frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

"They're made out of weights"

https://maxleiter.com/blog/weights
58•MaxLeiter•3h ago

Comments

turtleyacht•1h ago
Numbers that dream.
CSSer•1h ago
It works until they get to the sentience part. Neat idea!
margalabargala•49m ago
Even there it works a bit.

> These models are the only other things we've ever met that can hold a conversation, and they're made out of weights

Is a fair point.

RodgerTheGreat•23m ago
Not especially. Depending on where you set your standards for "holding a conversation" you can satisfy the requirement with a classical markov chatterbot, a well-trained parrot, a copy of Eliza, or a telemarketer flowchart drawn on a sheet of paper. Only the markov bot is made out of "weights" in the sense of a statistical model.

Parrots are intelligent animals, albeit with a limited capacity for vocabulary and syntax compared to a human, and Eliza and the flowchart are made out of explicitly encoded rules and conversational tactics.

noosphr•56m ago
It's not often I see something that's fractally wrong but here we are.

There is a dictionary, it's called the tokenizer.

There are grammar rules, they are just very weak because the structure of human language is generally quite weak. When presented with languages which have strong consistent grammars the weights are very easily interpretable as a grammar: https://arxiv.org/abs/2201.02177

The point of the original short story is that the computational substrate doesn't matter when you have Turing completeness. This one seems to think that you don't need structure and interpretability just because you change substrates.

benlivengood•29m ago
I don't think the grokking paper is a great argument for the difference between weights and meat. E.g. https://en.wikipedia.org/wiki/Cortical_Labs learning to play Pong.

The tokenizer is, at best, a sensory mechanism as evidenced by 1) the random generation of the tokenization scheme, and 2) vastly different tokenization schemes produce virtually identical behavior. It'd be like if Noah Webster threw a bunch of movable type into a bucket (breaking some words in half) and then drew randomly to make the first English dictionary.

noosphr•11m ago
I'm kind of stunned that someone is using my work to tell me I'm wrong. I wrote the code for the dish brain pong and encoding information was a huge part of what that experiment was about.

So when I way that the grok paper and the pong paper fundamentally agree I have some idea of what I'm talking about.

throw310822•22m ago
oofbey•9m ago
I love this. For anybody not getting the joke, it’s riffing on the classic 1990s essay “They’re made out of meat.”

https://web.mit.edu/people/dpolicar/writing/prose/text/think...

> There are grammar rules

And they're made out of weights.

U.S. to Dismantle System Tracking Atlantic Currents That Are at Risk of Collapse

https://e360.yale.edu/digest/trump-ooi-amoc
76•rguiscard•2h ago•34 comments

Elixir v1.20: Now a gradually typed language

https://elixir-lang.org/blog/2026/06/03/elixir-v1-20-0-released/
584•cloud8421•7h ago•220 comments

I built a vulnerable app and spent $1,500 seeing if LLMs could hack it

https://kasra.blog/blog/i-spent-1500-seeing-if-llms-could-hack-my-app/
67•jc4p•2h ago•28 comments

Gemma 4 12B: A unified, encoder-free multimodal model

https://blog.google/innovation-and-ai/technology/developers-tools/introducing-gemma-4-12b/
717•rvz•10h ago•293 comments

The ways we contain Claude across products

https://www.anthropic.com/engineering/how-we-contain-claude
50•jbredeche•2h ago•20 comments

"They're made out of weights"

https://maxleiter.com/blog/weights
62•MaxLeiter•3h ago•10 comments

I was recently diagnosed with anti-NMDA receptor encephalitis

https://burntsushi.net/encephalitis/
518•Tomte•12h ago•160 comments

Artificial intelligence is not conscious – Ted Chiang

https://www.theatlantic.com/philosophy/2026/06/no-artificial-intelligence-is-not-conscious/687378/
268•lordleft•9h ago•481 comments

Uber's $1,500/month AI limit is a useful signal for AI tool pricing

https://simonwillison.net/2026/Jun/3/uber-caps-usage/
380•pdyc•14h ago•494 comments

DaVinci Resolve 21

https://www.blackmagicdesign.com/products/davinciresolve/whatsnew
400•pentagrama•12h ago•188 comments

Meteor Explodes over Massachusetts

https://www.nbcboston.com/news/local/meteor-explodes-over-massachusetts-what-we-know-and-where-it...
59•1970-01-01•2d ago•35 comments

The Ü Programming Language

https://github.com/Panzerschrek/U-00DC-Sprache/
33•deterministic•3h ago•15 comments

Failing grades soar with AI usage, dwindling math skills in Berkeley CS classes

https://www.dailycal.org/news/campus/academics/failing-grades-soar-as-professors-see-greater-ai-u...
18•littlexsparkee•2h ago•3 comments

Ableton Extensions SDK

https://www.ableton.com/en/live/extensions/
74•bennett_dev•6h ago•31 comments

Gooey: A GPU-accelerated UI framework for Zig

https://github.com/duanebester/gooey
141•ksec•9h ago•49 comments

A Post-Quantum Future for Let's Encrypt

https://letsencrypt.org/2026/06/03/pq-certs
224•SGran•11h ago•133 comments

Patching my guitar amp's firmware

https://mforney.org/blog/2026-05-28-patching-my-guitar-amps-firmware.html
41•birdculture•3d ago•7 comments

Launch HN: Hyper (YC P26) – Company brain to power agentic development

54•shalinshah•9h ago•55 comments

Journey to JPEG XL: open-source experiments shaped the future of image coding

https://opensource.googleblog.com/2026/06/journey-to-jpeg-xl-how-open-source-experiments-shaped-t...
46•ledoge•5h ago•23 comments

ESP32-S31

https://www.espressif.com/en/products/socs/esp32-s31
256•volemo•10h ago•144 comments

Stop Killing Games

https://jxself.org/stop-killing-games.shtml
198•amcclure•2d ago•192 comments

A Mathematician's Lament – Paul Lockhart (2002) [pdf]

https://worrydream.com/refs/Lockhart_2002_-_A_Mathematician%27s_Lament.pdf
42•xeonmc•5h ago•1 comments

Algorithmic Theming Engines

https://www.smashingmagazine.com/2026/05/building-self-correcting-color-systems-contrast-color/
6•mooreds•2d ago•0 comments

Self-hosted dev sandboxes with preview URLs (Docker, Go, no K8s)

https://github.com/tastyeffectco/sandboxes
56•tastyeffectco•7h ago•12 comments

Mathematicians issue warning as AI rapidly gains ground

https://www.science.org/content/article/mathematicians-issue-warning-ai-rapidly-gains-ground
182•pseudolus•16h ago•230 comments

Skyvern (YC S23) Is Hiring Open-Source Loving DevRel Engineers

https://www.ycombinator.com/companies/skyvern/jobs/1qRTlVx-founding-developer-marketing-open-sour...
1•suchintan•9h ago

Embryos shape their limbs: a key discovery of "genetic brakes"

https://nouvelles.umontreal.ca/en/article/2026/06/02/how-embryos-shape-their-limbs-a-key-discover...
56•gmays•9h ago•4 comments

A Man Who Reads Books for a Living (One Every Two Days)

https://lithub.com/the-man-who-reads-books-for-a-living-one-every-two-days/
79•gmays•6h ago•59 comments

PlayStation Architecture

https://www.copetti.org/writings/consoles/playstation/
264•gregsadetsky•16h ago•50 comments

Show HN: Bio Glyph – Turn Your Face into a One-Line Drawing

https://bio.bairui.dev/
6•subairui•2h ago•2 comments