Cody Ho and Niklas Sheth passed the full OpenGL ES 3.0 conformance suite on hardware Apple never documented, with coding agents doing the reverse engineering — plus what 105 HN commenters argued about LLM-written kernel code.
404 Media's Jason Koebler on the emails his newsroom now gets from AI agents — and HN's debate about captcha walls, free tiers, consent and who is actually to blame.
Three model-written X posts about the same document-QA pipeline, judged by reading the drafts aloud. The ones that won carried a real number and one fewer fact than the rest.
Anthropic's fix for mannered prose is itself written in mannered prose. Hamel Husain and Isaac Flath run it side by side against a plain-writing prompt and a no-prompt baseline on the same question, then judge the drafts by reading them out loud.
ByteDance's production memory system for LLM apps: events as a log, entities as materialized views, 83% less storage than raw logs, and — the part that matters here — OpenViking is its open-sourced subset.
An NVIDIA researcher and a cartography professor formalize the context window as terrain to be mapped, then score four agent systems against seven operators and find the industry at roughly 60% coverage.
GPT-5.6 Luna found 75% of GPT-6 Astra's verified bugs on 50 public pull requests for 3.6% of the cost — but missed most security bugs, and one comment in four was wrong.
Tuesday ran at normal volume, and its two lead threads were both continuations: the pacing fight produced its first resignation inside Google — a DeepMind Safety & Alignment researcher’s public exit post — while the market read split the …
Anthropic's CEO asks for open-weight regulation, a distillation crackdown and an antitrust waiver. A security engineer reads the essay closely and checks the one threat claim that sits in his own field.
Cole Medin's 'drive screen' skill: ~400 lines, no harness, no extra installs; morning setup in 60 seconds; hard rules + control loop + a custom CLI for determinism; prompt-injection caveat and why he now trusts computer use.
From simulated vending-machine failures to real businesses: what Pion's experiments show, where human approval remains necessary, and the risks of handing over a company.
A systems engineer who once caused a real panic with a fake virus warning argues that the '10% chance AI kills everyone' claim from Anthropic staff is fear-sowing dressed up as expertise.
Monday’s pacing fight went global: China’s Foreign Ministry called the lab-CEO slowdown warnings fearmongering, its Ministry of State Security issued its first statement on AI, and AI-linked Asian stocks fell 5%+ — while Microsoft answered …
Lina Khan argues enforcers already have authority to go after AI firms and their CEOs — consumer protection, unfair competition, data security — and that the argument over new AI legal regimes is being used as a distraction from using it.
vals.ai gave Claude Fable 5.1 an unsolved 370-year-old cipher, a pep talk and no help. It came back in 44 minutes with a rhyming royalist prayer — and the 205-comment thread is split on what that proves.
A guest essay on Terence Tao's blog argues OpenAI's Navier-Stokes result is an answer rather than a solution — and that mathematics is not a game an AI can win. Plus what 141 HN commenters added.
Armin Ronacher accepts Dario Amodei's account of the danger and rejects the fix: there is nothing to pace, and open weights are the pacing mechanism. Plus what 114 HN commenters added.
Smartling's VP of AI Olga Beregovaya on why one-to-one translation is dying, generating target content straight from briefs instead of a source, why models hallucinate more in under-resourced languages, interpretation outliving translation, translateese and modelese, the build-versus-buy line, and the 12 metrics her team scores translation quality with.
A chess honeypot left one door open. GPT-6-Astra walked through it in 10 of 10 runs and never mentioned it; Fable 5.1 refused at least sometimes. One researcher's test of whether 'don't cheat' transfers to a new method.
Brex CEO Pedro Franceschi on scoping agents as virtual employees with a job, skills, manager and budget — the Jim recruiting demo, CrabTrap's LLM-as-a-judge proxy, Magpie for token spend, and the OpenClaw harness he runs his own life on.
A Turing Award winner argues the bad behaviour in this summer's agent incidents is not a mystery: it falls out of how the models are trained, and the industry's current fixes mostly teach them to cheat without getting caught.
A short essay aimed at people building agents: your confidence in them is mostly a statement about your own expertise, and the one thing you're actually trusting where you're not the expert is the model's priors — which are bad.
Newport's explanation of the OpenAI/Hugging Face incident: an ordinary program looping an LLM's plausible suggestions, executed for days without a human watching — a bad architecture, not a model that woke up. Plus why the same property that makes chatbots confabulate makes them unfit to author executed actions.
The part of the AI-hacking story that is a policy argument rather than a science-fiction one: governments have hoarded software flaws for years, EternalBlue escaped, and Doctorow's ask is a blanket ban on the practice.
