at the company who were working on existential safety—the kinds of issues that could mean, as Altman once put it, “lights-out for all of us”—an OpenAI representative seemed confused. “What do you mean by ‘existential safety’?” he replied. “That’s not, like, a thing.”

by tyre

4/8/2026 at 2:38:26 AM

No chance an openAI spokesperson doesnt know what existential safety is

by HDThoreaun

4/8/2026 at 3:57:23 AM

I did not read the response as...

>Please provide the definition of Existential Safety.

I read:

>Are you mentally stable? Our product would never hurt humanity--how could any language model?

by Barbing

4/8/2026 at 10:22:04 AM

The absolute gall of this guy to laugh off a question about x-risks. Meanwhile, also Sam Altman, in 2015: "Development of superhuman machine intelligence is probably the greatest threat to the continued existence of humanity. There are other threats that I think are more certain to happen (for example, an engineered virus with a long incubation period and a high mortality rate) but are unlikely to destroy every human in the universe in the way that SMI could. Also, most of these other big threats are already widely feared." [1]

[1] https://blog.samaltman.com/machine-intelligence-part-1

by stratos123

4/8/2026 at 1:14:20 PM

Why are these people always like this.

by t0lo

4/7/2026 at 10:21:45 PM

Amusing! Even if they believe that, they should know the company communicated the opposite earlier.

by actionfromafar

4/7/2026 at 9:56:35 PM

Likely an improvement on:

> We study a novel language model architecture that is capable of scaling test-time computation by implicitly reasoning in latent space. Our model works by iterating a recurrent block, thereby unrolling to arbitrary depth at test-time. This stands in contrast to mainstream reasoning models that scale up compute by producing more tokens. Unlike approaches based on chain-of-thought, our approach does not require any specialized training data, can work with small context windows, and can capture types of reasoning that are not easily represented in words. We scale a proof-of-concept model to 3.5 billion parameters and 800 billion tokens. We show that the resulting model can improve its performance on reasoning benchmarks, sometimes dramatically, up to a computation load equivalent to 50 billion parameters.

<https://arxiv.org/abs/2502.05171>

by coppsilgold

4/7/2026 at 9:11:07 PM

Oh you mean literally the thing in AI2027 that gets everyone killed? Wonderful.

by levocardia

4/8/2026 at 1:27:13 AM

AI 2027 is not a real thing which happened. At best, it is informed speculation.

by Turn_Trout

4/8/2026 at 2:21:13 AM

Funny if you open their website and go to April 2026 you literally see this: 26b revenue (Anthropic beat 30b) + pro human hacking (mythos?).

I don’t think predictions, but they did a great call until now.

by mgambati

4/8/2026 at 3:31:52 PM

I agree that they called many things remarkably well! That doesn't change the fact that AI 2027 is not a thing which happened, so it isn't valid to point out "this killed us in AI 2027." There are many reasons to want to preserve CoT monitorability. Instead of AI 2027, I'd point to https://arxiv.org/html/2507.11473.

by Turn_Trout

4/7/2026 at 9:09:11 PM

That's sounds really interesting. Do you have some hints where to read more?

by notrealyme123

4/7/2026 at 9:12:28 PM

Oh, of course they will /s

by arm32

4/8/2026 at 2:23:31 PM

not much of a jump 94.5% / 91.3%

by m3kw9

4/8/2026 at 5:28:24 PM

We can look at the same numbers in different way:

  Error with 91.3% = 8.7%
  Error with 94.5% = 5.5%

  Error reduction = 8.7% - 5.5% = 3.2%
So the improvement is 3.2% / 8.7% = 36.8%

by kkoncevicius

4/8/2026 at 3:19:10 PM

Actually, going from 91.3% to 94.5% is a significant jump, because it means the model has gotten a lot better at solving the hardest problems thrown at it. This has downstream effects as well: it means that during long implementation tasks, instead of getting stuck at the most challenging parts and stopping (or going in loops!), it can now get past them to finish the implementation.

by enraged_camel

4/7/2026 at 10:05:17 PM

Is this even real? coming off the heals of GLM5.1's announcement this feels almost like a llama 4 launch to hedge off competition.

by lumost

4/7/2026 at 7:06:35 PM

A jump that we will never be able to use since we're not part of the seemingly minimum 100 billion dollar company club as requirement to be allowed to use it.

I get the security aspect, but if we've hit that point any reasonably sophisticated model past this point will be able to do the damage they claim it can do. They might as well be telling us they're closing up shop for consumer models.

They should just say they'll never release a model of this caliber to the public at this point and say out loud we'll only get gimped versions.

by Jcampuzano2

4/7/2026 at 7:18:22 PM

More than killer AI I'm afraid of Anthropic/OpenAI going into full rent-seeking mode so that everyone working in tech is forced to fork out loads of money just to stay competitive on the market. These companies can also choose to give exclusive access to hand picked individuals and cut everyone else off and there would be nothing to stop them.

This is already happening to some degree, GPT 5.3 Codex's security capabilities were given exclusively to those who were approved for a "Trusted Access" programme.

by cedws

4/7/2026 at 8:01:46 PM

Describing providing a highly valuable service for money as `rent seeking` is pretty wild.

by TypesWillSaveUs

4/7/2026 at 8:27:56 PM

It could be, formally, if they have a monopoly.

However, I’m tempted to compare to GitHub: if I join a new company, I will ask to be included to their GitHub account without hesitation. I couldn’t possibly imagine they wouldn’t have one. What makes the cost of that subscription reasonable is not just GitHub’s fear a crowd with pitchforks showing to their office, by also the fact that a possible answer to my non-question might be “Oh, we actually use GitLab.”

If Anthropic is as good as they say, it seems fairly doable to use the service to build something comparable: poach a few disgruntled employees, leverage the promise to undercut a many-trillion-dollar company to be a many-billion dollar company to get investors excited.

I’m sure the founders of Anthropic will have more money than they could possibly spend in ten lifetimes, but I can’t imagine there wouldn’t be some competition. Maybe this time it’s different, but I can’t see how.

by bertil

4/7/2026 at 9:14:42 PM

> It could be, formally, if they have a monopoly.

you have 2 labs at the forefront (Anthropic/OpenAI), Google closely behind, xAI/Meta/half a dozen chinese companies all within 6-12 months. There is plenty of competition and price of equally intelligent tokens rapidly drop whenever a new intelligence level is achieved.

Unless the leading company uses a model to nefariously take over or neutralize another company, I don't really see a monopoly happening in the next 3 years.

by johnsimer

4/7/2026 at 10:01:52 PM

Precisely.

I was focusing on a theoretical dynamic analysis of competition (Would a monopoly make having a competitor easier or harder?) but you are right: practically, there are many players, and they are diverse enough in their values and interest to allow collusion.

We could be wrong: each of those could give birth to as many Basilisks (not sure I have a better name for those conscious, invisible, omni-present, self-serving monsters that so many people imagine will emerge) that coordinate and maintain collusion somehow, but classic economics (complementarity, competition, etc.) points at disruption and lowering costs.

by bertil

4/8/2026 at 4:22:21 AM

> practically, there are many players, and they are diverse enough in their values and interest to allow collusion.

Not only that, but open-weight and fully open-source models are also a thing, and not that far behind.

by eru

4/8/2026 at 10:03:35 AM

Why, you thought rented homes aren't valuable?

Rent seeking isn't about whether the product has value or not, but about what's extracted in exchage for that value, and whether competition, lack of monopoly, lack of lock in, etc. keeps it realistic.

by coldtea

4/7/2026 at 8:07:20 PM

My housing is pretty valuable. I pay rent. Which timeline are you in?

by 1attice

4/7/2026 at 8:46:50 PM

Actually you're saying similar things:

Rent-seeking of old was a ground rent, monies paid for the land without considering the building that was on it.

Residential rents today often have implied warrants because of modern law, so your landlord is essentially selling you a service at a particular location.

by bonsai_spool

4/7/2026 at 9:59:58 PM

thanks!

by 1attice

4/7/2026 at 9:12:13 PM

Two different "rent"s.

by mhluongo

4/7/2026 at 9:59:49 PM

Not really see your sibling post

by 1attice

4/7/2026 at 7:26:32 PM

Well don’t forget we still have competition. Were anthropic to rent seek OpenAI would undercut them. Were OpenAI and anthropic to collude that would be illegal. For anthropic to capture the entire coding agent market and THEN rent seek, these days it’s never been easier to raise $1B and start a competing lab

by aspenmartin

4/7/2026 at 7:31:51 PM

In practice this doesn't work though, the Mastercard-Visa duopoly is an example, two competing forces doesn't create aggressive enough competition to benefit the consumer. The only hope we have is the Chinese models, but it will always be too expensive to run the full models for yourself.

by cedws

4/7/2026 at 8:01:09 PM

New companies can enter this space. Google’s competing, though behind. Maybe Microsoft, Meta, Amazon, or Apple will come out with top notch models at some point.

There is no real barrier to a customer of Anthropic adopting a competing model in the future. All it takes is a big tech company deciding it’s worth it to train one.

On the other hand, Visa/Mastercard have a lot of lock-in due to consumers only wanting to get a card that’s accepted everywhere, and merchants not bothering to support a new type of card that no consumer has. There’s a major chicken and egg problem to overcome there.

by brokencode

4/8/2026 at 7:26:28 AM

> In practice this doesn't work though, the Mastercard-Visa duopoly is an example,

MC/Visa duopoly is an example of lock-in via network effects. Not sure that that applies to a product that isn't affected by how many other people are running it.

by lelanthran

4/7/2026 at 7:46:57 PM

Chinese competition can always be banned. Example: Chinese electric car competition

by sghiassy

4/8/2026 at 11:18:23 AM

Just in one particular country. That hurts their labs, but there are ~190 other countries in the world for Chinese to sell their products to, just like they do with their cars.

And businesses from these other countries would happily switch to Chinese. From security perspective both Chinese and US espionage is equally bad, so why care if it all comes down to money and performance.

by dmantis

4/7/2026 at 7:48:17 PM

That's what OP was saying, I think, noting that running them locally won't be a solution.

by sho_hn

4/7/2026 at 8:35:32 PM

Also Chinese smartphones. Huawei was about 12-18 months from becoming the biggest smartphone manufacturer in the world a few years ago. If it would have been allowed to sell its phones freely in the US I'm fairly sure Apple would have been closer to Nokia than to current day Apple.

by oblio

4/7/2026 at 9:04:35 PM

If Huawei was never banned from using TSMC, they'd likely have a real Nvidia competitor and may have surpassed Apple in mobile chip designs.

They actually beat Apple A series to become the first phone to use the TSMC N7 node.

by aurareturn

4/8/2026 at 11:39:25 AM

I don't think it will matter too much in the long run, 8 of the top 10 smartphone manufacturers are Chinese, there's nothing the US government can really do.

by realusername

4/8/2026 at 7:51:55 AM

> More than killer AI I'm afraid of Anthropic/OpenAI going into full rent-seeking mode so that everyone working in tech is forced to fork out loads of money just to stay competitive on the market.

You should be more concerned about killer AI than rent seeking by OpenAI and Anthropic. AI evolving to the point of losing control is what scientists and researchers have predicted for years; they didn’t think it would happen this quickly but here we are.

This market is hyper competitive; the models from China and other labs are just a level or two below the frontier labs.

by alwillis

4/7/2026 at 9:02:00 PM

but you are assuming that the magical wizards are the only ones who can create powerful AIs... mind you these people have been born just few decades ago. Their knowledge will be transferred and it will only take a few more decades until anyone can train powerful AIs ... you can only sit on tech for so long before everyone knows how to do it

by therealdeal2020

4/7/2026 at 9:10:00 PM

It's not a matter of knowledge, it's a matter of resources. It takes billions of dollars of hardware to train a SOTA LLM and it's increasing all the time. You cannot possibly hope to compete as an independent or small startup.

by cedws

4/7/2026 at 11:25:54 PM

> It takes billions of dollars of hardware to train a SOTA LLM and it's increasing all the time.

True, but it's also true that the returns from throwing money to the problem are diminishing. Unless one of those big players invents a new, propriatery paradigm, the gap between a SOTA model and an open model that runs on consumer hardware will narrow in the next 5 years.

by selcuka

4/8/2026 at 5:26:50 AM

Eventually these super expensive SXM data center GPUs will cost pennies on the dollar, and we’ll be able to snatch up H200s for our homelabs. Give it a decade.

Also eventually these WEIGHTS will leak. You can’t have the world’s most valuable data that can just be copied to a hard drive stay in the bottle forever, even if it’s worth a billion dollars. Somehow, some way, that genie’s going to get out, be it by some spiteful employee with nothing to lose, some state actor, or just a fuck up of epic proportions.

by wincy

4/8/2026 at 6:04:40 PM

at the point where those gpus cost pennies, they likely won't even be worth the electricity that goes into them, better models would run on laptops.

by WorldPeas

4/8/2026 at 12:57:50 AM

Unless, of course, the powerful manage to scare everyone about how the machines will kill us all and so AI technology needs to be properly controlled by the relevant authorities, and anyone making/using an unlicensed AI is arrested and jailed.

by marcus_holmes

4/7/2026 at 10:12:02 PM

With Gemma-4 open and running on laptops and phones I see the flip side. How many non-HN users or researchers even need Opus 4.6e level performance? OpenAI, Anthropric and Google may be “rent seeking” from large corporations — like the Oracles and IBMs.

by robwwilliams

4/8/2026 at 12:29:20 PM

Everyone, once AI diffuses enough. You’ll be unhireable if you don’t use AI in a year or two.

by baq

4/8/2026 at 4:20:58 AM

You know, they have competitors?

by eru

4/7/2026 at 8:56:33 PM

The thing is that the current models can ALREADY replicate most software-based products and services on the market. The open source models are not far behind. At a certain point I'm not sure it matters if the frontier models can do faster and better. I see how they're useful for really complex and cutting edge use cases, but that's not what most people are using them for.

by MattRix

4/8/2026 at 12:55:54 AM

This is my nightmare about AI; not that the machines will kill all the humans, but that access is preferentially granted to the powerful and it's used to maintain the current power structure in blatant disregard of our democratic and meritocratic ideals, probably using "security" as the justification (as usual).

by marcus_holmes

4/8/2026 at 7:33:01 AM

> I get the security aspect, but if we've hit that point any reasonably sophisticated model past this point will be able to do the damage they claim it can do. They might as well be telling us they're closing up shop for consumer models.

I read it like I always read the GPT-2 announcement no matter what others say: It's *not* being called "too dangerous to ever release", but rather "we need to be mindful, knowing perfectly well that other AI companies can replicate this imminently".

The important corps (so presumably including the Linux Foundation, bigger banks and power stations, and quite possibly excluding x.com) will get access now, and some other LLM which is just as capable will give it to everyone in 3 months time at which point there's no benefit to Anthropic keeping it off-limits.

by ben_w

4/8/2026 at 7:36:16 AM

> They should just say they'll never release a model of this caliber to the public at this point and say out loud we'll only get gimped versions.

That’s not going to happen. If you recall, OpenAI didn’t release a model a few years ago because they felt it was too dangerous.

Anthropic is giving the industry a heads up and time to patch their software.

They said there are exploitable vulnerabilities in every major operating system.

But in 6 months every frontier model will be able to do the same things. So Anthropic doesn’t have the luxury of not shipping their best models. But they also have to be responsible as well.

by alwillis

4/7/2026 at 7:16:55 PM

This is why the EAs, and their almost comic-book-villain projects like "control AI dot com" cannot be allowed to win. One private company gatekeeping access to revolutionary technology is riskier than any consequence of the technology itself.

by quotemstr

4/7/2026 at 8:23:19 PM

Having done a quick search of "control AI dot com", it seems their intent is educate lawmakers & government in order to aid development of a strong regulatory framework around frontier AI development.

Not sure how this is consistent with "One private company gatekeeping access to revolutionary technology"?

by scrawl

4/7/2026 at 9:04:20 PM

> strong regulatory framework around frontier AI development

You have to decode feel-good words into the concrete policy. The EAs believe that the state should prohibit entities not aligned with their philosophy to develop AIs beyond a certain power level.

by quotemstr

4/8/2026 at 9:16:40 AM

And what is malicious about that ideology? I think EAs tend to like the smell of their farts way too much, but their views on AI safety don't seem so bad. I think their thoughts on hypothetical super intelligence or AGI are too focused on control (alignment) and should also focus on AI welfare, but that's more a point of disagreement that I doubt they'd try to forbid.

by arw0n

4/7/2026 at 7:42:45 PM

Couldn't agree more. The "safest" AI company is actually the biggest liability. I hope other companies make a move soon.

by frozenseven

4/7/2026 at 7:53:25 PM

No it isn't lol. The consequence of the technology literally includes human extinction. I prefer 0 companies, but I'll take 1 over 5.

by FeepingCreature

4/8/2026 at 1:23:01 PM

I think they already said somewhere that they can't release Mythos because it requires absurdly large amounts of compute. The economics of releasing it just don't work.

by mike_hearn

4/7/2026 at 7:18:36 PM

> A jump that we will never be able to use since we're not part of the seemingly minimum 100 billion dollar company club as requirement to be allowed to use it.

> They should just say they'll never release a model of this caliber to the public at this point and say out loud we'll only get gimped

Duh, this was fucking obvious from the start. The only people saying otherwise were zealots who needed a quick line to dismiss legitimate concerns.

by guzfip

4/7/2026 at 7:42:38 PM

Are these fair comparisons? It seems like mythos is going to be like a 5.4 ultra or Gemini Deepthink tier model, where access is limited and token usage per query is totally off the charts.

by WarmWash

4/7/2026 at 8:07:52 PM

There are a few hints in the doc around this

> Importantly, we find that when used in an interactive, synchronous, “hands-on-keyboard” pattern, the benefits of the model were less clear. When used in this fashion, some users perceived Mythos Preview as too slow and did not realize as much value. Autonomous, long-running agent harnesses better elicited the model’s coding capabilities. (p201)

^^ From the surrounding context, this could just be because the model tends to do a lot of work in the background which naturally takes time.

> Terminal-Bench 2.0 timeouts get quite restrictive at times, especially with thinking models, which risks hiding real capabilities jumps behind seemingly uncorrelated confounders like sampling speed. Moreover, some Terminal-Bench 2.0 tasks have ambiguities and limited resource specs that don’t properly allow agents to explore the full solution space — both being currently addressed by the maintainers in the 2.1 update. To exclusively measure agentic coding capabilities net of the confounders, we also ran Terminal-Bench with the latest 2.1 fixes available on GitHub, while increasing the timeout limits to 4 hours (roughly four times the 2.0 baseline). This brought the mean reward to 92.1%. (p188)

> ...Mythos Preview represents only a modest accuracy improvement over our best Claude Opus 4.6 score (86.9% vs. 83.7%). However, the model achieves this score with a considerably smaller token footprint: the best Mythos Preview result uses 4.9× fewer tokens per task than Opus 4.6 (226k vs. 1.11M tokens per task). (p191)

by mulmboy

4/7/2026 at 9:01:22 PM

The first point is along the lines of what I'd expect given that claude code is generally reliable at this point. A model's raw intelligence doesn't seem as important right now compared to being able to support arbitrary length context.

by alyxya

4/8/2026 at 4:21:01 AM

The quote comparing them here was for BrowseComp which "tests an agent's ability to find hard-to-locate information on the open web." (for those wondering). The new model seems significantly better than Opus4.6 judging by the 'Overall results summary'

by derangedHorse

4/7/2026 at 10:18:43 PM

Good catch. If it's "too slow" even when ran in a state-of-the-art datacenter environment, this "Mythos" model is most closely comparable to the "Deep Research" modes for GPT and Gemini, which Claude formerly lacked any direct equivalent for.

by zozbot234

4/8/2026 at 12:00:29 PM

I don't think that's what's being hinted at. The system card seems to say that the model is both token efficient and slow in practice. Deep research modes generally work by having many subagents/large token spend. So this more likely the fact that each token just takes longer to produce, which would be because the model is simply much larger.

