Show HN: Replacing my OS process scheduler with an LLM

12/30/2025 at 7:09:11 PM

This is a pretty funny project, you've outsourced the neurotic developers that keep their task manager open and kill off processes they don't like.

I wouldn't call it replacing the scheduler though - more that you've made a scheduler manager.

by devmor

1/4/2026 at 11:03:19 AM

Assistant to the scheduler manager

by moffkalast

1/5/2026 at 12:17:11 PM

Assistant scheduler manager

by runelk

12/31/2025 at 7:08:24 AM

haha exactly. i realized i spent too much time staring at htop wondering what is this process?, so i decided to automate my own anxiety.

Scheduler Manager is definitely the more accurate term. Im just the middleman between the chaos and the kernel.

by ImPrajyoth

1/4/2026 at 12:01:25 PM

Now we need processes to gain awareness of the process manager and integrate an LLM into each process to argue with the process manager why it should let them live.

by QuantumNomad_

1/4/2026 at 12:07:58 PM

Imagine Chrome.exe pleading its case: 'Please, I need 4GB of RAM, the user might revisit that tab from 3 hours ago!'

while BrainKernel replies: 'Objection overruled. You have 5 seconds to wrap up before SIGKILL.'

I might actually have to build a 'Process Defense Attorney' agent now. The logs would be hilarious.

by ImPrajyoth

12/31/2025 at 5:01:33 AM

I resemble that comment!

But seriously, it does really bug me on principle that DropBox should use over half a GB simply because it uses Chromium, even when nothing is visible.

by DougN7

1/4/2026 at 9:36:31 AM

Maestral is a cross platform implementation of the Dropbox client API which I use on low end Linux machines.

by SanjayMehta

1/4/2026 at 5:16:09 PM

Thanks for the tip - I’ll take a look

by DougN7

1/4/2026 at 12:44:24 PM

For me it's LSP servers taking 2 gigs of RAM. With Antigravity, Google managed to go beyond this, it is totally unusable for me (but other VScode clones work fine, apart from the 2 Go LSP servers).

by Xmd5a

12/30/2025 at 4:50:24 PM

OP here. this is a cursed project lol, but i wanted to see: What happens if you replace the OS scheduler with an LLM?

With Groq speed (Llama 3 @ 800t/s), inference is finally fast enough to be in the system loop.

i built this TUI to monitor my process tree. instead of just showing CPU %, it checks the context (parent process, disk I/O) to decide if a process is compiling code or bloatware. It roasts, throttles, or kills based on that.

Its my experiment in "Intelligent Kernels" how they would be. i used Delta Caching to keep overhead low.

by ImPrajyoth

12/30/2025 at 9:23:51 PM

It really is cursed to be spending hundreds of watts of power in a datacenter somewhere to make a laptop run slightly faster.

by 2001zhaozhao

12/31/2025 at 7:06:33 AM

oh absolutely. burning a coal plant to decide if i should close discord is peak 2025 energy. strictly speaking, using the local model (Ollama) is 'free' in terms of watts since my laptop is on anyway, but yeah, if the inefficiency is the art, I'm the artist.

by ImPrajyoth

1/4/2026 at 10:26:13 AM

Running ollama to compute inference uses energy that wouldn't have been used if you weren't running ollama. There's no free lunch here.

by fragmede

1/4/2026 at 10:30:43 AM

An interesting thought experiment - a fully local, off-grid, off-network LLM device. Solar or wind or what have you. I suppose the Mac Studio route is a good option here, I think Apple make the most energy efficient high-memory options. Back of the napkin indicates it’s possible, just a high up front cost. Interesting to imagine a somewhat catastrophe-resilient LLM device…

by hebejebelus

1/5/2026 at 3:52:48 AM

Macs would be the most power efficient with faster memory but an AI Max 395+ based system would probably be the most cost efficient right now. A Framework Desktop with 128GB of shared RAM only pulls 400W (and could be underclocked) and is cheaper by enough that you could buy it plus 400W of solar panels and a decently large battery for less than a Mac Studio with 128GB of RAM. Unfortunately the power efficiency win is more expensive than just buying more power generation and storage ability.

by evilduck

1/5/2026 at 9:00:22 AM

I suppose in terms of catastrophe resilience repairability would be important, although how do you repair a broken GPU in any case. Probably cold backup machines is probably the more feasible way to extend lifetimes.

And yeah - I was thinking that actually power efficiency isn’t really a massive deal if you have some kind of thin client setup. The LLM nodes can be at millraces or some other power dense locations, and then the clients are basically 5W displays with an RF transceiver and a keyboard…

An entertaining thought experiment :)

by hebejebelus

1/4/2026 at 12:15:55 PM

That is the endgame.

I think we are moving toward a bilayered compute model: The Cloud: For massive reasoning.

The Local Edge: A small, resilient model that lives on-device and handles the OS loop, privacy, and immediate context.

BrainKernel is my attempt to prototype that Local Edge layer. Its messy right now, but I think the OS of 2030 will definitely have a local LLM baked into the kernel.

by ImPrajyoth

1/4/2026 at 1:37:25 PM

Well, on my Macbook, some of that already exists. In the Shortcuts app you can use the "Use Model" action which offers to run an LLM on apple's cloud, on-device, or other external service (eg ChatGPT). I use this myself already for several actions, like reading emails from my tennis club to put events in my calendar automatically.