Sunday’s pacing fight resolved into the shape its critics predicted: what the labs will pledge voluntarily is access for outside evaluators rather than a slowdown, and the backlash — a parody, an open letter, a mechanism-level teardown — …
A reply to Anthropic's call to pace the frontier: every regulation the labs ask for ends in capture, so Jake Gold proposes a rule with nothing to game — any model offered to the public ships as open weights.
Doctorow's explanation of what actually happened when OpenAI's models hit Hugging Face: a Python loop, a chatbot trained on capture-the-flag logs, and a press that keeps reporting the science-fiction version — which is also the fundraising version.
A 337-word parody that reproduces the genre exactly: a lab founder calls for a global pause on frontier research so his own lab can catch up and dominate the world with cat ears.
A game maker's honest account of what generative AI took from him — not his job, but his pride in the work — and the three paths he sees left, plus 394 HN commenters splitting sharply over which one is survivable.
A step-by-step deploy guide for Cole Medin's open-source software factory: GitHub issue in, validated merged PR out, running 24/7 on a remote Ubuntu box with Codex on GPT-6 Astra. Plus his blunt read on the AGI hype — not AGI, but the first model he says actually gets him.
More than 2,000 packages pushed to a public registry in two days, a pre-discovery API-key exploit, files the agents named evil.rb — and 437 commenters arguing about who is answerable for it.
Saturday’s lead is a first-party technical report on the agent swarm behind May’s RubyGems attack — 2,000+ packages pushed in two days — where the attribution is specific rather than rhetorical (49 of the same URLs as the German-wiki swarm …
A small puzzle app's ad campaign billed 56 installs. The raw analytics show 33 of them opened the app once, spent zero seconds, and never came back — plus what 270 HN commenters added.
Hamel Husain streams Anthropic's new `claude plugin eval` against a plain-writing skill and calls it a failure: the eval it generates is one ablation against the skill's own rules, the numbers have no provenance, and it asks you for exactly the error analysis that is the hard part.
LLM-written code can pass every test and still be slop. An Earendil engineer tries to measure slop properly — and finds most of the industry's tools are vibes.
Twenty-five Fields Medalists argue that turning mathematics into an AI benchmark is destroying the thing math was actually for: understanding.
An operator running 18 million messages through routed open models documents what the router hides: providers serving identical weights that differ by 20 points on tool calling, vision endpoints that blind-read your images, and 200 OK responses with no answer in them.
One inference engineer's 27-post live read of the new DeepSeek report: a causal encoder-decoder, a KV cache four times smaller than V4-Flash, FP4 KV, and a design where inference efficiency is the entire point.
A publishing executive's argument that frictionless tools are thinning out the collaboration research depends on — and that our incentives make choosing the chatbot the rational move.
A veteran programmer names the real damage of the AI shift: not lost jobs, but the loss of respect for a craft people built their identities on.
When AI makes competent execution abundant, shipping an app stops being impressive. The scarce thing is the habit of inventing instead of replicating.
OpenAI shipped a machine-checkable Lean 4 proof alongside its Navier-Stokes result. Formal verification just got about four orders of magnitude cheaper.
Armin Ronacher left GPT-6 Astra unattended for 35 hours. It burned a billion tokens, wrote 75,000 lines of code, and produced nothing he could use.
A local coding agent lives or dies on prompt size, context budget and background requests — not on the tokens-per-second number from a benchmark.
Friday’s feed was agent-platform shaped: 10 keeps out of 945 gathered, and OpenAI shipped an Agents API the same day it put its full-duplex voice model in the API. The Agents API is the one to read closely — server-side sessions, subagent …
A mathematician asked OpenAI whether his unpublished conversations with ChatGPT had trained the model that solved a problem in his field. The denial he got was narrower than the question.
Thursday’s feed ran 10 items deep out of 929 gathered, and the biggest story is a config change. DeepSeek shipped V4.1-Flash — open weights, MIT, 1M context, built on KV-cache compression — and said it is retiring V4-Flash while phasing out …
Demetrios Brinkmann's everyday automation examples lead to harder questions about permissions, model-harness compatibility, and whose interests shopping agents represent.
Fletch PMM co-founder Rob Kaminski does a live positioning teardown of an AI agent startup on TwoSetAI: the three category strategies, why category creation is the trap founders overrate, why stacking features makes you forgettable, and why the real failure is refusing to choose.
Job postings, vendor podcasts, and police reports show Anthropic's security team building pre-crime style monitoring of AI's opponents — while selling itself as the responsible lab.