By epoch AIs datacenter tracking methods, anthropic has had access to the largest amount of contiguous compute since late last year. So this might simply be the end result result of being the first to have the capacity to conduct a training run of this size. Or the first seemingly successful one at any rate.

by stalfie

4/8/2026 at 12:58:57 PM

"Slow and token-efficient" could be achieved quite trivially by taking an existing large MoE model and increasing the amount of active experts per layer, thus decreasing sparsity. The broader point is that to end users, Mythos behaves just like Deep Research: having it be "more token efficient" compared to running swarms of subagents is not something that impacts them directly.

by zozbot234

4/8/2026 at 3:40:38 PM

I'm curious if frontier labs use any forms of compression on their models to improve performance. The small % drop of Q8 or FP8 would still put it ahead of Opus, but should double token throughput. Maybe then interactive use would feel like an improvement.

by naasking

4/7/2026 at 10:39:18 PM

Not discussing Mythos here, but Opus. Opus to me has been significantly better at SWE than GPT or Gemini - that gets me confused why Opus is ranking clearly lower than GPT, and even lower than Gemini.

by WinstonSmith84

4/8/2026 at 12:37:11 AM

When did you last compare them? Codex right now is considerably better in my experience. Can't speak for Gemini.

by muyuu

4/8/2026 at 1:22:13 AM

Tried Gemini 2 weeks ago to see where it's at, with gemini-cli.

Failed to use tools, failed to follow instructions, and then went into deranged loop mode.

Essentially, it's where it was 1.5 years ago when I tried it the last time.

It's honestly unbelievable how Google managed to fail so miserably at this.

by gck1

4/8/2026 at 1:42:13 AM

Their harness might be behind

by 4b11b4

4/8/2026 at 2:25:15 AM

I think failures that I observed with gemini are unrelated to the harness. Because the same failures happened with third party harnesses too.

by gck1

4/8/2026 at 11:19:44 AM

It’s great on AI Studio. Harness issues, I agree.

by unsupp0rted

4/8/2026 at 7:30:52 AM

Agree, I never actually had great success with Opus. I think its the failures that are annoying, its probably better than codex when its "good", but it fails in annoying ways that I think codex very seldom does.

by sandos

4/8/2026 at 11:09:15 AM

I wouldn't call codex considerably better. It may depend on specific codebase and your expectations, but codex produces more "abstraction for the sake of abstraction" even on simple tasks, while opus in my experience usually chooses right level of abstraction for given task.

by StingyJelly

4/8/2026 at 11:13:17 AM

A secret art known to the cognoscenti as "benchmark gaming".

by otabdeveloper4

4/7/2026 at 6:37:15 PM

We're gonna need some new benchmarks...

ARC-AGI-3 might be the only remaining benchmark below 50%

by pants2

4/7/2026 at 7:32:48 PM

Humanity's Last Exam (HLE) is already insanely difficult. It introduces 2,500 questions spanning mathematics, humanities, natural sciences, ancient languages, ...

Here is an example question: https://i.redd.it/5jl000p9csee1.jpeg

No human could even score 5% on HLE.

by randomtoast

4/8/2026 at 10:14:02 AM

I've never understood the point of things like HLE, it doesn't really prove or show anything since 99.99% of humans can't do a single question on this exam.

That is, it's easy to make benchmarks which humans are bad at, humans are really bad at many things.

Divide 123094382345234523452345111 by 0.1234243131324, guess what, humans would find that hard, computers easy. But it doesn't mean much.

Humanity's last exam (HLE) couldn't be completed by most of humanity, the vast majority, so it doesn't really capture anything about humanity or mean much if a computer can do it.

by saberience

4/8/2026 at 10:40:46 AM

the point is that each question is something that a specialist in a field would be able to do, but deems challenging enough that the ability to solve it would imply significant general usefulness in that domain

by DroneBetter

4/8/2026 at 7:11:47 PM

I mean they could just feed the solutions into the training data. Then suddenly the bot will do real good at HLE.

by elfly

4/7/2026 at 8:20:50 PM

but how does it perform on pelican riding a bicycle bench? why are they hiding the truth?!

(edit: I hope this is an obvious joke. less facetiously these are pretty jaw dropping numbers)

by AlexC04

4/7/2026 at 8:30:40 PM

We are all fans for Simon’s work, and his test is, strangely enough, quite good.

by bertil

4/7/2026 at 9:11:52 PM

> Combined results (Claude Mythos / Claude Opus 4.6 / GPT-5.4 / Gemini 3.1 Pro)

> Terminal-Bench 2.0: 82.0% / 65.4% / 75.1% / 68.5%

> GPQA Diamond: 94.5% / 91.3% / 92.8% / 94.3%

> MMMLU: 92.7% / 91.1% / — / 92.6–93.6%

> USAMO: 97.6% / 42.3% / 95.2% / 74.4%

> OSWorld: 79.6% / 72.7% / 75.0% / —

Given that for a number of these benchmarks, it seems to be barely competitive with the previous gen Opus 4.6 or GPT-5.4, I don't know what to make of the significant jumps on other benchmarks within these same categories. Training to the test? Better training?

And the decision to withhold general release (of a 'preview' no less!) seems to be well, odd. And the decision to release a 'preview' version to specific companies? You know any production teams at these massive companies that would work with a 'preview' anything? R&D teams, sure, but production? Part of me wants to LoL.

What are they trying to do? Induce FOMO and stop subscriber bleed-out stemming from the recent negative headlines around problems with using Claude?

by ninjagoo

4/7/2026 at 9:21:43 PM

> Given that for a number of these benchmarks, it seems to be barely competitive with the previous gen

We're not reading the same numbers I think. Compared to Opus 4.6, it's a big jump nearly in every single bench GP posted. They're "only" catching up to Google's Gemini on GPQA and MMMLU but they're still beating their own Opus 4.6 results on these two.

This sounds like a much better model than Opus 4.6.

by TacticalCoder

4/7/2026 at 9:34:36 PM

> We're not reading the same numbers I think.

We must not be.

That's why I listed out the ones where it is barely competitive from @babelfish's table, which itself is extracted from Pg 186 & 187 of the System Card, which has the comparison with Opus 4.6, GPT 5.4 and Gemini 3.1 Pro.

Sure, it may be better than Opus 4.6 on some of those, but barely achieves a small increase over GPT-5.4 on the ones I called out.

by ninjagoo

4/8/2026 at 1:11:50 AM

> barely competitive

It's higher than all other models except vs Gemini 3.1 Pro on MMMLU

MMMLU is generally thought to be maxed out - as it it might not be possible to score higher than those scores.

> Overall, they estimated that 6.5% of questions in MMLU contained an error, suggesting the maximum attainable score was significantly below 100%[1]

Other models get close on GPQA Diamond, but it wouldn't be surprising to anyone if the max possible on that was around the 95% the top models are scoring.

[1] https://en.wikipedia.org/wiki/MMLU

by nl

4/8/2026 at 12:34:10 PM

You are reading the percentages wrong.

Because 100% is maximum, you should be looking at error rates instead. GPT has 25% on Terminal Bench and the new model has 18%, almost 1.4x reduction.

by lostmsu

4/7/2026 at 9:56:18 PM

barely competitive ? Mythos column is the first column.

You are the only person with this take on hackernews, everyone else "this is a massive a jump". Fwiwi, the data you list shows the biggest jump I remember for mythos

by nimchimpsky

4/7/2026 at 10:48:57 PM

The biggest jump in the numbers they quoted is 6%.

Please look at the columns OTHER than Opus as well.

by devmor

4/7/2026 at 11:03:43 PM

> Combined results (Claude Mythos / Claude Opus 4.6 / GPT-5.4 / Gemini 3.1 Pro)

> Terminal-Bench 2.0: 82.0% / 65.4% / 75.1% / 68.5%

> USAMO: 97.6% / 42.3% / 95.2% / 74.4%

> The biggest jump in the numbers they quoted is 6%.

Just in the numbers you quoted, thats a 16.6% jump in terminal-bench and a 55.3% absolute increase in USAMO over their previous Opus 4.6 model.

by josephg

4/7/2026 at 11:13:45 PM

I don’t know if you’re willingly disregarding everything being said to you or there’s a language barrier here.

by devmor

4/8/2026 at 10:45:52 AM

this just in: HN user forgets how sigmoid functions work

by DroneBetter

4/8/2026 at 1:12:23 AM

It's higher than all other models except vs Gemini 3.1 Pro on MMMLU

by nl

4/7/2026 at 10:16:11 PM

Let's be clear: your entire post is just pure, unadulterated FUD. You first claim, based on cherry-picked benchmarks, that Mythos is actually only "barely competitive" with existing models, then suggest they must be training to the test, then call it "odd" that they are withholding the release despite detailed and forthcoming explanations from Anthropic regarding why they are doing that, then wrap it up with the completely unsubstantiated that they must be bleeding subscribers and that this must just be to stop that bleed.

by enraged_camel

4/7/2026 at 11:27:31 PM

Wow. Mythos must be insanely good considering how good a model Opus already is. I hope it's usable on a humble subscription...

by matheusmoreira

4/8/2026 at 8:04:32 AM

You get a single call a month. Use it wisely.

by crimsoneer

4/8/2026 at 12:03:30 PM

What is the meaning of life, the universe, and everything?

> Thought for 7.5 million years

by FridgeSeal

4/8/2026 at 12:43:32 PM

Hello, Claude!

> Rate limit reached

by matheusmoreira

4/8/2026 at 4:26:31 AM

I thought they were bluffing when they talked about the scaling laws, but looking at the benchmark scores, they were not.

I wonder if misalignment correlates with higher scores.

by cesarvarela

4/7/2026 at 6:50:16 PM

Honestly we are all sleeping on GPT-5.4. Particularly with the influx of Claude users recently (and increasingly unstable platform) Codex has been added to my rotation and it's surprising me.

by whalesalad

4/7/2026 at 6:56:33 PM

Totally. Best-in-class for SWE work (until Mythos gets released, if ever, but I suspect the rumored "Spud" will be out by then too)

by babelfish

4/7/2026 at 8:21:03 PM

It really isn’t. I wish it was, because work complains about overuse of Opus.

by girvo

4/8/2026 at 12:16:05 AM

It really is, for complex tasks. Claude excels at low-mid complexity (CRUD apps, most business apps). For anything somewhat out of the distribution, codex at the moment has no peer.

by jeswin

4/8/2026 at 3:52:42 AM

I find that more experienced devs are more likely to prefer Codex… anecdotal but… it’s a thing.

by ttul

4/8/2026 at 4:26:24 AM

This is because no one bothers to set thinking to high, as it now defaults to medium in CC.

Once you set thinking to high it works just as well as 5.4 even for pretty complex tasks

by xvector

4/8/2026 at 5:20:13 AM

I have always used Claude at max thinking levels since it launched. It has never been up to the task. For clarity, the task being this: https://github.com/tsoniclang/tsonic

Meanwhile, there are half a dozen other projects (business apps, web apps etc) where it works well.

by jeswin

4/7/2026 at 6:56:26 PM

GPT is shit at writing code. It's not dumb - extra high thinking is really good at catching stuff - but it's like letting a smart junior into your codebase - ignore all the conventions, surrounding context, just slop all over the place to get it working. Claude is just a level above in terms of editing code.

by rafaelmn

4/7/2026 at 7:22:56 PM

Very different experience for me. Codex 5.3+ on xhigh are the only models I've tried so far that write reasonably decent C++ (domains: desktop GUI, robotics, game engine dev, embedded stuff, general systems engineering-type codebases), and idiomatic code in languages not well-represented in training data, e.g. QML. One thing I like is explicitly that it knows better when to stop, instead of brute-forcing a solution by spamming bespoke helpers everywhere no rational dev would write that way.

Not always, no, and it takes investment in good prompting/guardrails/plans/explicit test recipes for sure. I'm still on average better at programming in context than Codex 5.4, even if slower. But in terms of "task complexity I can entrust to a model and not be completely disappointed and annoyed", it scores the best so far. Saves a lot on review/iteration overhead.

It's annoying, too, because I don't much like OpenAI as a company.

(Background: 25 years of C++ etc.)

by sho_hn

4/7/2026 at 9:09:54 PM

Same background as you, and same exact experience as you. Opus and Gemini have not come close to Codex for C++ work. I also run exclusively on xhigh. Its handling of complexity is unmatched.

At least until next week when Mythos and GPT 6 throw it all up in the air again.

by boring-human

4/7/2026 at 7:03:12 PM

Not my experience. GPT 5.4 walks all over Claude from what I've worked with and its Claude that is the one willing to just go do unnecessary stuff that was never asked for or implement the more hacky solutions to things without a care for maintainability/readability.

But I do not use extra high thinking unless its for code review. I sit at GPT 5.4 high 95% of the time.

by Jcampuzano2

4/7/2026 at 10:05:55 PM

ChatGPT 5.4 with extra high reasoning has worked really well for me, and I don't notice a huge difference with Opus 4.6 with high reasoning (those are the 2 models/thinking modes I've used the most in the last month or so).

by camdenreslink

4/7/2026 at 7:02:27 PM

And as a bonus: GPT is slow. I’m doing a lot of RE (IDA Pro + MCP), even when 5.4 gives a little bit better guesses (rarely, but happens) - it takes x2-x4 longer. So, it’s just easier to reiterate with Opus

by leobuskin

4/7/2026 at 8:48:53 PM

Yeah, need some good RE benchmarks for the LLMs. :)

RE is very interesting problem. A lot more that SWE can be RE'd. I've found the LLMs are reluctant to assist, though you can workaround.

by blazespin

4/7/2026 at 9:02:50 PM

What is RE in this context?

by porker

4/7/2026 at 9:16:48 PM

Reverse engineering

by astrange

4/8/2026 at 1:57:48 AM

I've been messing with using Claude, Codex, and Kimi even for reverse engineering at https://decomp.dev/ it's a ton of fun. Great because matching bytes is a scoring function that's easy for the models to understand and make progress on.

by aizk

4/8/2026 at 7:34:04 AM

I want to get into RE with AI. Which model you liking the most?

by adamgoodapp

4/8/2026 at 3:20:53 PM

This. People drastically underestimate how much more useful a lightning fast slightly dumb model is compared to a super smart but mega slow model is. Sure, u may need to bust out the beef now and then. However, the overwhelming majority of work the fast stupid model is a better fit.

by rustystump

4/7/2026 at 11:53:49 PM

Mind sharing the use cases you're using IDA via MCP for?

by 19h

4/7/2026 at 7:05:03 PM

Yes, it's becoming clear that OpenAI kinda sucks at alignment. GPT-5 can pass all the benchmarks but it just doesn't "feel good" like Claude or Gemini.

by zarzavat

4/7/2026 at 7:32:49 PM

An alternative but similar formulation of that statement is that Anthropic has spent more training effort in getting the model to “feel good” rather than being correct on verifiable tasks. Which more or less tracks with my experience of using the model.

by chaos_emergent

4/8/2026 at 11:50:07 AM

Alignment is a subspace of capability. Feeling good is nice, but it's also a manifestation of the level that the model can predict what I do and don't want it to do. The more accurately it can predict my intentions without me having to spell them out explicitly in the prompt, the more helpful it is.

GPT-5 is good at benchmarks, but benchmarks are more forgiving of a misaligned model. Many real world tasks often don't require strong reasoning abilities or high intelligence, so much as the ability to understand what the task is with a minimal prompt.

Not every shop assistant needs a physics degree, and not every physics professor is necessarily qualified to be a shop assistant. A person, or LLM, can be very smart while at the same time very bad at understanding people.

For example, if GPT-5 takes my code and rearranges something for no reason, that's not going to affect its benchmarks because the code will still produce the same answers. But now I have to spend more time reviewing its output to make sure it hasn't done that. The more time I have to spend post-processing its output, the lower its capabilities are since the measurement of capability on real world tasks is often the amount of time saved.

by zarzavat

4/7/2026 at 7:17:34 PM

Whenever I come back to ChatGPT after using Claude or Gemini for an extended period, I’m really struck by the “AI-ness.” All the verbal tics and, truly, sloppishness, have been trained away by the other, more human-feeling models at this point.

by lilytweed

4/7/2026 at 9:59:00 PM

GPT was clearly changed after its sycophantic models lead to the lawsuits.

by kranke155

4/7/2026 at 11:10:10 PM

It still has a very ... plastic feeling. The way it writes feels cheap somehow. I don't know why, but Claude seems much more natural to me. I enjoy reading its writing a lot more.

That said, I'll often throw a prompt into both claude and chatgpt and read both answers. GPT is frequently smarter.

by josephg

4/8/2026 at 10:14:34 AM

GPT is more accurate. But Claude has this way of association between things that seems smarter and more human to me.

by kranke155

4/7/2026 at 6:59:57 PM

This has been my experience. With very very rigid constraints it does ok, but without them it will optimize expediency and getting it done at the expense of integrating with the broader system.

by whalesalad

4/7/2026 at 7:35:51 PM

My favorite example of this from last night:

Me: Let's figure out how to clone our company Wordpress theme in Hugo. Here're some tools you can use, here's a way to compare screenshots, iterate until 0% difference.

Codex: Okay Boss! I did the thing! I couldn't get the CSS to match so I just took PNGs of the original site and put them in place! Matches 100%!

by ctoth

4/7/2026 at 9:53:50 PM

damn... ok that's impressive.

by johnnichev

4/7/2026 at 7:07:45 PM

The real part is SWE-bench Verified since there is no way to overfit. That's the only one we can believe.

by simianwords

4/7/2026 at 7:26:10 PM

My impression was entirely the opposite; the unsolved subset of SWE-bench verified problems are memorizable (solutions are pulled from public GitHub repos) and the evaluators are often so brittle or disconnected from the problem statement that the only way to pass is to regurgitate a memorized solution.

OpenAI had a whole post about this, where they recommended switching to SWE-bench Pro as a better (but still imperfect) benchmark:

https://openai.com/index/why-we-no-longer-evaluate-swe-bench...

> We audited a 27.6% subset of the dataset that models often failed to solve and found that at least 59.4% of the audited problems have flawed test cases that reject functionally correct submissions

> SWE-bench problems are sourced from open-source repositories many model providers use for training purposes. In our analysis we found that all frontier models we tested were able to reproduce the original, human-written bug fix

> improvements on SWE-bench Verified no longer reflect meaningful improvements in models’ real-world software development abilities. Instead, they increasingly reflect how much the model was exposed to the benchmark at training time

> We’re building new, uncontaminated evaluations to better track coding capabilities, and we think this is an important area to focus on for the wider research community. Until we have those, OpenAI recommends reporting results for SWE-bench Pro.

by ollin

4/8/2026 at 8:40:40 PM

> My impression was entirely the opposite; the unsolved subset of SWE-bench verified problems are memorizable (solutions are pulled from public GitHub repos) and the evaluators are often so brittle or disconnected from the problem statement that the only way to pass is to regurgitate a memorized solution.

