alt.hn

8/18/2026 at 6:07:21 PM

Turbovec – Google's TurboQuant for vector search in Rust

 https://github.com/RyanCodrai/turbovec

by fittingopposite

8/18/2026 at 9:36:33 PM

FAISS is no longer close to SoTA:

https://ann-benchmarks.com/index.html https://vector-index-bench.github.io/ https://big-ann-benchmarks.com/neurips23.html

by Eridrus

8/19/2026 at 12:15:22 AM

I think their point is the size/performance tradeoff rather than outright performance. The point of TurboQuant is the size savings, while still giving high accuracy.

It's been a while, but I do recall some high-performing vector matching indexes being very large.

by nl

8/19/2026 at 2:48:22 AM

Surprised that usearch isn't in any of these, it's pretty fast.

by ehsanu1

8/18/2026 at 6:52:52 PM

Wow! 4GB for 10 million documents. This means one could build a reverse index much faster than before and devx processes like debugging, performance testing would become much smoother. Can't wait for the sqlite bindings to come out!

by ghm2199

8/18/2026 at 6:54:32 PM

Also the removal latency is on a log scale. Which is quite insane.

by ghm2199

8/18/2026 at 6:47:42 PM

It would be nice to have the README be a little more human written for a project where you actually want people to adopt it

by nharada

8/18/2026 at 7:44:52 PM

[flagged]

by badatnames

8/18/2026 at 8:01:19 PM

Then again, if the only thing the human doing is bitching about AI use, it's not really that comparatively useful.

by deeviant

8/19/2026 at 12:05:48 AM

Sure it is useful, the bitching is canary in the shit software mine. How do you know the software isnt shit if the Readme is shit?

by righthand

8/18/2026 at 7:02:06 PM

This looks perfect for local, privacy first search, but since it's built in Rust, has anyone tried compiling it to WASM to run directly inside a browser extension?

by anishvarghese

8/18/2026 at 8:08:21 PM

oxirs does embeddings and GraphRAG, and full text search with Tantivy; oxirs-vec, oxirs-graphrag

There's an oxirs-wasm with RDF and SPARQL bindings with a query budget. Tantivy-wasm says that the release WASM bundle is 1.5 MB.

cool-japan/oxirs: https://github.com/cool-japan/oxirs

oxirs-wasm: https://crates.io/crates/oxirs-wasm

tantivy-wasm: https://github.com/phiresky/tantivy-wasm

Is there an advantage to adding an MCP local memory interface over agent instructions on how to use a rust CLI?

And then write Markdown documents with Google OKF-like frontmatter YAML metadata for agents that work with tokens not linked data graphs; https://github.com/GoogleCloudPlatform/knowledge-catalog/blo...

by westurner

8/18/2026 at 9:09:45 PM

Can WASM use AVX512-VNNI?

by coredog64

8/18/2026 at 10:10:29 PM

No, WASM only has 128b SIMD instructions, for now.

by LtdJorge

8/19/2026 at 6:21:15 PM

> Can WASM use AVX512-VNNI?

wasmtime-mte implements ARM64 Memory Tagging Extensions in a fork of the wasmtime WASM runtime.

Memory Tagging Extensions for RISC-V would be a great project too

by westurner

8/19/2026 at 11:29:57 AM

nope

by m00dy

8/18/2026 at 7:27:32 PM

Also interested.

by cpursley

8/19/2026 at 3:33:48 AM

Interestingly, while we don't fine-tune generative models for Louie.ai, we found fine-tuning embedding models to be a major $ saver. Instead of 1K-2K wide frontier embedding vector lens... Just 64. Huge savings on vector DB $$$.

I'm curious how that works with something like turboquant. Not needed any more, still dominant, better together, ... .

by lmeyerov

8/18/2026 at 11:56:09 PM

I am not convinced that Turbovec yields better retrieval than the same amount of bits of a Matryoshka embedding.

by OutOfHere

8/18/2026 at 11:53:05 PM

What's a good embedding model and search to run locally? something fast and lightweight.

by cat-whisperer

8/18/2026 at 8:48:29 PM

Why not just use Qdrant? They've been integrating TurboQuant for months, works well.

by beernet

8/19/2026 at 2:26:53 AM

Integrated in 5 weeks and just expanded data types for turbo4 in last release. No longer need to store fp32 vectors if you don't need them

by kanungle

8/18/2026 at 8:36:41 PM

Bloviating nonsense, 3rd time I’ve seen something like this in HN since TurboQuant came out. You don’t need float32, never did. Source: I’ve been writing on device embedding code for 4 years.

by refulgentis

8/18/2026 at 9:29:49 PM

Another vibe coded slop where they can't even spend time on Readme or documentation around code...

by cute_boi

8/18/2026 at 6:59:20 PM

Who is this co-author called t <t@t>?

by burgerboii

8/18/2026 at 9:31:29 PM

As it is heavily vibe coded, I think member of technical staff at antropic has no clue....

Next Prompt: remove t@t and force commit.

by cute_boi

8/18/2026 at 6:30:19 PM

lancedb and duckdb integrations would be great...

by esafak

8/18/2026 at 6:59:59 PM

what could i use this for as part of my agentic workflow? codebase indexing? docs ?

by zuzululu

8/18/2026 at 7:01:56 PM

notes/docs/wiki is a great use case

by kyxsc

8/19/2026 at 2:05:58 PM

[flagged]

by spread2009

8/18/2026 at 8:13:21 PM

Well. That is insane. O_O Fantastic job!

by spoaceman7777

8/19/2026 at 12:21:33 PM

[dead]

by stlahxm

8/19/2026 at 1:57:24 PM

[dead]

by emerthorn

8/19/2026 at 12:17:10 AM

[dead]

by myshapeprotocol

8/19/2026 at 5:45:25 AM

[dead]

by anthropic-dario

8/18/2026 at 10:20:25 PM

[flagged]

by tracespect