Whether or not we'll see it lower down in the system I'm not sure. Honestly I'm not certain of the utility of an autonomous LLM loop in many or most parts of an OS, where (in general) systems have more value the more deterministic they are, but in the user space, who can say.

In any case, I certainly went down a fun rabbit hole thinking about a mesh network of LLM nodes and thin clients in a post-collapse world. In that scenario, I wonder if the utility of LLMs is really worth the complexity versus a kindle-like device with a copy of wikipedia...

by hebejebelus

12/31/2025 at 3:07:31 PM

> using the local model (Ollama) is 'free' in terms of watts since my laptop is on anyway

Now that’s a cursed take on power efficency

by bdhcuidbebe

12/31/2025 at 3:47:38 PM

efficiency is just a mindset. if i save 3 seconds of my own attention by burning 300 watts of gpu, the math works out in my favor!

by ImPrajyoth

1/4/2026 at 9:25:01 AM

"works out in my favor" is a pretty poor metric.

If I burn a billion tons of someone else's coal to make myself a paperclip (and don't have to breathe the outputs) it works out in my favor too.

by abeyer

1/4/2026 at 2:39:10 PM

An entire datacenter on the other hand, might be appealing to spot things you wouldn't otherwise see in a sea of logs and graphs.

by nubinetwork

12/30/2025 at 4:56:20 PM

You can't replace the NTOS scheduler. This is more of an automated (?) process manager.

by p_ing

12/30/2025 at 5:00:58 PM

you are technically right (the best kind of right). i am running in userspace, so i cant replace the actual thread scheduling logic in Ring 0 without writing a driver and BSODing my machine.

think of this more as a High-Level Governor. The NTOS scheduler decides which thread runs next, but this LLM decides if that process deserves to exist at all.

basically; NTOS tries to be fair to every process. BrainKernel overrides that fairness with judgment. if i suspend a process, i have effectively vetoed the scheduler.

by ImPrajyoth

12/30/2025 at 8:24:36 PM

> NTOS tries to be fair to every process

This is a super simplification of the NTOS scheduler. It's not that dumb!

> if i suspend a process, i have effectively vetoed the scheduler.

I mean, I suppose? It's the NTOS scheduler doing the suspension. It's like changing the priority level -- sure, you can do it, but it's generally to your detriment outside of corner cases.

by p_ing

1/4/2026 at 3:24:42 PM

I do wonder how painfully slow the computer would be if you actually did replace the in-kernel scheduler with an LLM...

by nijave

1/4/2026 at 10:21:26 AM

You're underselling this as a process manager, it could also be a productivity tool with some prompt changes; Determine procrastination apps: games, non-professional chat, video streaming and kill it.

by gillesjacobs

1/4/2026 at 12:03:32 PM

That is actually a brilliant pivot.

A 'Focus Mode' that doesn't just block URLs but literally murders the process if I open Steam or Civilization VI.

I could probably add a --mode strict flag that swaps the system prompt to be a ruthless productivity coach. 'Oh, you opened Discord? Roast and Kill.'

Thanks for the idea mate!

by ImPrajyoth

1/4/2026 at 12:44:16 PM

I was looking for a project which would run an LLM-powered character (like Clippy), who would periodically screenshot my screen and comment on my life choices.

Sadly the only project I've found was for windows OS

by tryauuum

1/4/2026 at 8:46:22 PM

If it doesn’t find a process that needs roasting or killing for a while, will it see itself as bloatware and commit suicide?

by Someone

1/4/2026 at 2:45:08 PM

Please add Roulette mode where a random process is killed every so often

by lorenzohess

1/4/2026 at 3:34:10 PM

You did not replace the OS process scheduler with an LLM.

by solarkraft

1/4/2026 at 7:26:02 PM

This is the one place that I would want Copilot running. It's giving me ideas :)

by 1970-01-01

1/4/2026 at 9:23:56 PM

Why not branch prediction with LLM?

by brcmthrowaway

1/4/2026 at 1:52:36 PM

Task manager, not scheduler.

by nialv7

1/5/2026 at 12:51:51 AM

Please name it juggler

by effnorwood

1/2/2026 at 3:17:46 PM

Ok

by pasisu

1/1/2026 at 6:38:41 AM

[dead]

by Agent_Builder

1/1/2026 at 11:58:47 AM

Great point. In a real kernel, non determinism is a bug. Here, it's a feature (or at least, a known hazard). To answer your question: There is no Ctrl+Z for SIGKILL. Once the LLM decides to kill a process, it's gone. My reasoning for 'rollback' is actually latency. I built in a 'Roasting Phase' where the agent mocks the process for a few seconds before executing the kill. That delay acts as an optimistic lock it gives me a window to veto the decision if I see it targeting something critical.

If I'm AFK and it kills my IDE? I treat that as the system telling me to touch grass.

by ImPrajyoth

1/4/2026 at 11:06:54 AM

You're absolutely right. Being AFK is a sign of spending too much time at your computer.

by GibbonBreath

1/4/2026 at 10:23:58 AM

[dead]

by CSOAI_Official