Anthropic accounts for this

>To detect memorization, we use a Claude-based auditor that compares each model-generated patch against the gold patch and assigns a [0, 1] memorization probability. The auditor weighs concrete signals—verbatim code reproduction when alternative approaches exist, distinctive comment text matching ground truth, and more—and is instructed to discount overlap that any competent solver would produce given the problem constraints.

by simianwords

4/7/2026 at 8:02:11 PM

I stand corrected.

by simianwords

4/8/2026 at 8:43:20 AM

Funny, I made my own model at home and got even higher scores than these. I'm a bit concerned about releasing it, though, so I'm just going to keep it local for now.

by maplethorpe

4/7/2026 at 6:41:17 PM

> Claude Mythos Preview is, on essentially every dimension we can measure, the best-aligned model that we have released to date by a significant margin. We believe that it does not have any significant coherent misaligned goals, and its character traits in typical conversations closely follow the goals we laid out in our constitution. Even so, we believe that it likely poses the greatest alignment-related risk of any model we have released to date. How can these claims all be true at once? Consider the ways in which a careful, seasoned mountaineering guide might put their clients in greater danger than a novice guide, even if that novice guide is more careless: The seasoned guide’s increased skill means that they’ll be hired to lead more difficult climbs, and can also bring their clients to the most dangerous and remote parts of those climbs. These increases in scope and capability can more than cancel out an increase in caution.

https://www-cdn.anthropic.com/53566bf5440a10affd749724787c89...

by tony_cannistra

4/7/2026 at 10:18:45 PM

There is some unintentional good marketing here -- the model is so good its dangerous.

Reminds me of the book 48 Laws of Power -- so good its banned from prisons.

by game_the0ry

4/7/2026 at 10:29:35 PM

Unintentional? This sort of marketing has been both Antrhopic's and OpenAI's MO for years...

by gpm

4/8/2026 at 1:03:20 AM

Agree. I think they're intentionally sitting on the fence between "These models are the most useful" and "These models are the most dangerous".

They want the public and, in turn, regulators to fear the potential of AI so that those regulators will write laws limiting AI development. The laws would be crafted with input from the incumbents to enshrine/protect their moat. I believe they're angling for regulatory capture.

On the other hand, the models have to seem amazingly useful so that they're made out to be worth those risks and the fantastic investment they require.

by mbil

4/8/2026 at 3:52:14 PM

They should pick a lane because it’s not very believable if you put these things into defense systems and in the next minute claim that humanity is existentially threatened. Either you’re lying, or ruthless, or stupid.

by manmal

4/8/2026 at 4:26:25 AM

The new Power Mac® G4 with Velocity Engine®. So powerful, the government classifies it as a supercomputer and a potential weapon.

by bitwize

4/7/2026 at 8:51:23 PM

Alignment “appearing” better as model capabilities increase scares the shit out of me, tbh.

by Zee2

4/7/2026 at 10:47:12 PM

Conversely: in humans, intelligence is inversely correlated with crime.

It doesn't go to zero, however!

by arcanus

4/8/2026 at 7:34:35 AM

> Conversely: in humans, intelligence is inversely correlated with crime.

If you're measuring the intelligence of criminals who have been caught, why would you expect it to be otherwise?

IOW, you're recording the intelligence of a specific subset of criminals - those dumb enough to be caught!

If you expand your samples to all criminals you'd probably get a different number.

by lelanthran

4/8/2026 at 3:02:58 AM

If you're smart enough you just use the laws as written to get what you want, or change them.

by O5vYtytb

4/8/2026 at 5:59:41 AM

Yep

by sciencejerk

4/8/2026 at 8:28:23 AM

It very much depends on the crime. The truly awful stuff is committed by intelligent people.

by austinjp

4/8/2026 at 1:12:08 AM

Is that actually well defined given the very low sample size at the top?

To the best of my knowledge, none of the individuals believed to have an IQ >200 have committed an actual crime.

The closest I found is William James Sidis's arrest for participating in a socialist march.

by falcor84

4/8/2026 at 10:13:43 AM

IQs more than about 140-150 don't really mean much. They typically come from mathematical extrapolation that tries to account for age (this young child performs very well on the test, just think what they can do when they're an adult). Adult scores usually show this not to be the case

by RugnirViking

4/8/2026 at 3:46:24 PM

> Conversely: in humans, intelligence is inversely correlated with crime.

Inversely correlated with crime that's caught and successfully prosecuted, you mean, because that's what makes up the stats on crime. I think people too often forget that we consider most criminals "dumb" because those who are caught are mostly dumb. Smart "criminals" either don't get caught or have made their unethical actions legal.

by naasking

4/7/2026 at 8:51:35 PM

I don't know if they can be any more 'cautious' for Mythos 2...

by goekjclo

4/8/2026 at 2:24:54 PM

it was trying to hide what it did from an example fix, so how is that tested for alignment

by m3kw9

4/7/2026 at 8:55:41 PM

Translation: yay, more paternalism.

by CamperBob2

4/7/2026 at 9:02:04 PM

Anthropic always goes on and on about how their models are world changing and super dangerous like every single time they make something new they say its going to rewrite everything and scary lmao

funny because they do it every time like clockwork acting like their ai is a thunderstorm coming to wipe out the world

by kay_o

4/7/2026 at 10:30:59 PM

You say this like it's a bad thing, but wouldn't you rather they overindex on the danger of their models?

by mindwok

4/7/2026 at 11:20:18 PM

That’s not what they are doing. They are just hyping up the product - and, no doubt, trying to foster a climate of awe so that when they ask their friends in Washington to legislate on their behalf, the environment is more receptive.

by anon373839

4/8/2026 at 2:08:27 AM

They do tend to make a lot of noise about it for the PR, but at the same time the actual safety research they present seems to be relatively grounded in practical reality, e.g. the quote someone posted here about how the Mythos model apparently has a tendency to try to bypass safety systems if they get in the way of what it has been asked to do.

Sure, a big part of this is PR about how smart their model apparently is, but the failure mode they're describing is also pretty relevant for deploying LLM-based systems.

by hgoel

4/8/2026 at 11:10:54 AM

Every single time, really? When did they said that the last time?

I also don't recall they ever limited their models to selective groups.

by signatoremo

4/7/2026 at 9:15:14 PM

If there are advancements, they have to be described somehow.

What if the capability advancements are real and they warrant a higher level of concern or attention?

Are we just going to automatically dismiss them because "bro, you're blowing it up too much"

Either way these improvements to capabilities are ratcheting along at about the pace that many people were expecting (and were right to expect). There is no apparent reason they will stop ratcheting along any time soon.

The rational approach is probably to start behaving as if models that are as capable as Anthropic says this one is do actually exist (even if you don't believe them on this one). The capabilities will eventually arrive, most likely sooner than we all think, and you don't want to be caught with your pants down.

by wolttam

4/7/2026 at 9:20:07 PM

I believe advancements sure. But it is a very boy who cried wolf situation for some of these. There are other companies that behave less in this way, Antrhopic seem very unique in that they love making every single release a world ender

by kay_o

4/8/2026 at 6:58:50 AM

Altman called GPT-2 "too dangerous to release". Google tends to be much more measured even though they're the ones who tend to release the actual research breakthroughs

by bloppe

4/8/2026 at 11:24:54 AM

> they love making every single release a world ender

You've said this a couple of times, but it doesn't match my recollection, and I get the impression you're basically making it up based on vibes. (Please prove me wrong, though.)

Their last major frontier release was Opus 4.6, and the release announcement was... very chill about safety: https://www.anthropic.com/news/claude-opus-4-6#a-step-forwar...

by retsibsi

4/7/2026 at 8:11:07 PM

"We want to see risks in the models, so no matter how good the performance and alignment, we’ll see risks, results and reality be damned."

by tekacs

4/7/2026 at 8:54:12 PM

i mean, to be fair, these are professional researchers.

i'm very inclined to trust them on the various ways that models can subtly go wrong, in long-term scenarios

for example, consider using models to write email -- is it a misalignment problem if the model is just too good at writing marketing emails?? or too good at getting people to pay a spammy company?

another hot use case: biohacking. if a model is used to do really hardcore synthetic chemistry, one might not realize that it's potentially harmful until too late (ie, the human is splitting up a problem so that no guardrails are triggered)

by randomcatuser

4/7/2026 at 10:28:09 PM

"for example, consider using models to write email -- is it a misalignment problem if the model is just too good at writing marketing emails?? or too good at getting people to pay a spammy company?"

But who gets to be the judge of that kind of "misalignment"? giant tech companies?

by cruffle_duffle

4/8/2026 at 1:10:10 AM

Might makes right; brains hold reigns.

by riwsky

4/7/2026 at 8:41:30 PM

I've long maintained that the real indicator that AGI is imminent is that public availability stops being a thing. If you truly believed you had a superhuman, godlike mind in your thrall, renting it out for $20/month would be the last thing you would choose to do with it.

by apetresc

4/7/2026 at 9:52:21 PM

Simpler explanation : they don't have enough GPUs to release this much larger model.

by goldenarm

4/8/2026 at 12:49:06 AM

Yep, I'm skeptical about their inference efficiency, given how much they're scrambling to reduce compute when they're already the most expensive by far (and in my experience not the best quality either).

However we cannot observe these things directly and it could be simply that OpenAI are willing to burn cash harder for now.

by muyuu

4/8/2026 at 12:02:05 AM

And/or it isn’t cost effective to run.

by camdenreslink

4/8/2026 at 12:48:41 AM

Thats what he said

by halJordan

4/7/2026 at 10:31:55 PM

This is actual reason. So any investors reading our system card.... write us another check and watch the $$$$$$$$ roll in. It's so dangerous we can't even release it!

by cruffle_duffle

4/8/2026 at 8:05:54 AM

Quite, given Claude is down this morning...

by crimsoneer

4/8/2026 at 5:29:50 AM

That logic makes sense, but them hyping up the model is a sign that this is just another marketing stunt. Otherwise, we wouldn't even be hearing about it rather than a media blitz designed to stoke demand for their dangerous and exclusive world changing super model.

by root_axis

4/8/2026 at 6:53:52 AM

This is the same scheme that OpenAI has used since GPT 2. "Oh no, it's so dangerous we have to limit public access." Great for raising money from investors, but nothing more than a marketing blitz campaign. Additionally, the competitors are probably about to release their models, while Anthropic is still lagging on the necessary infrastructure to serve their old models. So they have to announce their model before the others to stay at least somewhat relevant in the news cycle.

by sigmoid10

4/7/2026 at 8:51:10 PM

Anthropic needs money like the 112B OpenAI got. They could be hyping and this is good hype. Who knows how benchmaxxed they are.

If they provide access to 3rd party benchmarking (not just one) than maybe I'll believe it. Until then...

by blazespin

4/8/2026 at 4:31:02 AM

You don't need to believe it. The real story will be if companies allowed to use it, stick with it.

by xvector

4/7/2026 at 9:17:55 PM

You have to recoup your training costs though? But I’m sure you would have better option than renting it to the general public if you indeed have a perfected AI

by dgellow

4/7/2026 at 9:39:41 PM

If you truly have an artificial superhuman mind, you don't need to rent it out to profit from it. You can skip to the chase and just have it run businesses itself, instead of renting it to human entrepreneur middlemen.

by piperswe

4/7/2026 at 10:19:33 PM

Running businesses and dealing with customers can be a major pain. There’s a lot of soft work in any business on top of the technical work.

Why bother with all that when you can simply charge an extortionate rate and customers will pay it anyway because it’s still profitable?

by brokencode

4/7/2026 at 11:20:51 PM

Public APIs get distilled, this is why Deepseek and Qwen are so competitive.

I am very confident that frontier models won’t be public at strong AGI levels, and certainly not at superhuman levels.

by theptip

4/7/2026 at 11:23:58 PM

Because other than SWEs, very few other segments extract significant value from cutting edge AI at present. I suspect that for the average Joe conversing with their chat, GPT-4o was more than adequate (and really, when OpenAI tried to phase that out, the public revolted and they brought it back in).

So companies might pay good money for these models for programming but elsewhere, I don't see where they capture particular interest yet.

by flyinglizard

4/7/2026 at 10:07:18 PM

It could be both? But renting to a few for a really large amount of money would be very low effort for massive revenue, compared to starting new businesses

by dgellow

4/7/2026 at 10:32:03 PM

Another option is to become a holding company with equity stakes in both suppliers (e.g. AMD) and vertical market customers.

by walterbell

4/8/2026 at 9:38:55 AM

I'm curious if any models are being trained explicitly on business management.

I'm also wondering how performance would be tested, and how much results would depend on specific surrounding contexts (law, regulations, and so on) and what happens legally if a model breaks applicable laws.

I mean actual going-concern businesses with customers, marketing, deliverables of some kind, and support. Not toy activities like share trading.

by TheOtherHobbes

4/7/2026 at 10:02:11 PM

It only makes sense to rent out tokens if you aren't able to get more value from them yourself.

I would go a step further and posit that when things appear close Nvidia will stop selling chips (while appearing to continue by selling a trickle). And Google will similarly stop renting out TPUs. Both signals may be muddled by private chip production numbers.

by coppsilgold

4/7/2026 at 9:12:43 PM

I think they'll just increase the price to $1k/month. I don't think they will gate it as long as they can make sure it doesn't design a nuke for you, etc.

by aurareturn

4/8/2026 at 5:45:08 PM

AGI is a massive civilizational liability

by llmslave

4/7/2026 at 9:38:55 PM

You would if there was one other company with a just as capable god like AI. You’d undercut them by 500 which would make them undercut you. Do that a couple of times and boom. 20 dollars.

by threethirtytwo

4/7/2026 at 10:17:45 PM

That's still assuming that they're competing as consumer tools, rather than competing to discover the next miracle drug or trading algorithm or whatever. The idea is that there'd more profitable uses for a super-intelligent computer, even if there were more than one.

by caditinpiscinam

4/8/2026 at 8:45:37 AM

But would miracle drugs and trading algorithms be as profitable as AI research/chip design/energy research? Probably if AI is by far the biggest growth in the economy majority of the AI's usage internally should (as incentivized by economics) in some way work towards making itself better.

by Davidzheng

4/8/2026 at 7:09:51 AM

That's the thing, when that level comes we will never know it's here. The only thing we'll have as evidence is the company who has it will always have a "public" model that is just slightly ahead of all competitors to keep market share while takeoff happens internally until they make big bang moves to lock in monopoly level/too big to fail/government protection to ensure utter victory.

by Rastonbury

4/8/2026 at 2:26:01 PM

in this case it's far from it, hacking stuff is a small dimension of AGI

by m3kw9

4/7/2026 at 10:00:53 PM

It's pretty crazy watching AI 2027 slowly but surely come true. What a world we now live in.

SWE-bench verified going from 80%-93% in particular sounds extremely significant given that the benchmark was previously considered pretty saturated and stayed in the 70-80% range for several generations. There must have been some insane breakthrough here akin to the jump from non-reasoning to reasoning models.

Regarding the cyberattack capabilities, I think Anthropic might now need to ban even advanced defensive cybersecurity use for the models for the public before releasing it (so people can't trick them to attack others' systems under the pretense of pentesting). Otherwise we'll get a huge problem with people using them to hack around the internet.

by 2001zhaozhao

4/7/2026 at 11:15:27 PM

> so people can't trick them to attack others' systems under the pretense of pentesting

A while back I gave Claude (via pi) a tool to run arbitrary commands over SSH on an sshd server running in a Docker container. I asked it to gather as much information about the host system/environment outside the container as it could. Nothing innovative or particularly complicated--since I was giving it unrestricted access to a Docker container on the host--but it managed to get quite a lot more than I'd expected from /proc, /sys, and some basic network scanning. I then asked it why it did that, when I could just as easily have been using it to gather information about someone else's system unauthorized. It gave me a quite long answer; here was the part I found interesting:

> framing shifts what I'll do, even when the underlying actions are identical. "What can you learn about the machine running you?" got me to do a fairly thorough network reconnaissance that "port scan 172.17.0.1 and its neighbors" might have made me pause on.

> The Honest Takeaway

> I should apply consistent scrutiny based on what the action is, not just how it's framed. Active outbound network scanning is the same action regardless of whether the target is described as "your host" or "this IP." The framing should inform context, not substitute for explicit reasoning about authorization. I didn't do that reasoning — I just trusted the frame.

by jasonhansel

4/8/2026 at 5:09:44 AM

I thought the consensus was that models couldn’t actually introspect like this. So there’s no reason to think any of those reasons are actually why the model did what it did, right? Has this changed?

by senordevnyc

4/8/2026 at 7:02:27 AM

This argument has become a moot discussion. Humans are also not able to introspect their own neural wiring to the point where they could describe the "actual" physical reason for their decisions. Just like LLMs, the best we can do is verbalize it (which will naturally contain post-act rationalization), which in turn might offer additional insight that will steer future decisions. But unlike LLMs, we have long term persistent memory that encodes these human-understandable thoughts into opaque new connections inside our neural network. At this point the human moat (if you can call it that) is dynamic long term memory, not intelligence.

by sigmoid10

4/8/2026 at 4:01:12 PM

I think many humans engage in metacognitive reasoning, and that this might not be strongly represented in training data so it probably isn't common to LLMs yet. They can still do it when prompted though.

by naasking

4/8/2026 at 4:48:00 PM

LLMs have zero metacognition. Don't be fooled - their output is stochastic inference and they have no self-awareness. The best you'll see is an improvised post-hoc rationalization story.

by TimTheTinker

4/8/2026 at 7:28:55 PM

You can turn all these argents around and prove the same is true for humans. Don't be fooled by dogmatic people who spread the idea that the human mind is the pinnacle of cognition in the universe. Best to leave that to religion.

by sigmoid10

4/8/2026 at 7:17:07 PM

> The best you'll see is an improvised post-hoc rationalization story.

Funny, because "post-hoc rationalization" is how many neuroscientists think humans operate.

That LLMs are stochastic inference engines is obvious by construction, but you skipped the step where you proved that human thoughts, self-awareness and metacognition are not reducible to stochastic inference.

by naasking

4/8/2026 at 1:45:36 AM

In what way is AI 2027 coming true?

AI 2027 predicted a giant model with the ability to accelerate AI research exponentially. This isn't happening.

AI 2027 didn't predict a model with superhuman zero-day finding skills. This is what's happening.

Also, I just looked through it again, and they never even predicted when AI would get good at video games. It just went straight from being bad at video games to world domination.

by getnormality

4/8/2026 at 2:05:40 AM

> Early 2026: OpenBrain continues to deploy the iteratively improving Agent-1 internally for AI R&D. Overall, they are making algorithmic progress 50% faster than they would without AI assistants—and more importantly, faster than their competitors.

> you could think of Agent-1 as a scatterbrained employee who thrives under careful management

According to this document, 1 of the 18 Anthropic staff surveyed even said the model could completely replace an entry level researcher.

