frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

The One-Step Trap (In AI Research)

http://incompleteideas.net/IncIdeas/OneStepTrap.html
20•jxmorris12•2h ago

Comments

ssivark•1h ago
Ha, interesting. I wasn't aware of Sutton's blog post, but if I might make a shameless plug, we demonstrated [1] exactly this problem (see section 4.4.3), and how multi-step world models (using diffusion models as the substrate) could be one potential answer.

Since then, I have come to like temporally-abstract models more and more. Rolling out in time -- either step-by-step or many steps at once -- suffers from the tyranny of the specific. For long horizon planning with agents, I care (often only approximately) about where I can end up, and seldom about exactly when I end up there. Successor features, GVFs, Forward-Backward representations, and the like seem like they have an elegant approach for structuring thinking at a "high level", instead of generating exponentially large search trees by rolling out microscopic world models.

[1] https://arxiv.org/abs/2410.05364 (funnily, from around the same time / few months after Sutton's blog post)

mxwsn•49m ago
This is the same reasoning behind why Yann Lecun thought test-time scaling would not work for LLMs: compounding error.

Instead, the more tokens LLMs use, the better their performance on many tasks. LLMs can self-correct, evidenced by the power of getting models to question themselves by emitting "Wait," in S1. https://arxiv.org/abs/2501.19393

gnabgib•43m ago
(2024)

Tiny Emulators

https://floooh.github.io/tiny8bit-preview/index.html
17•naves•28m ago•0 comments

Claude Code sends 33k tokens before reading the prompt; OpenCode sends 7k

https://systima.ai/blog/claude-code-vs-opencode-token-overhead
256•systima•2h ago•132 comments

I love LLMs, I hate hype

https://geohot.github.io//blog/jekyll/update/2026/07/12/i-love-llms.html
173•therepanic•2h ago•87 comments

Old and new apps, via modern coding agents

https://terrytao.wordpress.com/2026/07/11/old-and-new-apps-via-modern-coding-agents/
372•subset•9h ago•104 comments

I Learned to Read Again

https://substack.magazinenongrata.com/p/how-i-learned-to-read-again
32•georgex7•2h ago•4 comments

The One-Step Trap (In AI Research)

http://incompleteideas.net/IncIdeas/OneStepTrap.html
20•jxmorris12•2h ago•3 comments

Against Usefulness

https://www.motivenotes.ai/p/against-usefulness
52•supo•3h ago•11 comments

LARP – Revenue infrastructure for serious founders

https://www.larp.website/
75•BerislavLopac•3h ago•14 comments

Automation Without Understanding

https://arxiv.org/abs/2607.06377
65•root-parent•3h ago•33 comments

Why write code in 2026

https://softwaredoug.com/blog/2026/07/09/write-code
51•softwaredoug•2d ago•98 comments

Don't you mean extinct?

https://fabiensanglard.net/extinct/index.html
150•zdw•5h ago•78 comments

Migrating a production AI agent to GPT-5.6: 2.2x faster, 27% cheaper

https://ploy.ai/blog/migrating-a-production-ai-agent-to-gpt-5-6
38•brryant•3h ago•7 comments

Defining new Jax types with hijax

https://docs.jax.dev/en/latest/hijax_types.html
6•fhchl•56m ago•0 comments

Show HN: Nectar, a Rust-like React that compiles to WebAssembly

https://buildnectar.com
20•blakeburnette•6d ago•8 comments

Neocities: Create your own free website

https://neocities.org/
38•Tomte•1h ago•5 comments

How to read more books

https://scotto.me/blog/2026-07-12-how-to-read-more-books/
202•silcoon•5h ago•117 comments

The shingles vaccine may reduce the risk of dementia

https://www.economist.com/leaders/2026/07/09/a-no-brainer-for-protecting-your-brain
166•saikatsg•5h ago•121 comments

Why study Diophantine equations?

https://hidden-phenomena.com/articles/modular
52•mb1699•4h ago•17 comments

Deir El-Medina Strikes

https://en.wikipedia.org/wiki/Deir_el-Medina_strikes
30•mooreds•5d ago•4 comments

Theo de Raadt: "You've been smoking something mind altering" (2007)

https://marc.info/?l=openbsd-misc&m=119318909016582
64•turrini•4h ago•56 comments

Irish datacenters now guzzle 23% of the country's electricity

https://www.theregister.com/on-prem/2026/07/11/irish-datacenters-now-guzzle-23-of-the-countrys-el...
23•Bender•35m ago•12 comments

The power of collaboration: How we can reduce traffic congestion

https://research.google/blog/the-power-of-collaboration-how-we-can-reduce-traffic-congestion/
41•raahelb•5h ago•30 comments

The Seed Beneath the Snow

https://eli.li/the-seed-beneath-the-snow
4•surprisetalk•3d ago•0 comments

Understanding the Odin programming language

https://odinbook.com/
131•AlexeyBrin•8h ago•70 comments

Can We Understand How Large Language Models Reason?

https://cacm.acm.org/news/can-we-understand-how-large-language-models-reason/
46•adunk•2h ago•45 comments

Show HN: Shirei, cross-platform GUI framework in native Go

https://github.com/hasenj/go-shirei/
66•hsn915•3h ago•35 comments

Ghostel.el: Terminal emulator powered by libghostty

https://dakra.github.io/ghostel/
244•signa11•11h ago•42 comments

Vint Cerf, “father of the Internet”, is retiring

https://techcrunch.com/2026/06/30/the-father-of-the-internet-is-finally-retiring/
262•compiler-guy•2d ago•150 comments

What xAI's Grok build CLI sends to xAI: A wire-level analysis

https://gist.github.com/cereblab/dc9a40bc26120f4540e4e09b75ffb547
374•jhoho•19h ago•148 comments

"Stop Cutting Down Flock Cameras " (Mass Surveillance Is Good) [video]

https://www.youtube.com/watch?v=su-Ce4PlNz0
7•Bender•12m ago•0 comments