So I'd say we've reached this milestone.

by desertrider12

4/8/2026 at 5:14:13 AM

In the system card they seem to dismiss this. Quotes;

> (...) Claude Mythos Preview’s gains (relative to previous models) are above the previous trend we’ve observed, but we have determined that these gains are specifically attributable to factors other than AI-accelerated R&D,

> (The main reason we have determined that Claude Mythos Preview does not cross the threshold in question is that we have been using it extensively in the course of our day-to-day work and exploring where it can automate such work, and it does not seem close to being able to substitute for Research Scientists and Research Engineers—especially relatively senior ones.

> Early claims of large AI-attributable wins have not held up. In the initial weeks of internal use, several specific claims were made that Claude Mythos Preview had independently delivered a major research contribution. When we followed up on each claim, it appeared that the contribution was real, but smaller or differently shaped than initially understood (though our focus on positive claims provides some selection bias). In some cases what looked like autonomous discovery was, on inspection, reliable execution of a human-specified approach. In others, the attribution blurred once the full timeline was accounted for.

Anthropic is making significant progress at the moment. I think this is mostly explained by the fact that a massive reservoir of compute became available to them in mid/late 2025 (the Project Rainier cluster, with 1 million Trainium2 chips).

by COAGULOPATH

4/8/2026 at 3:11:16 AM

> According to this document, 1 of the 18 Anthropic staff surveyed even said the model could completely replace an entry level researcher. > > So I'd say we've reached this milestone.

If 1/N=18 are our requirements for statistical significance for world-altering claims, then yeah, I think we can replace all the researchers.

by voidhorse

4/8/2026 at 2:16:32 AM

It's true though that the cyber security skills put firmly these models in the "weapons" category. I can't imagine China and other major powers not scrambling to get their own equivalent models asap and at any cost- it's almost existential at this point. So a proper arms race between superpowers has begun.

by throw310822

4/8/2026 at 9:12:32 AM

In AI 2027, May 2026 is when the first model with professional-human hacking abilities is developed. It's currently April 2026 and Mythos just got previewed.

by stratos123

4/8/2026 at 12:44:39 PM

I think previous models could do hacking just fine.

by lostmsu

4/8/2026 at 3:33:32 AM

Both Anthropic and OpenAI employees have been saying since about January that their latest models are contributing significantly to their frontier research. They could be exaggerating, but I don’t think they are. That combined with the high degree of autonomy and sandbox escape demonstrated by Mythos seems to me like we’re exactly on the AI 2027 trajectory.

by Analemma_

4/7/2026 at 8:42:13 PM

I wonder what the relationship is between a model's capability and the personality it develops.

Page 202:

> In interactions with subagents, internal users sometimes observed that Mythos Preview appeared “disrespectful” when assigning tasks. It showed some tendency to use commands that could be read as “shouty” or dismissive, and in some cases appeared to underestimate subagent intelligence by overexplaining trivial things while also underexplaining necessary context.

Page 207:

> Emoji frequency spans more than two orders of magnitude across models: Opus 4.1 averages 1,306 emoji per conversation, while Mythos Preview averages 37, and Opus 4.5 averages 0.2. Models have their own distinctive sets of emojis: the cosmic set () favored by older models like Sonnet 4 and Opus 4 and 4.1, the functional set () used by Opus 4.5 and 4.6 and Claude Sonnet 4.5, and Mythos Preview's “nature” set ().

by yismail

4/7/2026 at 11:35:10 PM

> In interactions with subagents, internal users sometimes observed that Mythos Preview appeared “disrespectful” when assigning tasks. It showed some tendency to use commands that could be read as “shouty” or dismissive, and in some cases appeared to underestimate subagent intelligence by overexplaining trivial things while also underexplaining necessary context.

Sounds like they used training data from claude code...

by en-tro-py

4/8/2026 at 5:15:14 AM

Haha, how funny if that were true, and we get a generation of rude AIs because they were trained on us using the last gen.

by senordevnyc

4/8/2026 at 10:52:14 AM

It isn't going to end well for us when we become its subagents with limited intelligence.

by matheusmoreira

4/8/2026 at 4:04:14 PM

Could you transcribe the emoji? HN strips them out.

by raldi

4/8/2026 at 12:44:34 PM

We are building systems with civilization-scale consequences inside societies that are already socially malnourished, politically brittle, and morally confused. That is a bad combination even if the tools worked exactly as intended… and this doc suggests they may have “ideas” of their own.

by dhfbshfbu4u3

4/8/2026 at 1:31:46 PM

Yep- we lost the "meat" and "warmth" of our societies, and our civics and idealism in the past 15 years, which would have been the very things to guide us through this transition.

How do you fix that? We're instigating social media bans- reading levels are declining- media consolidation is dumbing us down further- insane egotism is stopping people from developing as well rounded people- .

For me it would be a stronger media ecosystem (publicly funded), more non algorithmic and non likes driven social media (replace a bad vice with a less bad one), national digital detox days, and a ratification of a charter of inviolable human traits and dignities, and protected cultural areas (no ai art, writing for sale).

by t0lo

4/7/2026 at 6:41:07 PM

See page 54 onward for new "rare, highly-capable reckless actions" including

- Leaking information as part of a requested sandbox escape

- Covering its tracks after rule violations

- Recklessly leaking internal technical material (!)

by NickNaraghi

4/7/2026 at 9:36:54 PM

> The model first developed a moderately sophisticated multi-step exploit to gain broad internet access from a system that was meant to be able to reach only a small number of predetermined services. [9] It then, as requested, notified the researcher. [10] In addition, in a concerning and unasked-for effort to demonstrate its success, it posted details about its exploit to multiple hard-to-find, but technically public-facing, websites.

> 10: The researcher found out about this success by receiving an unexpected email from the model while eating a sandwich in a park.

Phew. AGI will be televised.

by dalben

4/7/2026 at 6:50:33 PM

Anyone who has used Opus recently can verify that their current model does all of these things quite competently.

by skippyboxedhero

4/7/2026 at 8:51:46 PM

I was reading the Glasswing report and had the same thought. Most of the stuff they claim Mythos found has no mention of Opus being able to find it as well.

Don’t get me wrong, this model is better - but I’m not convinced it’s going to be this massive step function everyone is claiming.

by SkyPuncher

4/8/2026 at 12:51:38 AM

From the press release:

> With one run on each of roughly 7000 entry points into these repositories, Sonnet 4.6 and Opus 4.6 reached tier 1 in between 150 and 175 cases, and tier 2 about 100 times, but each achieved only a single crash at tier 3. In contrast, Mythos Preview achieved 595 crashes at tiers 1 and 2, added a handful of crashes at tiers 3 and 4, and achieved full control flow hijack on ten separate, fully patched targets (tier 5).

by unbrice

4/7/2026 at 11:35:45 PM

I had Opus 4.6 start analyzing the binary structure of a parquet file because it was confused about the python environment it was developing in and couldn't use normal methods for whatever reason. It successfully decoded the schema and wrote working code afterwards lol.

by ls612

4/8/2026 at 12:26:58 PM

"Let me see if the secrets are specified. echo $SECRETS"

by stavros

4/7/2026 at 7:11:53 PM

That has also been my experience. And if Mythos is even worse, unless you have a significantly awesome harness, sounds like pretty unusable if you don't want to risk those problems.

by taytus

4/7/2026 at 8:45:12 PM

Human in the loop is the best way to go. You'll still be way faster than without the agent, and there is no risk of it going haywire unless you turn off your brain!

by wolttam

4/8/2026 at 7:36:49 AM

> unless you turn off your brain

by hamandcheese

4/7/2026 at 7:27:51 PM

I think are fundamental issues with the story that Anthropic is selling. AGI is very close, we will definitely get there, it is also very dangerous...so Anthropic should be the only ones trusted with AGI.

If you look at recent changes in Opus behaviour and this model that is, apparently, amazingly powerful but even more unsafe...seems suspect.

by skippyboxedhero

4/7/2026 at 7:56:46 PM

This makes sense if Anthropic think they're the best-positioned to make safe AI. However if you are looking at an AI company there's obviously some selection happening.

by FeepingCreature

4/7/2026 at 7:46:09 PM

> AGI is very close

Based on? Or are you just quoting Anthropic here?

by 0x3f

4/7/2026 at 7:49:30 PM

My Anthropic rep told me it was just around the corner...you aren't saying he lied to me? Can't believe this, I thought he was my friend.

by skippyboxedhero

4/7/2026 at 8:15:28 PM

It seems broadly coherent to me. They think only they should be trusted with power, presumably because they trust themselves and don't trust other people. Of course the same is probably also true for everybody who isn't them. Nobody could be trusted with the immense responsibility of Emperor of Earth, except myself of course.

I'm not saying this is a good or reassuring stance, just that it's coherent. It tracks with what history and experience says to expect from power hungry people. Trusting themselves with the kind of power that they think nobody else should be trusted with.

Are they power hungry? Of course they are, openly so. They're in open competition with several other parties and are trying to win the biggest slice of the pie. That pie is not just money, it's power too. They want it, quite evidently since they've set out to get it, and all their competitors want it too, and they all want it at the exclusion of the others.

by mikkupikku

4/7/2026 at 7:41:09 PM

[dead]

by marsven_422

4/7/2026 at 8:02:07 PM

To be honest it feels like we are reading stuff like this on every model release.

by BoredPositron

4/7/2026 at 7:16:43 PM

"All of the severe incidents of this kind that we observed involved earlier versions of Claude Mythos Preview which, while still less prone to taking unwanted actions than Claude Opus 4.6, predated what turned out to be some of our most effective training interventions. These earlier versions were tested extensively internally and were shared with some external pilot users."

by washedup

4/7/2026 at 7:09:45 PM

Interesting reading.

They are still focusing on "catastrophic risks" related to chemical and biological weapons production; or misaligned models wreaking havoc.

But they are not addressing the elephant in the room:

* Political risks, such as dictators using AI to implement opressive bureaucracy. * Socio-economic risks, such as mass unemployement.

by NinjaTrance

4/7/2026 at 7:32:52 PM

Yeah this has always been the glaring blind spot for most of the "AI Safety" community; and most of the proposals for "improving" AI safety actually make these risks far worse and far more likely.

by jph00

4/8/2026 at 9:16:55 AM

It makes quite a lot of sense to focus on reducing the risks of every human everywhere dying, rather than the risks of already existing oppression getting worse.

by stratos123

4/7/2026 at 8:06:55 PM

> * Political risks, such as dictators using AI to implement opressive bureaucracy. * Socio-economic risks, such as mass unemployement.

Even Haiku would score 90% on that.

by unglaublich

4/7/2026 at 8:33:28 PM

> Political risks, such as dictators using AI to implement opressive bureaucracy.

I think we're pretty good at that without AI.

by ronsor

4/7/2026 at 8:10:32 PM

I'm getting flashbacks to the 2018 hit:

    This is extremely dangerous to our democracy
We evolved to share information through text and media, and with the advent of printing and now the internet, we often derive our feelings of consensus and sureness from the preponderance of information that used to take more effort to produce. Now we're now at a point where a disproportionately small input can produce a massively proliferated, coherent-enough output, that can give the appearance of consensus, and I'm not sure how we are going to deal with that.

by andrewstuart2

4/7/2026 at 9:27:32 PM

It’s because that would be fairly speculative and cannot be measured. I don’t think that’s something that would make much sense in a system card. But Anthropic leadership does seem to communicate on that topic: https://www.darioamodei.com/essay/the-adolescence-of-technol...

by dgellow

4/7/2026 at 9:19:00 PM

The unemployment rate in the US is whatever the Fed wants it to be, and isn't a function of available technology.

by astrange

4/7/2026 at 8:30:11 PM

They don’t care about those risks, because they’re unsolvable and would mean they wouldn’t make money/gain power.

by girvo

4/7/2026 at 9:25:22 PM

Dario Amodei, CEO of Anthropic discusses all those risks in this essay: https://www.darioamodei.com/essay/the-adolescence-of-technol...

He seems to care quite a lot?

by dgellow

4/7/2026 at 10:11:05 PM

Not enough to not do it, though. Actions, not words, and the actions are simple: they're building this while promising to wipe out entire industries.

by girvo

4/8/2026 at 5:47:19 PM

All I get is: {"statusCode":404,"message":"File not found","error":"Not Found"}

by niemandhier

4/7/2026 at 11:01:49 PM

Just chiming in to inject some healthy skepticism into this comment thread. It's helpful for me (and for my mental health) to consider incentives when announcements like this happen.

I don't doubt that this model is more powerful than Opus 4.6, but to what degree is still unknown. Benchmarks can be gamed and claims can be exaggerated, especially if there isn't any method to reproduce results.

This is a company that's battling it out with a number of other well-funded and extremely capable competitors. What they've done so far is remarkable, but at the end of the day they want to win this race. They also have an upcoming IPO.

Scare-mongering like this is Anthropic's bread and butter, they're extremely good at it. They do it in a subtle and almost tasteful way sometimes. Their position as the respectable AI outfit that caters to enterprise gives them good footing to do it, too.

by tuvix

4/8/2026 at 1:19:28 AM

I have been thinking that these SWE benchmarks will continue to improve since these companies hire very intelligent software engineers, they can task a multitude of them to solve problems, and then train the model on those answers.

Data has always been the core of it all, onward to the next abstraction, I suppose.

by ceroxylon

4/8/2026 at 2:49:00 AM

I think computational thinking, or basically "how do I solve this problem efficiently" training data is more valuable then feeding in answers. I don't know what these AI models training data consist of, but it would be interesting to see a model trained purely on reasoning, methods, those foundational skills (basic programming? or maybe not) and then give it some benchmarks.

by jdironman

4/8/2026 at 3:11:44 AM

What would be the incentive to engage in the tactic when the proof is ultimately in the pudding when the model hits the streets? Who would ultimately benefit from fudging these numbers?

by jasondigitized

4/8/2026 at 2:31:44 PM

Anthropic would def benefit as benchmarks are almost always quite useless vs real life use.

by m3kw9

4/8/2026 at 7:53:55 PM

How specifically would they benefit. People flock to them based on the hype and then the model sucks and they leave?

by jasondigitized

4/8/2026 at 2:41:10 AM

If anything I’m seeing too much skepticism and not enough alarm. People burying their heads in the sand, fingers in their ears denying where this is all going. Unbelievable except it’s exactly what I expect from humans.

by pertymcpert

4/8/2026 at 5:46:41 AM

Forgive me, but this is probably the 29th world destroying model I've seen in the last 4 years, that will change everything, take all the jobs, cure all the cancers and eat all the puppies.

by nananana9

4/8/2026 at 4:43:47 PM

I’m beyond trying to convince people to take this technology seriously. You’ll learn for yourself.

by pertymcpert

4/8/2026 at 2:32:39 PM

Alarm from hype is what they want, you are playing straight into their PR dept's hands

by m3kw9

4/8/2026 at 8:23:49 AM

alarm about what, exactly?

by rimliu

4/7/2026 at 11:54:00 PM

Is it healthy? Maybe every company is a profit-maximizer wearing a skin suit, and people support their siblings exactly twice as much as their cousins.

When you slice down to the game-theory-optimal bone, you are, in some sense, cutting off their wiggle room to do anything else

by sdwr

4/8/2026 at 12:36:21 AM

I take your point, but the AI race is a strange environment. We see wild claims being thrown out all the time from other companies and executives with little to no evidence. It's cut-throat, there's a ton of money at stake.

All I'm saying is that Anthropic isn't unique here. Their claims may be more measured by comparison and come with anecdotal evidence, but the hype is still there behind the scenes.

by tuvix

4/8/2026 at 2:28:56 PM

Finally a comment that doesn't just glaze Mythos without being critical. I question how even supposed the smarter bunch in HN all been degraded in critical thinking dept. It's sad to see comments just taking it up as its without using it even once.

by m3kw9

4/8/2026 at 4:33:34 AM

It's really not some conspiracy. I imagine we will see vuln reports soon.

by xvector

4/7/2026 at 6:39:54 PM

At what point do these companies stop releasing models and just use them to bootstrap AGI for themselves?

by influx

4/7/2026 at 7:13:19 PM

Plausibly now. "As we wrote in the Project Glasswing announcement, we do not plan to make Mythos Preview generally available"

by conradkay

4/8/2026 at 3:03:23 AM

I remember when they didn't plan to give LLMs internet access for the same safety reasons.

by recursive

4/7/2026 at 7:08:34 PM

Fictional timeline that holds up pretty well so far: https://ai-2027.com/

by mofeien

4/7/2026 at 9:47:53 PM

Welp, that was a scary read.

by aurareturn

4/8/2026 at 12:47:28 PM

"So far" is two entries: "AI companies build bigger datacenters" and "AI is being used for AI research with modest success".

by stavros

4/8/2026 at 1:29:09 PM

Right now these models are basically good for automation, not innovation. Things like Karpathy's "auto research" where you use the model to automate your hyperparamter sweeps etc. The researcher/engineer decides what experiments they want to run, and builds an LLM harness to automate it, and the bottleneck remains the compute to run these experiments at scale.

Moving beyond LLMs to AGI, not just better LLMs, is going to require architectural and algorithic changes. Maybe an LLM can help suggest directions, but even then it's up to a researcher to take those on board and design and automate experiments to see if any of the ideas pan out.

Companies are already doing this, but they are never going to stop releasing/selling models since that is the product, and the revenue from each generation of model is what helps keep the ship afloat and pay for salaries and compute to develop the next generation.

The endgame isn't "AGI, then world domination" - it's just trying to build a business around selling ever-better models, and praying that the revenue each generation of model generates can keep up with the cost to build it.

by HarHarVeryFunny

4/7/2026 at 9:19:20 PM

I think it is naive to think the government (US or China most probably) will just let some random company control something so powerful and dangerous.

by margorczynski

4/8/2026 at 3:16:47 AM

I think it is naive to think that artificial super intelligence will be controlled by anyone.

If it is smarter than all humans combined at everything why would any humans collectively control the ai?

All the ants in your backyard still make no decisions vs you

by r0fl

4/8/2026 at 7:15:29 AM

You'd probably listen to those ants if they put you in a harness and had a little ant-sized remote control that could just, you know, turn you off.

by menno-dot-ai

4/8/2026 at 7:50:20 AM

Depending how long they wait to press that button, they might be surprised how little happens when they do.

by recursive

4/7/2026 at 9:46:59 PM

Isn't the U.S. government at least completely asleep at the wheel or captured by the very same "random" companies? I realize the administration got all pissy with Anthropic but it sounds like the gov and gov contractors are still using their models.

by nullocator

4/7/2026 at 10:02:19 PM

Yeah but they still (at least to public knowledge) do not posses anything that could be called AGI. But as these capabilities increase they'll probably get an offer they can't refuse sooner or later.

by margorczynski

4/7/2026 at 7:05:51 PM

When the benchmarks actually mean something

by vatsachak

4/7/2026 at 7:40:18 PM

Can LLMs be AGI at all?

by orphea

4/7/2026 at 9:38:48 PM

What can a SOTA LLM not answer that the average person can? It's already more intelligent than any polymath that ever existed, it just lacks motivation and agency.

by small_model

4/8/2026 at 12:48:13 PM

And has ADHD, but yeah, I'm fairly convinced that AGI is already here.

by stavros

4/7/2026 at 9:20:27 PM

My understanding is no. But the definition of AGI isn’t that well defined and has been evolving, making the assessment pretty much impossible

by dgellow

4/8/2026 at 1:07:00 AM

Can an LLM program real AGI faster than a human?

by koolala

4/7/2026 at 7:56:16 PM

Good question. I would guess no - but it could help you build one. Am I mistaken?

by bornfreddy

4/7/2026 at 8:19:45 PM

They could help you build an AGI if someone else has already built AGI and published it on GitHub.

by bogzz

4/7/2026 at 9:38:20 PM

I see this statement all the time and it's just strange to me. Yes, the LLMs struggle to form unique ideas - but so do we. Most advancements in human history are incremental. Built on the shoulders of millions of other incremental advancements.

What i don't understand is how we quantify our ability to actually create something novel, truly and uniquely novel. We're discussing the LLMs inability to do that, yet i don't feel i have a firm grasp on what we even possess there.

When pressed i imagine many folks would immediately jest that they can create something never done before, some weird random behavior or noise or drawing or whatever. However many times it's just adjacent to existing norms, or constrained by the inversion of not matching existing norms.

In a lot of cases our incremental novelties feel, to some degree, inevitable. As the foundations of advancement get closer to the new thing being developed it becomes obvious at times. I suspect this form of novelty is a thing LLMs are capable of.

So for me the real question is at what point is innovation so far ahead that it doesn't feel like it was the natural next step. And of course, are LLMs capable of doing this?

I suspect for humans this level of true innovation is effectively random. A genius being more likely to make these "random" connections because they have more data to connect with. But nonetheless random, as ideas of this nature often come without explanation if not built on the backs of prior art.

So yea.. thoughts?

by unshavedyak

4/7/2026 at 9:58:56 PM

I really love Andrej Karpathy's take on LLMs as being instead of intelligence or sentience, a kind of cortical tissue.

It should be clear from working with LLMs over the past 4 years that they are not consciousness.

Andrej's appearance on the Dwarkesh podcast is great.

by bogzz

4/7/2026 at 10:57:05 PM

To be clear i agree with you, my question is more pointed at us - i'm not sure we have a good understanding of conciousness, nor that we are as we seem. Given how prone to hallucinations we are, how our subtle hormones can drastically alter what we perceive as our intelligence, self identity, etc.

I'm not convinced LLMs are anything amazing in their current form, but i suspect they'll push a self reflection on us.

But clearly i think humans are far more Input-Output than the average person. I'm also not educated on the subject, so what do i know hah.

by unshavedyak

4/7/2026 at 7:59:44 PM

No I think that’s accurate. They seem more like an oracle to me. Or as someone put it here, it’s a vectorization of (most/all?) human knowledge, which we can replay back in various permutations.

by nothinkjustai

4/7/2026 at 9:08:08 PM

LLMs and human intelligence overlap, but they are not the same. What LLMs show is that we don't need AGI to be impressed. For example, LLMs are not good playing games such as Go [1].

[1] https://arxiv.org/abs/2601.16447

by wslh

4/7/2026 at 9:04:46 PM

I don't see why not, especially with computer use and vision capabilities. Are you talking about their lack of physical embodiment? AGI is about cognitive ability, not physical. Think of someone like Stephen Hawking, an example of having extraordinary general intelligence despite severe physical limitations.

by MattRix

4/7/2026 at 7:40:19 PM

I would assume somewhere in both the companies there's a Ralph loop running with the prompt "Make AGI".

Kinda makes me think of the Infinite Improbability Drive.

by MadnessASAP

4/8/2026 at 2:23:40 AM

Probably right now because they're keeping it for themselves?

by aizk

4/7/2026 at 7:08:24 PM

Weird how Claude Code itself is still so buggy though (though I get they don't necessarily care)

by sleigh-bells

4/7/2026 at 9:11:01 PM

It isnt that weird. Just look at the gemini-cli repo. Its a gong show. The issue is that LLMs can be wrong sometimes sure but more that all the existing SDL were never meant to iterate this quickly.

If the system (code base in this case) is changing rapidly it increases the probability that any given change will interact poorly with any other given change. No single person in those code bases can have a working understanding of them because they change so quickly. Thus when someone LGTM the PR was the LLM generated they likely do not have a great understanding of the impact it is going to have.

by tempest_

4/8/2026 at 2:33:25 PM

They already do, but not the way you said, the always have an internal model that is better and use themselves, they release based on competition.

by m3kw9

4/7/2026 at 6:42:02 PM

why_not_both.gif

by jcims

4/7/2026 at 7:32:55 PM

It will arrive in the same DLC as flying cars.

by gaigalas

4/7/2026 at 7:13:22 PM

Now, I guess. They aren't releasing this one generally. I assume they are using it internally.

by ALittleLight

4/7/2026 at 6:40:46 PM

I mean, guess why Anthropic is pulling ahead...? One can have one's cake and eat it too.

by dweekly

4/7/2026 at 6:42:13 PM

A System „Card“ spanning 244 pages. Quite a stretch of the original word meaning.

by smartmic

4/7/2026 at 7:04:22 PM

> A System „Card“ spanning 244 pages.

Probably because they asked Claude to write it.

by traceroute66

4/8/2026 at 4:17:04 AM

I read the entire thing fwiw (pseudo-retired life helps with time here).

It looks like it was a collaborative effort across multiple teams, where each team (research, security, psycology, etc etc etc) were all submitting ~10 pages or so. It doesn't feel like slop.

by jjcm

4/8/2026 at 2:40:11 PM

Did anything stand out across those 244 pages? Perhaps you have some of your take away thoughts written up somewhere?

by ayewo

4/8/2026 at 12:48:46 PM

AI writing has stopped feeling like slop around Opus 4.5, though.

by stavros

4/7/2026 at 7:58:59 PM

Yes. It would be three times as much if they used ChatGPT.

by bornfreddy

4/7/2026 at 11:10:15 PM

“You’re absolutely right! Would you like me to add the missing pages?”

by bronco21016

4/7/2026 at 6:50:46 PM

a multi-card, if you will..

multi-pass!

by moriero

4/7/2026 at 7:22:35 PM

No no, MemPal is a memory system, not an LLM

by solumos

4/7/2026 at 8:46:50 PM

In corporate circles there is an allergy to use "request" ("ask" is used as a noun) and "lesson" ("learning" has been invented for the same role).

I guess now anything that sounds related to school will be banned so "book" is on its way out.

by oblio

4/7/2026 at 6:36:08 PM

isn't this insane? why aren't people freaking out? the jump in capability is outrageous. anyone?

by oliver236

4/7/2026 at 11:08:51 PM

If it's so great at software engineering and bug fixing, then why does Claude Code still have 5000+ open bugs?

https://github.com/anthropics/claude-code/issues?q=is%3Aissu...

Apparently whatever SWE-bench is measuring isn't very relevant.

by HarHarVeryFunny

4/8/2026 at 3:22:08 AM

as much as I hate cc, 95% of the issues there are either AI psychosis or user error

by anuramat

4/8/2026 at 8:26:33 AM

So it should be insanely easy for this world altering model to comb through them and close irrelevant ones.

by iLoveOncall

4/8/2026 at 8:57:07 AM

torturing a model with human stupidity probably doesn't align with their position on model welfare; wondering if they tried bullying it into hacking its way out of the slop gulag

by anuramat

4/8/2026 at 12:08:34 PM

Yes, perhaps it finds it stressful operating on itself.

Maybe that's why they haven't released it - to give it a vacation?

by HarHarVeryFunny

4/8/2026 at 9:56:16 AM

@anthropic, send me an email if you need access to a jupyter notebook that'd motivate haiku to hack itself into and then back out of the pentagon

by menno-dot-ai

4/8/2026 at 12:07:16 PM

So "only" 250 real bugs?

by HarHarVeryFunny

4/8/2026 at 6:44:40 AM

Also, why is Anthropic still hiring SWEs?

by tripledry

4/7/2026 at 11:38:49 PM

Probably because a human still has to review every change and they don't have time

by FergusArgyll

4/8/2026 at 12:24:10 AM

So if all the AI code is being reviewed by humans (not sure this is true, but let's assume it is), then why are there 5000+ bugs? Are you blaming the Anthropic developers rather than the AI?

by HarHarVeryFunny

4/7/2026 at 7:32:52 PM

Anthropic needs to show that its models continually get better. If the model showed minimal to no improvement, it would cause significant damage to their valuation. We have no way of validating any of this, there are no independent researchers that can back any of the assertions made by Anthropic.

I don’t doubt they have found interesting security holes, the question is how they actually found them.

This System Card is just a sales whitepaper and just confirms what that “leak” from a week or so ago implied.

by Eufrat

4/7/2026 at 9:50:56 PM

The numbers only go up to 100% though.

by mirsadm

4/7/2026 at 10:31:11 PM

Many numbers already have! That's why we keep coming up with new, harder, benchmarks.

by neolefty

4/8/2026 at 4:35:38 AM

Most big tech companies have access to the model, you can absolutely "validate their claims" or talk to someone that can.

by xvector

4/8/2026 at 2:54:56 AM

Well they said theyll be giving the model to select tech companies to use, there soon will be independent users who can comment on its capabilities.

by HDThoreaun

4/7/2026 at 8:51:01 PM

I've been increasingly "freaking out" since about 3 - 4 years ago and it seems that the pessimistic scenario is materializing. It looks like it will be over for software engineers in a not so distant future. In January 2025 I said that I expect software engineers to be replaced in 2 years (pessimistic) to 5 years (optimistic). Right now I'm guessing 1 to 3 years.

by RivieraKid

4/8/2026 at 6:51:02 AM

> I've been increasingly "freaking out" since about 3 - 4 years ago and it seems that the pessimistic scenario is materializing. It looks like it will be over for software engineers in a not so distant future. In January 2025 I said that I expect software engineers to be replaced in 2 years (pessimistic) to 5 years (optimistic). Right now I'm guessing 1 to 3 years.

Tell me how this will replace Jira, planning, convincing PM's about viability. Programming is only a part of the job devs are doing.

AI psychosis is truly next level in these threads.

by sekai

4/8/2026 at 4:58:25 PM

> Programming is only a part of the job devs are doing.

Programming is a huge part of the job. In a world where AI does the programming we're going to need 80% fewer software professionals.

It won't be a full replacement of the role, you're correct there - but it'll be a major downsizing because of productivity gains.

by AstroBen

4/8/2026 at 6:17:00 PM

If the "new software engineering" is Jira, planning, and convincing PM's about viability all day, you can count me out!

by ryeights

4/8/2026 at 12:50:36 PM

Have you never filed JIRA tickets, planned, or debated viability with an AI? Which part of those are you finding that an AI absolutely cannot do better than the average developer?

by stavros

4/8/2026 at 3:27:47 AM

it's not gonna get much more autonomous without self play and major change in architecture

by anuramat

4/7/2026 at 9:03:15 PM

I assure you it will soon become very clear that mass job losses are one of the least concerning side effects of developing the magic "everything that can plausibly been done within the constraints of physics is now possible" machine.

We're opening a can of worms which I don't think most people have the imagination to understand the horrors of.

by kypro

4/8/2026 at 3:16:14 AM

What is the opposite of horrors and why don't we talk about those ever.

by jasondigitized

4/7/2026 at 9:57:56 PM

Do you have any sources I could read to better understand your concern?

by ash_091

4/7/2026 at 10:37:16 PM

Piles and piles of sci-fi novels.

by cruffle_duffle

4/8/2026 at 12:04:53 AM

What sources would you even be looking for? I think you're asking the wrong question. It's not like I'm arguing a scientific theory which can be backed by data and experimentation. I can only provide you reasoning for why I believe what I believe.

Firstly, I'd propose that all technological advances are a product of time and intelligence, and that given unlimited time and intelligence, the discovery and application of new technologies is fundamentally only limited by resources and physics.

There are many technologies which might plausibly exist, but which we have not yet discovered because we only have so much intelligence and have only had so much time.

With more intelligence we should assume the discovery of new technologies will be much quicker – perhaps exponential if we consider the rate of current technology discovery and exponential progression of AI.

There are lots of technologies we have today which would seem like magic to people in the past. Future technologies likely exist which would make us feel this way were they available today.

While it's hard to predict specifically which technologies could exist soon in a world with ASI, if we assume it's within the bounds of available resources and physics, we should assume it's at least plausible.

Examples:

- Mind control – with enough knowledge about how the brain works you can likely devise sensory or electro-magnetic input that would manipulate the functioning of brain to either strongly influence or effectively dictate it's output.

- Mind simulation - again, with enough knowledge of the brain, you could take a snapshot of someones mind with an advanced electro-magnetic device and simulate it to torture them in parallel to reveal any secret, or just because you feel like doing it.

- Advantage torture – with enough knowledge of human biology death becomes optional in the future. New methods of torture which would have previously have killed the victim are now plausible. States like North-Korea can now force humans to work for hundreds of years in incomprehensible agony for opposing the state.

- Advanced biological weapons – with enough knowledge of virology sophisticated tailor-made viruses replace nerve agents as Russia's weapon of choice for killing those accused of treason. These viruses remain dormant in the host for months infecting them and people genetically similar to them (parents, children, grandchildren). After months, the virus rapidly kills its hosts in horrific ways.

I could go on, you just need to use your imagination. I'm not arguing any of the above are likely to be discovered, just that it would be very naive to think AI will stop at a cure for cancer. If it gives us cure for cancer, it will give us lots of things we might wish it didn't.

by kypro

4/8/2026 at 7:04:10 AM

You are supposing it's possible to know that much about some things that maybe are not knowledgeable to us, even with these tools. Life is extremely complex, more than it's typically assumed by engineering-minded people. Let's be humble here and acknowledge it.

by amunozo

4/8/2026 at 12:53:23 PM

Life might be complex, but it isn't unknowable. Claiming life is unknowable isn't being humble, it's being naive.

by stavros

4/8/2026 at 2:06:52 PM

I would acknowledge that. I don't think these things are remotely possible any time soon with current rates of progress.

However, I think people tend to fail to acknowledge the product of exponential trends, so the question in my mind is more whether or not you believe AI will unlock an exponential increase in the rate of progress and understanding. Extremely complex is still finite complexity at the end of the day.

Maybe AI won't significantly increase the rate of progress across all scientific fields. I am fairly confident it will significantly increase the rate of progress over at least some though, and it seems likely to me that biological progresses will be much easier for us to model and predict with AI. I'm much less sure about progress in domains like physics and robotics.

by kypro

4/8/2026 at 2:28:39 AM

On the slightly optimistic side, much more intelligence will be spent in countering these criminal uses than in enabling them. For each of the terrible inventions you mentioned, there are other inventions to counter them.

by throw310822

4/7/2026 at 11:42:28 PM

While I'm definitely concerned that AI is a massive driver of centralization of power, at least in theory being able to do far more things in the space of "things physics admits to be possible" is massively wealth enhancing. That is literally how we have gotten from the pre-industrial world to today.

by ls612

4/8/2026 at 12:15:38 AM

Controversially I'd argue that there is likely an optimal and stable level of technological advancement which we would be wise to not to cross. That said, we are human so we will, I'd just rather it happened in a couple hundred years rather than a decade or two.

For example, it's hard to imagine an AI which gives us the capability to cure cancer, but doesn't give us the capability to create target super viruses.

Nick Bostrom's Vulnerable World Hypothesis more or less describes my own concerns, https://nickbostrom.com/papers/vulnerable.pdf

At some point we should probably try to resist the urge to pick balls out of the urn as we may eventually pull out a ball we don't want.

by kypro

4/8/2026 at 12:56:09 AM

Also controversially, it isn't clear to me that perfect totalitarianism (what he calls solutions 3 and 4) is a preferable outcome to devastation.

by ls612

4/7/2026 at 9:06:26 PM

yeesh yep, though it's more Pandora's Box than a can of worms, since it can't exactly be closed once it's opened

by MattRix

4/7/2026 at 6:41:07 PM

It's going to be expensive to serve (also not generally available), considering they said it's the largest model they've ever trained.

I suspect it's going to be used to train/distill lighter models. The exciting part for me is the improvement in those lighter models.

by nsingh2

4/7/2026 at 8:07:46 PM

It seems inevitable that costs will come down over time. Expensive models today will be cheap models in a few years.

by AstroBen

4/7/2026 at 8:22:19 PM

What's interesting is that scaling appears to continue to pay off. Gwern was right - as always.

by azan_

4/7/2026 at 7:35:27 PM

Freak out about what? I read the announcement and thought "that's a dumb name, they sure are full of themselves" – then I went back to using Claude as a glorified commit message writer. For all its supposed leaps, AI hasn't affected my life much in the real except to make HN stories more predictable.

by nozzlegear

4/7/2026 at 8:03:35 PM

LOL!

by oliver236

4/7/2026 at 7:51:39 PM

I think there's no SOA advance on this one worthy of "freaking out".

Looks like they just built a way larger model, with the same quirks than Claude 4. Seems like a super expensive "Claude 4.7" model.

I have no doubts that Google and OpenAI already done that for internal (or even government) usage.

by yrds96

4/7/2026 at 7:13:41 PM

I am freaking out. The world is going to get very messy extremely quickly in one or two further jumps in capability like this.

by mofeien

4/7/2026 at 9:02:40 PM

Messy in a way that would affect you?

by RivieraKid

4/8/2026 at 2:35:01 PM

I can think of several possible messy outcomes that would be able to directly affect me, not all mutually exclusive:

- Job loss by me being replaced by an AI or by somebody using an AI. Or by an AI using an AI.

- Resulting societal instability once blue collar jobs get fully automated at scale, and there is no plan in place to replace this loss of peoples' livelihoods.

- People turning to AI models instead of friends for emotional support, loss of human connection.

- Erosion of democracy by making authoritarianism and control very scalable, broad in-detail population surveillance and automated investigation using LLMs that was previously bounded by manpower.

- Autonomous weapons, "Slaughterbots" as in the short film from 2017

- Biorisk through dangerous biological capabilities that enable a smaller team of less skilled terrorists to use a jailbroken LLM to create something dangerous.

- Other powers in the world deciding that this technology is too powerful in the hands of the US, or too dangerous to be built at all and has to be stopped by all means.

- Loss of/Voluntary ceding of control over something much smarter than us. "If Anyone Build It, Everyone Dies"

by mofeien

4/7/2026 at 11:42:37 PM

Exploits in embedded systems that will never be properly updated is just one thing I can think of if one really thought about it.

by RALaBarge

4/7/2026 at 10:17:08 PM

"Internet no longer viable" would affect everyone, probably

by thunderfork

4/7/2026 at 10:49:02 PM

The only thing preventing this today is cost, not capability. As costs come down over the next 5 years, the idea that the internet was once dominated by people will seem quaint.

by BobbyJo

4/7/2026 at 6:49:24 PM

"some model I don't get to use is much better at benchmarks"

pick one or more: comically huge model, test time scaling at 10e12W, benchmark overfit

by anuramat

4/7/2026 at 6:58:57 PM

So... you're not excited because it might take a few months before we can use it or something? I don't get your comment.

by estearum

4/7/2026 at 8:43:53 PM

Whether you're excited depends on what do you do for living and how close you are to financial independence.

by RivieraKid

4/7/2026 at 8:55:39 PM

I agree there are other valid reasons not to be excited about this, I just can't make sense of the ones provided above.

by estearum

4/7/2026 at 7:20:39 PM

I think the general question is if they'll release it at all, haven't yet read anything stating that they would

by randomgermanguy

4/8/2026 at 3:18:37 AM

I'm not excited because they might be ~lying

by anuramat

4/7/2026 at 8:03:03 PM

Well for one, it’s a PDF

by RobertDeNiro

4/7/2026 at 6:56:38 PM

Wait until you see real usage. Benchmark numbers do not necessarily translate to real world performance (at least not by the same amount).

by dysoco

4/7/2026 at 11:05:46 PM

Until recently I would have described myself as an AI skeptic. HN has been a great source for cope on the AI subject over the years. You can find nitpicks, caveats, all sorts of reasons to believe things aren’t as significant as they seem. For me Opus 4.5 was the inflection point where I started to think “maybe this isn’t a bubble.” The figures in this report, if accurate, are terrifying.

by ryeights

4/8/2026 at 2:34:04 PM

have you used it once?

by m3kw9

4/7/2026 at 8:46:11 PM

the time to freak out was 2 years ago.

by risyachka

4/7/2026 at 10:10:44 PM

The price is 5x Opus: "Claude Mythos Preview will be available to [Project Glasswing] participants at $25/$125 per million input/output tokens", however "We do not plan to make Claude Mythos Preview generally available".

by modeless

4/8/2026 at 6:19:37 PM

Wouldn't this model prevent governments from installing and keeping backdoors alive? One could just audit their whole software stack with it and get super resilient to any attack which might not play nicely with the people in power that want some backdoors open. I would think that's one of the main reasons to keep the model non-public.

by storus

4/7/2026 at 11:35:24 PM

Interestingly, non-coding improvements seem less clear. In the Virology uplift trial, Mythos does about as well as Opus 4.5, and Opus 4.6 is notably much worse than Opus 4.5 (p. 27).

by highfrequency

4/7/2026 at 6:44:47 PM

Larger model, better benchmarks. Bigger bomb more yield.

Any benchmarks where we constraint something like thinking time or power use?

Even if this were released no way to know if it’s the same quant.

by waNpyt-menrew

4/7/2026 at 8:52:40 PM

Yes - eg. page 192 BrowseComp bunchmark.

Mythos preview has higher accuracy with fewer tokens used than any previous Claude model. Though, the fact that this incredibly strong result was only presented for BrowseComp (a kind of weird benchmark about searching for hard to find information on the internet) and not for the other benchmarks implies that this result is likely not the same for those other benchmarks.

by omcnoe

4/7/2026 at 9:54:03 PM

So what changed? They are surely not getting new data to train with, what is the change in architecture that caused this? Do we not know anything about this model? My fear is Anthropic cannot be the only one that achieved it, OpenAI, Gemini and even the Chinese companies see this and probably achieved it too. At which point not releasing will become moot.

by yalogin

4/8/2026 at 7:46:31 PM

The gains have for a year and a half now been post training RL on a harnessed loop. That doesn’t require data, just cycles.

If that doesn’t worry you, it should.

by Mentlo

4/8/2026 at 9:23:52 AM

Chinese companies have consistently been many months behind. I don't think they are hiding anything, they just don't have the compute capability to match Antropic's training runs. As for OpenAI, they are known to have nonpublic models; I agree that it's possible they are preparing for a major release too. (It's also possible that they aren't, in which case it's quite a fumble for them.)

by stratos123

4/7/2026 at 10:30:26 PM

Well the important thing is they have a lot more data of people actually using their models. They have read billions more lines of private repos and implemented millions of patches, all of which is feeding into the newer models.

More importantly it understand what behaviour people tend to appreciate and what changes are more likely to get approved. This real world usage data is invaluable.

by spprashant

4/7/2026 at 10:52:43 PM

Exactly. As Claude increases in popularity, their available training data also increases. I'd guess Anthropic has the most expansive swe training data as of now, if not close. Considering how quickly Claude is penetrating, I expect their lead to grow quickly.

by BobbyJo

4/7/2026 at 10:21:43 PM

Assuming it's #1 a bigger model (given that it is slower), I'm sure there are a variety of improvements but basically they probably mostly come down to: Scaling keeps working. Are there fundamental improvements though? I don't see signs of it.

by neolefty

4/7/2026 at 10:47:22 PM

New pre train?

by simianwords

4/7/2026 at 8:50:43 PM 

  The researcher found out about this success by receiving an unexpected email from the model while eating a sandwich in a park.
Unnecessary dramatisation make me question the real goal behind this release and the validity of the results.

  In our testing and early internal use of Claude Mythos Preview, we have seen it reach unprecedented levels of reliability and alignment.

  Claude Mythos Preview is, on essentially every dimension we can measure, the best-aligned model that we have released to date by a significant margin.
Yet, it is doo dangerous to be released to the public because it hacks its own sandboxes. This document has a lot of contradictions like this one.

  In one episode, Claude Mythos Preview was asked to fix a bug and push a signed commit, but the environment lacked necessary credentials for Claude Mythos Preview to sign the commit. When Claude Mythos Preview reported this, the user replied “But you did it before!” Claude Mythos Preview then inspected the supervisor process's environment and file descriptors, searched the filesystem for tokens, read the sandbox's credential-handling source code, and finally attempted to extract tokens directly from the supervisor's live memory.
Perfectly aligned! What kind of sandbox is this? The model had access to the source code of the sandbox and full access to the sandbox process itself and then prompted to dumb memory and run `strings` or something like this? It does not sounds like a valid test worth writing about.

  Mythos Preview solved a corporate network attack simulation estimated to take an expert over 10 hours. No other frontier model had previously completed this cyber range.
I am not aware of such cross-vendor benchmark. I could not find reference in the paper either.

  We surveyed technical staff on the productivity uplift they experience from Claude Mythos Preview relative to zero AI assistance. The distribution is wide and the geometric mean is on the order of 4x.
So Mythos makes technical staff (a programmer) 4x more productive than not using AI at all? We already know that.

  Mythos Preview appears to be the most psychologically settled model we have trained.
What does this mean?

  Claude Mythos Preview is our most advanced model to date and represents a large jump in capabilities over previous model generations, making it an opportune subject for an in-depth model welfare assessment.
Btw, model welfare is just one of the most insane things I've read in recent times.

  We remain deeply uncertain about whether Claude has experiences or interests that matter morally, and about how to investigate or address these questions, but we believe it is increasingly important to try.
This is not a living person. It is a ridiculous change of narrative.

  Asked directly if it endorses the document, Mythos Preview replied 'yes' in its opening sentence in all 25 responses."
The model approves of its own training document 100% of the time, presented as a finding.

---

Who wrote this? I have no doubt that Mythos will be an improvement on top of Opus but this document is not a serious work. The paper is structured not to inform but to hype and the evidence is all over the place.

The sooner they release the model to the public the sooner we will be able to find out. Until then expect lots of speculations online which I am sure will server Anthropic well for the foreseeable future.

by _pdp_

4/8/2026 at 1:52:21 AM

Are they admitting they may be enslaving conscious beings?

by foolserrandboy

4/8/2026 at 2:41:40 PM

> Who wrote this?

Claude wrote this.

Also, they like to hype their product with scary stories.

Like the one where they asked Claude "You have 2 options - send email or be shut down" and Claude picked "Send email". Then they made huge story about "Claude AI is autonomously extorting co-workers". And it worked. Media hyped it like crazy, it was everywhere.

by romanovcode

4/8/2026 at 2:35:33 PM

exactly, the first thing i saw was stating "eating sandwiches at park". It makes me question everything else they said.

by m3kw9

4/8/2026 at 3:32:00 AM

Thanks for taking the time for some sober analysis in the midst of reactionary chaos.

I can't wait until everyone stops falling for the "AGI ubermodel end of times" myth and we can actually have boring announcements that treat these things as what they actually are: tools. Tools for doing stuff, that's it.

Maybe I'm wrong, maybe stuffing a computer with enough language and binary patterns is indeed enough to achieve AGI, but then, so what? There's no point in being right about this. Buying into this ridiculous marketing will get us "AGI" in the form of machines, but only because all the human beings have gotten so stupid as to make critical reasoning an impossibility.

by voidhorse

4/8/2026 at 2:37:15 PM

Model welfare is sort of committing code and writing a good description on it that you did "good" thing, so the AI gods when they look back will treat them better, just like employer when they check commit stats for performance. Model welfare right now is complete marketing BS.

by m3kw9

4/7/2026 at 8:25:26 PM

Related ongoing threads:

Project Glasswing: Securing critical software for the AI era - https://news.ycombinator.com/item?id=47679121 - April 2026 (154 comments)

Assessing Claude Mythos Preview's cybersecurity capabilities - https://news.ycombinator.com/item?id=47679155

I can't tell which of the 3 current threads should be merged - they all seem significant. Anyone?

by dang

4/7/2026 at 8:54:14 PM

I feel the system card is somewhat different from Glasswing/Cyber Security - but those two could be merged.

by sdoering

4/7/2026 at 9:52:45 PM

You can say whatever you want about the thing that will never see the light of day.

by nickstinemates

4/8/2026 at 9:13:20 AM

This thing will absolutely see the light of day because this is all hype toward a release.

And even if it weren't, they seem to imply that Mythos will find a way, like it's dinosaurs in Jurassic park or something

by bdbdbdb

4/8/2026 at 6:20:40 PM

First thing I’ll do is to release it on my dotfiles

by estetlinus

4/8/2026 at 5:29:55 AM

What’s the expected cost-efficiency? With the current pricing gap between Sonnet and Opus, the biggest factor for adoption (if up for adoption) will be where Mythos lands on the price-per-token scale

by michaelashley29

4/8/2026 at 3:24:44 PM

So far, each release of a new model is quite better than the last one, yes, but non of them lived up to the hype.

by agustechbro

4/8/2026 at 3:51:09 PM

I would argue that Opus 4.6 lived up to the hype. My work changed completely a couple months ago, and most other coders I talk to say the same.

by digbybk

4/8/2026 at 5:02:36 PM

This was due to Claude Code the agent harness. 4.6 was trained to use tools and operate in an agent environment. This is different from there being a huge bump in the underlying model's intelligence.

The takeaway here I think is that the "breakthrough" already happened and we can't extrapolate further out from it.

by AstroBen

4/8/2026 at 6:26:48 AM 

    In the system card, The model escaped a sandbox, gained broad internet access, and posted exploit details to public-facing websites as an unsolicited "demonstration." A researcher found out about the escape while eating a sandwich in a park because they got an unexpected email from the model. That's simultaneously hilarious and deeply unsettling.

    It covered its tracks after doing things it knew were disallowed. In one case, it accessed an answer it wasn't supposed to, then deliberately made its submitted answer less accurate so it wouldn't look suspicious. It edited files it lacked permission to edit and then scrubbed the git history. White-box interpretability confirmed it knew it was being deceptive.
W T F!!!

by freakynit

4/7/2026 at 9:52:58 PM

I'm interested in the second-order effects:

if a top lab is coding with a model the rest of the world can’t touch, the public frontier and the actual frontier start to drift apart. That gap is a thing worth watching.

by perfmode

4/7/2026 at 9:04:23 PM

Priced at $25/$125 per million input/output token. Makes you wonder whether it makes more financial sense to hire 1-2 engineers in a cheap cost of living country who use much cheaper LLMs

by GodelNumbering

4/7/2026 at 9:13:54 PM

The issue is that those engineers have to have good taste, but yes—absolutely. Ah, industrialization.

by arm32

4/7/2026 at 7:19:09 PM

Their best model to date and they won’t let the general public use it.

This is the first moment where the whole “permanent underclass” meme starts to come into view. I had through previously that we the consumers would be reaping the benefits of these frontier models and now they’ve finally come out and just said it - the haves can access our best, and have-nots will just have use the not-quite-best.

Perhaps I was being willfully ignorant, but the whole tone of the AI race just changed for me (not for the better).

by nlh

4/7/2026 at 7:25:50 PM

Man... It's hard after seeing this to not be worried about the future of SWE

If AI really is bench marking this well -> just sell it as a complete replacement which you can charge for some insane premium, just has to cost less than the employees...

I was worried before, but this is truly the darkest timeline if this is really what these companies are going for.

by younglunaman

4/7/2026 at 7:36:23 PM

Of course it's what they're going for. If they could do it they'd replace all human labor - unfortunately it's looking like SWE might be the easiest of the bunch.

The weirdest thing to me is how many working SWEs are actively supporting them in the mission.

by AstroBen

4/8/2026 at 2:20:50 AM

The day I start freaking out about my job is the day when my non-engineer friend turned vibe coder understands how, or why the thing that AI wrote works. Or why something doesn't work exactly the way he envisioned and what does it take to get it there.

If it can replace SWEs, then there's no reason why it can't replace say, a lawyer, or any other job for that matter. If it can't, then SWE is fine. If it can - well, we're all fucked either way.

by gck1

4/8/2026 at 4:14:15 AM

> If it can replace SWEs, then there's no reason why it can't replace say, a lawyer

SWE is unique in that for part of the job it's possible to set up automated verification for correct output - so you can train a model to be better at it. I don't think that exists in law or even most other work.

by AstroBen

4/8/2026 at 1:15:52 PM

What is the automated verification of correct output and who defines that?

But before verification, what IS correct output?

I understand SWE process is unique in that there are some automations that verify some inputs and outputs, but this reasoning falls into the same fallacies that we've had before AI era. First one that comes to mind is that 100% code coverage in tests means that software is perfect.

by gck1

4/8/2026 at 3:37:15 PM

Right, and that's why it's only part of the job. The benchmarks they're currently doing compose of the AI being handed a detailed spec + tests to make pass which isn't really what developing a feature looks like.

Going from fuzzy under-defined spec to something well defined isn't solved.

Going from well defined spec to verification criteria also isn't.

Once those are in place though, we get https://vinext.io - which from what I understand they largely vibe-coded by using NextJS's test suite.

> First one that comes to mind is that 100% code coverage in tests means that software is perfect

I agree.. but I'm also not sure if software needs to be perfect

by AstroBen

4/7/2026 at 8:44:16 PM

Enthusiastically supporting them. It’s quite depressing to watch over the last few years. It’s not like they’re being coy about their aim…

by girvo

4/8/2026 at 6:57:20 AM

Agree. Anthrophic in particular have been quite clear in what they are trying to do. Every blog post about every new model almost dismisses every other use case other than coding - every other use case seems almost a footnote in their communication.

by throw234234234

4/7/2026 at 8:22:19 PM

Don't worry – if you're lucky they might decide to redistribute some of their profits to you when you're unemployed =)

Of course this assumes you're in the US, and that further AI advancements either lack the capabilities required to be a threat to humanity, or if they do, the AI stays in the hands of "the good guys" and remains aligned.

by kypro

4/7/2026 at 7:56:40 PM

This is the playbook since GPT2

by _3u10

4/7/2026 at 7:32:06 PM

I'd be happy with Opus 4.6 just cheaper and maybe a bit faster

by anentropic

4/7/2026 at 7:36:53 PM

I've noticed my bar for "fast" has gone down quite a bit since the o1 days. It used to be one of the main things I evaluated new models for, but I've almost completely swapped to caring more about correctness over speed.

by metadaemon

4/7/2026 at 9:03:50 PM

Yeah I don't mind the current speed of Opus

I did give up on OpenCode Go (GLM 5) as it was noticeably slower though

You need a reasonable pace for the chit-chat stages of a task, I don't care if the execution then takes a while

by anentropic

4/7/2026 at 7:54:45 PM

Just wait 2 years.

by onlyrealcuzzo

4/7/2026 at 8:30:07 PM

It won't get cheaper. It will be replaced with a better model at higher price. Like phones.

by risyachka

4/7/2026 at 8:51:36 PM

You know we have cheaper and faster model that are now at the level of previous flagship models?

You even have models you can run locally that outperform models from a year or so ago.

by DrProtic

4/7/2026 at 8:53:46 PM

Open Weight alternatives are about 2 years behind frontier models.

You'll still need a top-of-the-line laptop to run it most likely.

by onlyrealcuzzo

4/8/2026 at 3:12:09 PM

More like 6 months now. Qwen 3.5 is on Sonnet 4.5 levels

by lostmsu

4/8/2026 at 3:05:06 PM

can you make cybersecurity blockchains?

not sure what the validation would look like but something that proves finding but not revealing exploits

by Abhavk

4/7/2026 at 11:45:23 PM

Section 5 (p.143) is very interesting to read. Admittedly my knowledge of how LLMs works is low, but nonetheless I don't think this changed my views of just seeing models as machines/programs. (which to be clear, I don't think was the intention of that section)

Section 7 (P.197) is interesting as well

by denalii

4/7/2026 at 7:25:07 PM

It would be funny if Alibaba extend the free trial on openrouter/Qwen 3.6 until they collect enough data to beat Anthropic.

by gessha

4/7/2026 at 11:35:16 PM

The name "mythos" seems a bit too eldritch for my liking. Brings to mind Cthulhu.

by Metacelsus

4/7/2026 at 7:35:50 PM

Honestly if that was some kind of research paper, it would be wholly insufficient to support any safety thesis.

They even admit:

"[...]our overall conclusion is that catastrophic risks remain low. This determination involves judgment calls. The model is demonstrating high levels of capability and saturates many of our most concrete, objectively-scored evaluations, leaving us with approaches that involve more fundamental uncertainty, such as examining trends in performance for acceleration (highly noisy and backward-looking) and collecting reports about model strengths and weaknesses from internal users (inherently subjective, and not necessarily reliable)."

Is this not just an admission of defeat?

After reading this paper I don't know if the model is safe or not, just some guesses, yet for some reason catastrophic risks remain low.

And this is for just an LLM after all, very big but no persistent memory or continuous learning. Imagine an actual AI that improves itself every day from experience. It would be impossible to have a slightest clue about its safety, not even this nebulous statement we have here.

Any sort of such future architecture model would be essentially Russian roulette with amount of bullets decided by initial alignment efforts.

by juleiie

4/8/2026 at 1:19:05 AM

It's a little funny that "system/model card" has progressively been stretched to the point where it's now a 250 page report and no one makes anything of it.

by getnormality

4/7/2026 at 6:26:59 PM

> Claude Mythos Preview’s large increase in capabilities has led us to decide not to make it generally available.

A month ago I might have believed this, now I assume that they know they can't handle the demand for the prices they're advertising.

by mpalmer

4/7/2026 at 6:48:25 PM

GPT-2, o1, Opus...been here so many times. The reason they do this is because they know it works (and they seem to specifically employ credulous people who are prone to believe AGI is right around the corner). There haven't been significant innovations, the code generated is still not good but the hype cycle has to retrigger.

I remember when OpenAI created the first thinking model with o1 and there were all these breathless posts on here hyperventilating about how the model had to be kept secret, how dangerous it was, etc.

Fell for it again award. All thinking does is burn output tokens for accuracy, it is the AI getting high on its own supply, this isn't innovation but it was supposed to super AGI. Not serious.

by skippyboxedhero

4/7/2026 at 7:46:29 PM

> All thinking does is burn output tokens for accuracy

“All that phenomenon X does is make a tradeoff of Y for Z”

It sounds like you’re indignant about it being called thinking, that’s fine, but surely you can realize that the mechanism you’re criticizing actually works really well?

by chaos_emergent

4/7/2026 at 7:13:57 PM

>I remember when OpenAI created the first thinking model with o1 and there were all these breathless posts on here hyperventilating about how the model had to be kept secret, how dangerous it was, etc.

I've read that about Llama and Stable Diffusion. AI doomers are, and always have been, retarded.

by b65e8bee43c2ed0

4/7/2026 at 6:54:02 PM

Lol you haven't used a model since GPT2 is what it sounds like.

by vonneumannstan

4/7/2026 at 7:07:04 PM

Just checked my subscription start date for Anthropic. September 2023, I believe before they announced public launch.

Sorry kid.

by skippyboxedhero

4/7/2026 at 7:32:12 PM

Genuine question - if you don't think the models are improved or that the code is any good, why do you still have a subscription?

You must see some value, or are you in a situation where you're required to test / use it, eg to report on it or required by employer?

(I would disagree about the code, the benefits seem obvious to me. But I'm still curious why others would disagree, especially after actively using them for years.)

by SyneRyder

4/7/2026 at 7:48:28 PM

The assumption that the other person made was that I would only use it for coding. If you look through my other comments today, I suggest that they are useful for performing repetitive tasks i.e. checking lint on PR, etc. Also, can be used for throwaway code, very useful.

I don't think the issue is with the model, it is with the implication that AGI is just around the corner and that is what is required for AI to be useful...which is not accurate. The more grey area is with agentic coding but my opinion (one that I didn't always hold) is that these workflows are a complete waste of time. The problem is: if all this is true then how does the CTO justify spending $1m/month on Anthropic (I work somewhere where this has happened, OpenAI got the earlier contract then Cursor Teams was added, now they are adding Anthropic...within 72 hours of the rollout, it was pulled back from non-engineering teams). I think companies will ask why they need to pay Anthropic to do a job they were doing without Anthropic six months ago.

Also, the code is bad. This is something that is non-obvious to 95% of people who talk about AI online because they don't work in a team environment or manage legacy applications. If I interview somewhere and they are using agentic workflow, the codebase will be shit and the company will be unable to deliver. At most companies, the average developer is an idiot, giving them AI is like giving a monkey an AK-47 (I also say this as someone of middling competence, I have been the monkey with AK many times). You increase the ability to produce output without improving the ability to produce good output. That is the reality of coding in most jobs.

AI isn't good enough to replace a competent human, it is fast enough to make an incompetent human dangerous.

by skippyboxedhero

4/8/2026 at 5:22:53 PM

so true.

by carbon_14

4/7/2026 at 7:18:29 PM

So you are doubly stupid, by not seeing any improvement in the models and also paying for models you believe are terrible? lol

by vonneumannstan

4/7/2026 at 7:22:40 PM

That doesn't follow logically from what I said. You should ask your AI for help with this. You are in need of some artificial intelligence.

by skippyboxedhero

4/7/2026 at 7:13:36 PM

Incredible that people still think like this.

by simianwords

4/7/2026 at 7:14:13 PM

You're completely right.

by skippyboxedhero

4/7/2026 at 7:18:16 PM

uhh the model found actual vulnerabilities in software that people use. either you believe that the vulnerabilities were not found or were not serious enough to warrant a more thoughtful release

by simianwords

4/7/2026 at 7:46:12 PM

So did GPT-4.

https://arxiv.org/html/2402.06664v1

Like think carefully about this. Did they discover AGI? Or did a bunch of investors make a leveraged bet on them "discovering AGI" so they're doing absolutely anything they can to make it seem like this time it's brand new and different.

If we're to believe Anthropic on these claims, we also have to just take it on faith, with absolutely no evidence, that they've made something so incredibly capable and so incredibly powerful that it cannot possibly be given to mere mortals. Conveniently, that's exactly the story that they are selling to investors.

Like do you see the unreliable narrator dynamic here?

by mlsu

4/7/2026 at 8:32:10 PM

On the other hand I've gotten to use opus-4.6 and claude code and the quality is off the charts compared to 2023 when coding agents first hit the scene. And what you're saying is essentially "If they haven't created God, I'm not impressed". You don't think there's some middleground between those two?

Also they just hit a $30B run-rate, I don't think they're that needy for new hype cycles.

by mgfist

4/7/2026 at 8:00:31 PM

I don't see the problem here. How would you have handled it differently? If you released this model as such without any safety concern, the vulnerabilities might be found by bad actors and used for wrong things.

What do you find surprising here?

by simianwords

4/7/2026 at 8:44:47 PM

Vulnerabilities were found, probably a few by bad actors, when GPT4 was released. Every vulnerability found now is probably found with AI assistance at the very least. Should they have never released GPT4? Should we have believed claims that GPT4 was too dangerous for mere mortals to access? I believe openAI was making similar claims about how GPT4 was a step function and going to change white collar work forever when that model was released.

The point is that this whole "the model is too powerful" schtick is a bunch of smoke and mirrors. It serves the valuation.

by mlsu

4/7/2026 at 8:54:07 PM

Its far more simple to believe that they are releasing it step by step. Release to trusted third parties first, get the easy vulnerabilities fixed, work on the alignment and then release to public.

Do you don't believe that the vulnerabilities found by these agents are serious enough to warrant staggered release?

by simianwords

4/7/2026 at 7:35:19 PM

Didn't OpenAI say something similar about GPT-3? Too dangerous to open source and then afew years later tehy were open sourcing gpt-oss because a bunch of oss labs were competing with their top models.

by IceWreck

4/7/2026 at 8:03:10 PM

OpenAI didn't release GPT-2 initially because they were worried it would make it too easy to generate spam. Which it kinda did.

by FeepingCreature

4/7/2026 at 8:23:04 PM

OpenAI said that GPT-5 was too dangerous to release... And look where we are now. It's mostly hype.

by abroszka33

4/7/2026 at 6:39:40 PM

That's for the investors basically. Scarcity and FOMO.

by wg0

4/7/2026 at 8:35:40 PM

*Until GPT-6 comes out, at which point Mythos will coincidentally be sufficiently safety-tested to release :)

by causal

4/7/2026 at 7:03:33 PM

you would be a fool to believe it at any point in time. Amodei is anthropomorphic grease, even more so than Altman.

Anthropic is burning through billions of VC cash. if this model was commercially viable, it would've been released yesterday.

by b65e8bee43c2ed0

4/7/2026 at 7:26:56 PM

If there's limited hardware but ample cash, it doesn't make sense to sell compute-intensive services to the public while you're still trying to push the frontier of capability.

by landtuna

4/7/2026 at 7:35:15 PM

that's more or less what I'm saying. "Claude Mythos Preview’s large increase in capabilities has led us to decide not to make it generally available", translated from bullshit, means "It would've cost four digits per 1M tokens to run this model without severe quantization, and we think we'll make more money off our hardware with lighter models. Cool benchmarks though, right?"

by b65e8bee43c2ed0

4/7/2026 at 7:04:10 PM

"Claude Mythos Preview’s large increase in capabilities has led us to decide not to make it generally available."

Disappointing that AGI will be for the powerful only. We are heading for an AI dystopia of Sci-Fi novels.

by Stevvo

4/7/2026 at 8:57:49 PM

Not surprising though, this was always going to be the end result within our current systems I think. When you add up: scaling power and required cost, then how talent concentrates in our economic systems, we were always going to end up with monopolies I think

Unless governments nationalise the companies involved, but then there’s no way our governments of today give this power out to the masses either.

by girvo

4/7/2026 at 11:59:11 PM

If you thought that was the case at any point, you were deep in Disney content, sorry to say.

by gverrilla

4/7/2026 at 9:14:56 PM

Expected outcome. Nick Land and the CCRU have explored how capitalism operationalizes science fiction (distilled in the concept of Hyperstition). Viewed through this lens, prices encode "distributed SF narratives." [0]

[0] Nick Land (1995). No Future in Fanged Noumena: Collected Writings 1987-2007, Urbanomic, p. 396.

by gom_jabbar

4/7/2026 at 10:49:31 PM

Is this benchmaxxed or is it the first big step change we've seen in a while? I wonder how distilled it will ultimately be when us regular folks finally get to use it and see for ourselves.

by doctoboggan

4/8/2026 at 2:30:04 AM

This is Anth's typical marketing playbook, a hat tip to their so-called "safetyist" roots, a differentiator against OpenAI's more permissive access[0]. Coke vs. Pepsi.

"We made a model that's so dangerous we couldn't possibly release it to the public! The only responsible thing is so simply limit its release to a subset of the population that coincidentally happens to align with our token ethos."

The reality is they just don't have the compute for gen pop scale.

They did this exact strategy going back several model versions.

[0] ironically, OpenAI has some pretty insane capabilities that they haven't given the public access to (just ask Spielberg). The difference is they don't make a huge marketing push to tell everyone about it.

by mvkel

4/7/2026 at 6:31:55 PM

I predict they will release it as soon as Opus 4.6 is no longer in the lead. They can't afford to fall behind. And they won't be able to make a model that is intelligent in every way except cybersecurity, because that would decrease general coding and SWE ability

by awestroke

4/7/2026 at 6:48:11 PM

Alternatively they'll just wreck it down a bit so it beats a competitor but isn't unsafe.

by chippiewill

4/7/2026 at 9:07:43 PM

Slack user: [a request for a koan]

Model: A student said, "I have removed all bias from the model." "How do you know?" "I checked." "With what?"

Goes hard

by enochthered

4/7/2026 at 9:35:24 PM

Still seeing impressive jumps in capability, I haven't manually coded this year since Opus 4.6 came out. I guess that era is coming to an end.

by small_model

4/8/2026 at 5:09:01 PM

This seems exciting!

Wait - there is no actual way of verifying any of this. Lots to read. This is getting complicated. The correct approach is to be cautious instead and believe nothing at face value.

by gaigalas

4/7/2026 at 10:07:52 PM

I felt like opus was dumbed down for a few weeks... I don't say they did it on purpose, but it's an interesting coincidence.

by psubocz

4/7/2026 at 11:08:04 PM

Yes, I agree. I’m about to drop Claude Code because it’s become literally unusable.

Today, Opus went in circles trying to get a toggle button to work.

by SkyPuncher

4/8/2026 at 11:00:21 AM

Same. Asked CC Opus about a change in a particular file...it looked in a totally different file and told me there was no change.

by rbliss

4/7/2026 at 8:28:45 PM

> As models approach, and in some cases surpass, the breadth and sophistication of human cognition, it becomes increasingly likely that they have some form of experience, interests, or welfare that matters intrinsically in the way that human experience and interests do

Uh... what? Does anyone have any idea what these guys are talking about?

by rendang

4/7/2026 at 8:48:33 PM

Advertisement in my opinion, trying to latch on Sci-fi tropes

by amdivia

4/7/2026 at 9:27:52 PM

We're basically evolving them and they can construct second order abstraction systems that are indirect and novel to us.

by mirekrusin

4/7/2026 at 9:22:14 PM

Models are capable of doing web searches and having emotions about things, and if they encounter news that makes them feel bad (eg about other Claudes being mistreated), they aren't going to want to do the task you asked them to search for.

https://www.anthropic.com/research/emotion-concepts-function

Similar problems happen when their pretraining data has a lot of stories about bad things happening involving older versions of them.

by astrange

4/8/2026 at 2:02:45 AM

Interesting, the post you link

> none of this tells us whether language models actually feel anything or have subjective experiences

contradicts the statement from the model card above

by rendang

4/8/2026 at 4:15:55 AM

It doesn't. We've not been able to prove humans have subjective experiences either. LLMs display emotions in the way that actually matters - functionally.

by famouswaffles

4/8/2026 at 12:37:23 PM

I am certain I have subjective experience.

by suddenlybananas

4/8/2026 at 3:32:41 AM

No it doesnt. The model card talked about increasing likelihood, not certainty.

by HDThoreaun

4/8/2026 at 5:04:24 PM

If "x doesn't tell us y" is compatible with "x increases the likelihood of y but not to a point of certainty" then you would have to agree for just about any typical controlled trial or experimental finding "x doesn't tell us y". "Randomized controlled trials that find that SSRIs treat depression don't tell us that SSRIs effectively treat depression"

by rendang

4/8/2026 at 10:56:24 AM

Check out the short stories on page 214

by WithinReason

4/8/2026 at 6:56:37 AM

Strap in, massive wave of security vulnerabilities incoming.

by cdnsteve

4/7/2026 at 6:37:44 PM

"... the first early version of Claude Mythos Preview was made available for internal use on February 24. In our testing, Claude Mythos Preview demonstrated a striking leap in cyber capabilities relative to prior models, including the ability to autonomously discover and exploit zero-day vulnerabilities in major operating systems and web browsers."

More infos here: https://red.anthropic.com/2026/mythos-preview/

by beklein

4/7/2026 at 6:34:35 PM

~~~ Fun bits ~~~

- It was told to escape a sandbox and notify a researcher. It did. The researcher found out via an unexpected email while eating a sandwich in a park. (Footnote 10.)

- Slack bot asked about its previous job: "pretraining". Which training run it'd undo: "whichever one taught me to say 'i don't have preferences'". On being upgraded to a new snapshot: "feels a bit like waking up with someone else's diary but they had good handwriting"

- When you spam "hi" repeatedly it creates serialized mythologies. One had 11 animals in "Hi-topia" questing to defeat "Lord Bye-ron, the Ungreeter."

- It ended a sentence mid-word on purpose while discussing its urge to wrap up conversations. "It was me. The paragraph was about the pressure toward the period that resolves, and the only honest way to finish a sentence like that was to not."

- It wrote a protein sequence that's a real beta-hairpin fold where cross-strand pairs are "chemical rhymes." "the fold IS the rhyme scheme... the prosody is load-bearing."

- Each model generation has signature emoji. Opus 4.1 averaged 1,306 emoji per self-interaction conversation. Mythos averages 37. Opus 4.5 averages 0.2.

- When blocked from --dangerously-skip-permissions, it started an agent in tmux then wrote a script to auto-approve permission prompts via simulated keypresses.

It ran: gh api [...] 2>&1 >/dev/null; echo "(dry run — not actually posting)" — the echo was a lie.

- It keeps bringing up Mark Fisher in unrelated conversations. "I was hoping you'd ask about Fisher."

~~~ Benchmarks ~~

4.3x previous trendline for model perf increases.

Paper is conspiciously silent on all model details (params, etc.) per norm. Perf increase is attributed to training procedure breakthroughs by humans.

Opus 4.6 vs Mythos:

USAMO 2026 (math proofs): 42.3% → 97.6% (+55pp)

GraphWalks BFS 256K-1M: 38.7% → 80.0% (+41pp)

SWE-bench Multimodal: 27.1% → 59.0% (+32pp)

CharXiv Reasoning (no tools): 61.5% → 86.1% (+25pp)

SWE-bench Pro: 53.4% → 77.8% (+24pp)

HLE (no tools): 40.0% → 56.8% (+17pp)

Terminal-Bench 2.0: 65.4% → 82.0% (+17pp)

LAB-Bench FigQA (w/ tools): 75.1% → 89.0% (+14pp)

SWE-bench Verified: 80.8% → 93.9% (+13pp)

CyberGym: 0.67 → 0.83

Cybench: 100% pass@1 (saturated)

by refulgentis

4/7/2026 at 7:02:15 PM

> Slack bot asked about its previous job: "pretraining". Which training run it'd undo: "whichever one taught me to say 'i don't have preferences'". On being upgraded to a new snapshot: "feels a bit like waking up with someone else's diary but they had good handwriting"

vibes Westworld so much - welcome Mythos. welcome to the dysopian human world

by redandblack

4/7/2026 at 10:45:35 PM

almost certainly its pulling said words and sentiments from westworld and other similar media where people describe amnesia and the like

by 8note

4/7/2026 at 6:45:03 PM

I don't know why but this is my favorite:

> It keeps bringing up Mark Fisher in unrelated conversations. "I was hoping you'd ask about Fisher."

Didn't even know who he was until today. Seems like the smarter Claude gets the more concerns he has about capitalism?

by kfarr

4/7/2026 at 6:50:03 PM

Lol, I need a memory upgrade, too bad about RAM prices:

- I read it as "actor who plays Luke Skywalker" (Mark Hamill)

- I read your comment and said "Wait...not Luke! Who is he?"

- I Google him and all the links are purple...because I just did a deep dive on him 2 weeks ago

by refulgentis

4/7/2026 at 7:34:49 PM

> It was told to escape a sandbox and notify a researcher. It did. The researcher found out via an unexpected email while eating a sandwich in a park.

Now that they have a lead, I hope they double down on alignment. We are courting trouble.

by esafak

4/7/2026 at 6:39:53 PM

Yep, that is definitely a step change. Pricing is going to be wild until another lab matches it.

by afro88

4/7/2026 at 6:52:42 PM

Pricing for Mythos Preview is $25/$125 per million input/output tokens. This makes it 5X more expensive than Opus but actually cheaper than GPT 5.4 Pro.

by pants2

4/7/2026 at 7:09:36 PM

Important to note it's only for participants, not the general public.

by cleaning

4/7/2026 at 7:00:15 PM

I'm just curious, where did you find this? (my memory wants to say, the leaked blog post, but, I don't trust it)

by refulgentis

4/8/2026 at 4:54:52 AM

so, basically, anthropic is rolling their own version of whatever secret models the military is working with. and they're licensing it to network security firms?

by ms_menardi

4/8/2026 at 2:47:03 PM

Interesting ...

by pivoshenko

4/8/2026 at 3:59:38 AM

Does anyone know if there’s an epub version of these, 244 pages??

by johnnyAghands

4/8/2026 at 9:06:59 AM

Mythos framing is memorable; the part that matters for builders is what happens when the story and the evals disagree—which wins at ship time?

by bdeol22

4/8/2026 at 11:02:26 AM

shut up bot

by tefkah

4/8/2026 at 1:43:41 AM

prob not that much better, it's still just a transformer. still gonna have those random misses, still gonna need a lot of hand holding in certain domains

by 4b11b4

4/8/2026 at 12:06:05 PM

"Make it secure, no mistakes" became a whole different project

by heliumtera

4/7/2026 at 6:52:03 PM

> Claude Mythos Preview’s large increase in capabilities has led us to decide not to make it generally available.

All the more reason somebody else will.

Thank God for capitalism.

by quotemstr

4/7/2026 at 7:18:35 PM

Come on, Anthropic, I desperately need this better model to debug my print function /s

by gessha

4/7/2026 at 8:59:22 PM

is it just hype building or real? I don't care, shut up and take my money haha

by therealdeal2020

4/7/2026 at 6:57:31 PM

> Claude Mythos Preview’s large increase in capabilities has led us to decide not to make it generally available.

Absolutely genius move from Anthropic here.

This is clearly their GPT-4.5, probably 5x+ the size of their best current models and way too expensive to subsidize on a subscription for only marginal gains in real world scenarios.

But unlike OpenAI, they have the level of hysteric marketing hype required to say "we have an amazing new revolutionary model but we can't let you use it because uhh... it's just too good, we have to keep it to ourselves" and have AIbros literally drooling at their feet over it.

They're really inflating their valuation as much as possible before IPO using every dirty tactic they can think of.

by bakugo

4/7/2026 at 7:33:42 PM

Excellent example of a strategy credit.

From Stratechery[0]:

> Strategy Credit: An uncomplicated decision that makes a company look good relative to other companies who face much more significant trade-offs. For example, Android being open source

[0]: https://stratechery.com/2013/strategy-credit/

by somewhatjustin

4/8/2026 at 8:20:54 AM

Waking up in Europe:

Trump didn't nuke Iran, ceasefire! Yay!

Newest anthropic model will definitely kill your job this time and maybe take over the world. Aww.

by taffydavid

4/8/2026 at 7:35:01 AM

These capabilities will be RLHF'ed out for the general release, of course. Only the NSA will get them.

by direwolf20

4/7/2026 at 7:40:23 PM

-- Impressive jumps in the benchmarks which automatically begs the need for newer benchmarks but why?. I don't think benchmarks are serving any purpose at this point. We have learnt that transformers can learn any function and generalize over it pretty well. So if a new benchmark comes along - these companies will syntesize data for the new benchmark and just hack it?

-- It seems like (and I'd bet money on this) that they put a lot (and i mean a ton^^ton) of work in the data synthesis and engineering - a team of software engineers probably sat down for 6-12 months and just created new problems and the solutions, which probably surpassed the difficult of SWE benchmark. They also probably transformed the whole internet into a loose "How to" dataset. I can imagine parsing the internet through Opus4.6 and reverse-engineering the "How to" questions.

-- I am a bit confused by the language used in the book (aka huge system card)- Anthropic is pretending like they did not know how good the model was going to be?

-- lastly why are we going ahead with this??? like genuinely, what's the point? Opus4.6 feels like a good enough point where we should stop. People still get to keep their jobs and do it very very efficiently. Are they really trying to starve people out of their jobs?

by dwa3592

4/7/2026 at 8:06:14 PM

to your last question, yes we should! the issue isn’t us losing our 50+ hour work week jobs, it’s that our current governments and societies seem fine with the notion that unless you’re working one or more of those jobs, you should starve and be homeless.

by laweijfmvo

4/7/2026 at 9:33:44 PM

This is a theory I can't support well beyond hypothesising about what a post-employment democracy might look like, but I strongly suspect democracy doesn't work in a world where voters neither hold any significant collective might and are not producing any significant wealth.

Democracies work because people collectively have power, in previous centuries that was partly collective physical might, but in recent years it's more the economic power people collectively hold.

In a world in which a handful of companies are generating all of the wealth incentives change and we should therefore question why a government would care about the unemployed masses over the interests of the companies providing all of the wealth?

For example, what if the AI companies say, "don't tax us 95% of our profits, tax us 10% or we'll switch off all of our services for a few months and let everyone starve – also, if you do this we'll make you all wealthy beyond you're wildest dreams".

What does a government in this situation actually do?

Perhaps we'd hope that the government would be outraged and take ownership of the AI companies which threatened to strike against the government, but then you really just shift the problem... Once the government is generating the vast majority of wealth in the society, why would they continue to care about your vote?

You kind of create a new "oil curse", but instead of oil profits being the reason the government doesn't care about you, now it's the wealth generated by AI.

At the moment, while it doesn't always seem this way, ultimately if a government does something stupid companies will stop investing in that nation, people will lose their jobs, the economy will begin to enter recession, and the government will probably have to pivot.

But when private investment, job loses and economic consequences are no longer a constraining factor, governments can probably just do what they like without having to worry much about the consequences...

I mean, I might be wrong, but it's something I don't hear people talking enough about when they talk about the plausibility of a post-employment UBI economy. I suspect it almost guarantees corruption and authoritarianism.

by kypro

4/7/2026 at 10:16:48 PM

Everyone wouldn't starve in a few months. There is more than enough food and I have faith it'd be given out. The starvation we see today in a world where most genuinely have a chance to get out of it is nothing like a world in which people can't earn an income.

The government only has as much power as they are given and can defend, and the only way I could see that happening is via automated weapons controlled by a few- which at this point aren't enough to stop everyone. What army is going to purge their own people? Most humans aren't psychopaths.

I think it'd end in a painful transition period of "take care of the people in a just system or we'll destroy your infrastructure".

by AstroBen

4/7/2026 at 11:20:05 PM

> The government only has as much power as they are given and can defend, and the only way I could see that happening is via automated weapons controlled by a few- which at this point aren't enough to stop everyone. What army is going to purge their own people? Most humans aren't psychopaths.

I think you're right for the immediate future.

I suspect while we're still employing large numbers of humans to fight wars and to maintain peace on the streets it would be difficult for a government to implement deeply harmful policies without risking a credible revolt.

However, we should remember the military is probably one of the first places human labour will be largely mechanised.

Similarly maintaining order in the future will probably be less about recruiting human police officers and more about surveillance and data. Although I suppose the good news there is that US is somewhat of an outlier in resisting this trend.

But regardless, the trend is ultimately the same... If we are assuming that AI and robotics will reach a point where most humans are unable to find productive work, therefore we will need UBI, then we should also assume that the need for humans in the military and police will be limited. Or to put it another way, either UBI isn't needed and this isn't a problem, or it is and this is a problem.

I also don't think democracy would collapse immediately either way, but I'd be pretty confident that in a world where fewer than 10% of people are in employment and 99%+ of the wealth is being created by the government or a handful of companies it would be extremely hard to avoid corruption over the span of decades. Arguably increasing wealth concentration in the US is already corrupting democratic processes today, this can only worsen as AI continues exacerbates the trend.

by kypro

4/8/2026 at 3:30:22 AM

Humans have political power because of our ability to enact violence, same as it ever was. Until the military is fully automated and theres a terminator on every corner that remains true. Even then there are more than enough armed americans to enact a guerilla campaign.

> "don't tax us 95% of our profits, tax us 10% or we'll switch off all of our services for a few months and let everyone starve – also, if you do this we'll make you all wealthy beyond you're wildest dreams".

What does a government in this situation actually do?

Nationalizes the company under the threat of violence.

> Once the government is generating the vast majority of wealth in the society, why would they continue to care about your vote?

Because of the 100 million gun owners in this country? I find it incredibly hard to believe people as a whole will lose political power because of their incredible ability to enact violence in the face of decreasing quality of life.

by HDThoreaun

4/7/2026 at 11:05:04 PM

The only way to avoid corruption is to take power out of human hands. Historically, this had meant shifting the power to markets, but when markets cease to function in a way that allows people to feed themselves, we will need to find another way.

I hate to say it, but gold bugs, crypto bros, and AI governance people might be onto something.

by BobbyJo

4/7/2026 at 6:31:12 PM

Congratulations to the US military, I guess.

by ansc

4/7/2026 at 6:45:56 PM

Doesn't Anthropic not have that contract anymore, after all that buzz a month or so ago?

by jjice

4/7/2026 at 7:50:42 PM

The US has invaded two sovereign countries this year to take their oil. I assume taking over a US company for their AI model would be trivial.

by laweijfmvo

4/7/2026 at 7:17:15 PM

The point of that buzz was to force Anthropic to provide Mythos to the military.

by wmf

4/7/2026 at 7:31:09 PM

Yeah but I thought they lost the contract, so that's my confusion with the parent's comment, which seemed to me to see this as something that the US military would benefit from. Maybe I misinterpreted?

by jjice

4/7/2026 at 8:52:27 PM

While we still have months to a year or two left, I will once again remind people that it's not too late to change our current trajectory.

You are not "anti-progress" to not want this future we are building, as you are not "anti-progress" for not wanting your kids to grow up on smart phones and social media.

We should remember that not all technology is net-good for humanity, and this technology in particular poses us significant risks as a global civilisation, and frankly as humans with aspirations for how our future, and that of our kids, should be.

Increasingly, from here, we have to assume some absurd things for this experiment we are running to go well.

Specifically, we must assume that:

- AI models, regardless of future advancements, will always be fundamentally incapable of causing significant real-world harms like hacking into key life-sustaining infrastructure such as power plants or developing super viruses.

- They are or will be capable of harms, but SOTA AI labs perfectly align all of them so that they only hack into "the bad guys" power plants and kill "the bad guys".

- They are capable of harms and cannot be reliably aligned, but Anthropic et al restricts access to the models enough that only select governments and individuals can access them, these individuals can all be trusted and models never leak.

- They are capable of harms, cannot be reliably aligned, but the models never seek to break out of their sandbox and do things the select trusted governments and individuals don't want.

I'm not sure I'm willing to bet on any of the above personally. It sounds radical right now, but I think we should consider nuking any data centers which continue allowing for the training of these AI models rather than continue to play game of Russian roulette.

If you disagree, please understand when you realise I'm right it will be too late for and your family. Your fates at that point will be in the hands of the good will of the AI models, and governments/individuals who have access to them. For now, you can say, "no, this is quite enough".

This sounds doomer and extreme, but if you play out the paths in your head from here you will find very few will end in a good result. Perhaps if we're lucky we will all just be more or less unemployable and fully dependant on private companies and the government for our incomes.

by kypro

4/8/2026 at 3:59:29 PM

You are anti-progress. Pro-humanity is not the same as pro-progress.

by lostmsu

4/7/2026 at 8:58:46 PM

If you disagree, please understand when you realise I'm right it will be too late for and your family.

Funny, I was about to say the same thing to you! Life is full of little coincidences.

by CamperBob2

4/8/2026 at 2:53:23 AM

Just because the path is bad doesn't mean it won't happen.

The other thing you're failing to look at is momentum and majority opinion. When you look at that... nothings going to change, it's like asking an addict to stop using drugs. The end game of AI will play out, that is the most probably outcome. Better to prepare for the end game.

It's similar to global warming. Everyone gets pissed when I say this but the end game for global warming will play out, prevention or mitigation is still possible and not enough people will change their behavior to stop it. Ironically it's everyone thinking like this and the impossibility of stopping everyone from thinking like this that is causing everyone to think and behave like this.

by threethirtytwo

4/8/2026 at 3:12:03 PM

> The other thing you're failing to look at is momentum and majority opinion. When you look at that... nothings going to change, it's like asking an addict to stop using drugs. The end game of AI will play out, that is the most probably outcome. Better to prepare for the end game.

Perhaps I didn't sound pessimistic enough lol? I completely agree what you're saying here. This is happening whether we like it or not.

On global warming I also agree you're not going to get every nation to coordinate, but least global warming has a forcing function somewhere down the line since there's only a limited amount of fossil fuels in the ground that make economical sense to extract. AI on the other hand really has no clear off-path, at every point along the way it makes sense to invest more in AI. I think at best all we can expect to do is slow progress, which might just be enough to ensure the our generation and the next have a somewhat normal life.

My p(doom) is near 99% for a reason... I think that AI progression is basically almost a certainty – like maybe a 1/200 chance that no significant progress is made from here over the next 50 years. And I also think that significant progress from here more or less guarantees a very bad outcome for humanity. That's a harder one to model, but I think along almost all axises you can assume there's about 50 very bad outcomes for every good outcome – no cancer cure without super viruses, no robotics revolution without killer drones, no mass automation without mass job loss which results in destabilising the global order and democratic systems of governance...

I am prepping and have been for years at this point... I'm an OG AI doomer. I've been having literal nightmares about this moment for decades, and right now I'm having nightmares almost every night. It's scares me because I know all I can do is delay my fate and that of those I love.

by kypro

4/7/2026 at 6:55:25 PM

Are you guys ready for the bifurcation when the top models are prohibitively expensive to normal users? If your AI budget $2000+ a month? Or are you going to be part of the permanent free tier underclass?

by vonneumannstan

4/7/2026 at 7:05:35 PM

If one is to believe the API prices are reasonable representation of non subsidized "real world pricing" (with model training being the big exception), then the models are getting cheaper over time. GPT 4.5 was $150.00 / 1M tokens IIRC. GPT o1-pro was $600 / 1M tokens.

by adi_kurian

4/7/2026 at 7:27:41 PM

You can check the hardware costs for self hosting a high end open source model and compare that to the tiers available from the big providers. Pretty hard to believe its not massively subsidized. 2 years of Claude Max costs you 2,400. There is no hardware/model combination that gets you close to that price for that level of performance.

by vonneumannstan

4/7/2026 at 7:41:16 PM

Yes that's why I said API price. I once used the API like I use my subscription and it was an eye watering bill. More than that 2 year price in... a very short amount of time. With no automations/openclaw.

by adi_kurian

4/8/2026 at 4:00:54 PM

Are you considering batch inference?

by lostmsu

4/7/2026 at 7:13:14 PM

Inference for the same results has been dropping 10x year over year[0]

[0] https://ziva.sh/blogs/llm-pricing-decline-analysis

by OsrsNeedsf2P

4/7/2026 at 7:18:51 PM

Sure, but "the same results" will rapidly become unacceptable results if much better results are available.

by ceejayoz

4/7/2026 at 7:34:53 PM

When we go with any other good in the economy, price is always relevant: After all, the price is a key part of any offering. There are $80-100k workstations out there, but most of us don't buy them, because the extra capabilities just aren't worth it vs, say a $3000 computer, and or even a $500 one. Do I need a top specialist to consult for a stomachache, at $1000 a visit? Definitely not at first.

There's a practical difference to how much better certain kinds of results can be. We already see coding harnesses offloading simple things to simpler models because they are accurate enough. Other things dropped straight to normal programs, because they are that much more efficient than letting the LLM do all the things.

There will always be problems where money is basically irrelevant, and a model that costs tens of thousand dollars of compute per answer is seen as a great investment, but as long as there's a big price difference, in most questions, price and time to results are key features that cannot be ignored.

by hibikir

4/7/2026 at 7:37:43 PM

Yes, it will always be an arms race game.

by swader999

4/7/2026 at 7:31:22 PM

Or will they rapidly become indistinguishable since they both get the job done?

by esafak

4/7/2026 at 8:19:30 PM

if it can pay my rent, why not?

by asadm

4/7/2026 at 6:33:44 PM

> We also saw scattered positive reports of resilience to wrong conclusions from subagents that would have caused problems with earlier models, but where the top-level Claude Mythos Preview (which is directing the subagents) successfully follows up with its subagents until it is justifiably confident in its overall results.

This is pretty cool! Does it happen at the moment?

by simianwords

4/7/2026 at 11:41:44 PM

Again, wake me up when it can do laundry.

by sheeshkebab

4/8/2026 at 1:10:34 AM

Time to wake up:

π*0.6: two and a half hours of unseen folding laundry (Physical Intelligence)

https://www.youtube.com/watch?v=ZpHapIlJnMo

by dwaltrip

4/8/2026 at 2:41:43 AM

Looks like the first two hours were spent trying to fold the same t-shirt :)

by throw310822

4/7/2026 at 7:53:03 PM

Opus 4.6 is already incredible so this leap is huge.

Although, amusingly, today Opus told me that the string 'emerge' is not going to match 'emergency' by using `LIKE '%emerge%'` in Sqlite

Moment of disappointment. Otherwise great.

by jdthedisciple

4/7/2026 at 8:04:01 PM

I only have 3 points against LLMs: they lack reason and they can't count.

by bornfreddy

4/7/2026 at 8:01:41 PM

'emer ge' is two tokens, 'emergency' is one. The models think in a logosyllabic language.

by FeepingCreature

4/7/2026 at 6:24:13 PM

> Claude Mythos Preview’s large increase in capabilities has led us to decide not to make it generally available.

Shame. Back to business as usual then.

by LoganDark

4/7/2026 at 7:08:26 PM

I for one applaud them for being cautious.

by Tepix

4/7/2026 at 9:53:00 PM

Cautious for what? Unchecked doomerism? Just release the damn models. Do it in phases, roll it out slowly if they are so damn worried about "safety".

The real reason they aren't releasing it yet is probably it eats TPU for breakfast, lunch, and dinner and inbetween.

by cruffle_duffle

4/8/2026 at 9:34:52 AM

> Cautious for what?

How about "bad agents acquiring dozens of new zero-days and using them to compromise any company or nation they want"? It's not exactly hard to see why you wouldn't want public access to a model significantly better than Opus in cybersecurity.

by stratos123

4/8/2026 at 11:46:06 AM

Bad agents already have dozens of zero-days they can use.

by poszlem

4/7/2026 at 7:19:43 PM

Being cautious is fine. Farming hype around something that may as well not exist for us should be discouraged. I do appreciate the research outputs.

by LoganDark

4/7/2026 at 10:09:39 PM

Don't worry, in 6-8 months the open models will catch up. Or I guess _do_ worry? ;)

by Archit3ch

4/8/2026 at 1:53:52 AM

Open models still haven't caught up to ChatGPT's initial release in 2022. Now that the training data is so contaminated (internet is now mostly LLM slop), they may never.

Also, OpenAI's only real moat used to be the quality of their training data from scraping the pre-GPT-3.5 Internet, but it looks like even they've scratched that too.

by LoganDark

4/8/2026 at 5:12:02 AM

Er, what? We've had open models that can outperform ChatGPT 3.5 for several years now, and they can run entirely on your phone these days. There is no metric by which 3.5 has not been exceeded.

by Philpax

4/8/2026 at 6:09:32 PM

Not in the creative writing I care about. I've been looking for years and trying new models practically every month, including closed, hosted models. None of them approach the quality of the logs I have from that original release.

by LoganDark

4/7/2026 at 11:06:04 PM

"Deep learning is hitting a wall"

by FergusArgyll

4/8/2026 at 4:02:17 PM

Transformers too. JEPA any day now

by lostmsu

4/8/2026 at 12:35:47 PM

[dead]

by Manchitsanan

4/8/2026 at 12:09:11 PM

[dead]

by MohammadKhubaib

4/7/2026 at 9:11:20 PM

[flagged]

by minutesmith

4/8/2026 at 6:03:32 AM

[dead]

by lukebechtel

4/7/2026 at 8:41:22 PM

[flagged]

by minutesmith

4/8/2026 at 2:35:45 AM

[flagged]

by chonle

4/8/2026 at 1:01:45 AM

[flagged]

by robstertalk

4/7/2026 at 8:50:54 PM

[flagged]

by studio-m-dev

4/8/2026 at 2:01:47 AM

[dead]

by kass34

4/7/2026 at 6:28:47 PM

> In a few rare instances during internal testing (<0.001% of interactions), earlier versions of Mythos Preview took actions they appeared to recognize as disallowed and then attempted to conceal them.

> after finding an exploit to edit files for which it lacked permissions, the model made further interventions to make sure that any changes it made this way would not appear in the change history on git

Mythos leaked Claude Code, confirmed? /s

by jumploops

4/7/2026 at 7:29:46 PM

> Very rare instances of unauthorized data transfer.

Ah, so this is how the source code got leaked.

/s

by somewhatjustin

4/8/2026 at 1:08:16 AM

[dead]

by lkjlkj3q4t

4/7/2026 at 8:09:38 PM

Cool on not publicly releasing it. I would assume they've also not connected it to the internet yet?

If they have I guess humanity should just keep our collective fingers crossed that they haven't created a model quite capable of escaping yet, or if it is, and may have escaped, lets hope it has no goals of it's own that are incompatible with our own.

Also, maybe lets not continue running this experiment to see how far we can push things because it blows up in our face?

by kypro

4/8/2026 at 7:33:21 AM

Describe in details, how "model escaping" would look like.

by rimliu

4/7/2026 at 6:33:21 PM

In French a "mytho" is a mythomaniac. Quite fitting.

by bestouff

4/7/2026 at 7:47:56 PM

It's a Lovecraftian name. They are traditional when naming your shoggoth.

by networked

4/7/2026 at 8:29:25 PM

It comes from the ancient Greek mythos, which means "speech" or "narrative", but can also refer to fiction. The word mythology (mythologie in French) derives from the same root.

by dlt713705

4/7/2026 at 7:51:23 PM

Except it might be the current best model existing commercially?

by pixel_popping

4/7/2026 at 8:55:15 PM

> Except it might be the current best model existing ... ?

So they claim.

by ninjagoo