AI News Today

<p><strong><a href="https://martinalderson.com/posts/huggingface-openai-exploit/">The first known runaway AI agent - or a very bad marketing stunt?</a></strong></p> Martin Alderson's commentary on the <a href="https://simonwillison.net/2026/Jul/22/openai-cyberattack/">OpenAI accidental cyberattack against Hugging Face</a> includes a couple of details I hadn't considered.</p> <p>First, Hugging Face offers a truly rich target if you're trying to find potential vulnerabilities that require executin

Article URL: https://resobscura.substack.com/p/llm-based-educational-games-will Comments URL: https://news.ycombinator.com/item?id=49027196 Points: 1 # Comments: 0

Article URL: https://github.com/arizqi/cpubrrr Comments URL: https://news.ycombinator.com/item?id=49026595 Points: 1 # Comments: 0

Article URL: https://seldon-ai.com/blog/you-dont-need-an-llm-to-cluster-llm-traces Comments URL: https://news.ycombinator.com/item?id=49025349 Points: 2 # Comments: 0

Article URL: https://netflixtechblog.com/in-house-llm-serving-at-netflix-a5a8e799ea2c Comments URL: https://news.ycombinator.com/item?id=49025021 Points: 3 # Comments: 0

I love learning new concepts by playing around with them (think explorable explanations ( https://en.wikipedia.org/wiki/Explorable_explanation ), and I got more and more interested in running local LLMs on my own hardware. Since we live in the age of "speak, and your side project will be done" (joking, but LLMs are really helpfull for me when it comes to frontend stuff and for brainstorming in general!), I built the tool I needed myself :) WatchMachineGo is basically a visual simulator that simu

The excitement surrounding PrismML’s 1-bit/ternary Bonsai models has the industry closely watching how smartphone giants, particularly Apple, will implement LLMs on edge devices. Moving AI on-device is a brilliant and necessary strategy. It ensures absolute user privacy in alignment with EU regulations, fundamentally shifts the economics away from costly cloud inference, and paves the way for a significant hardware upgrade supercycle as users seek true AI-capable silicon. To create a smart on-de

Article URL: https://github.com/war851/Governed-Agent Comments URL: https://news.ycombinator.com/item?id=49020159 Points: 1 # Comments: 0

Article URL: https://github.com/akon47/NvChat Comments URL: https://news.ycombinator.com/item?id=49019051 Points: 1 # Comments: 0

Article URL: https://twitter.com/i/status/2080180914135667085 Comments URL: https://news.ycombinator.com/item?id=49018858 Points: 2 # Comments: 1

Article URL: https://ktransformers.net/en Comments URL: https://news.ycombinator.com/item?id=49018742 Points: 1 # Comments: 0

Article URL: https://medium.com/@v_regis/why-self-auditing-your-ai-model-violates-eu-ai-act-article-53-22f693fb64a7 Comments URL: https://news.ycombinator.com/item?id=49018523 Points: 1 # Comments: 0

Article URL: https://github.com/bossandboss/Ai-Firewall Comments URL: https://news.ycombinator.com/item?id=49018165 Points: 1 # Comments: 0

Article URL: https://github.com/MasihMoafi/Elpis Comments URL: https://news.ycombinator.com/item?id=49017977 Points: 1 # Comments: 0

Article URL: https://andytimm.github.io/posts/cs336/cs336_review.html Comments URL: https://news.ycombinator.com/item?id=49017816 Points: 1 # Comments: 0

Article URL: https://github.com/PxyUp/fitter Comments URL: https://news.ycombinator.com/item?id=49015561 Points: 1 # Comments: 0

Article URL: https://www.localmaxxing.com/en Comments URL: https://news.ycombinator.com/item?id=49015424 Points: 1 # Comments: 2

Health in ChatGPT now lets eligible U.S. users securely connect medical records and Apple Health to get more personalized insights and better understand their health.

<blockquote cite="https://twitter.com/tqbf/status/2080045032162173329"><p>I genuinely believe that if you took an open weights model from 2025 and built a pentest harness for it, it could do this kind of sandbox escape and scan/hack in most networks. This is only surprising because you assume OpenAI has sounder sandboxes.</p></blockquote> <p class="cite">— <a href="https://twitter.com/tqbf/status/2080045032162173329">Thomas Ptacek</a>, doesn't think <a href="https://simonwillison.net/2026/Jul/22

<p>This story is wild. The short version: OpenAI were running a cybersecurity test against an unreleased model, with the model's guardrail features turned off. Rather than solve the test, the model broke its way out of OpenAI's sandbox, then found exploits to break <em>in</em> to Hugging Face, all so it could cheat on the test by stealing the answers.</p> <p>Along the way it helped make the strongest case yet for how the imbalance of model availability is hurting our ability to secure our softwa

<p><strong><a href="https://dylancastillo.co/posts/pelicanmaxxing.html">Are AI labs pelicanmaxxing?</a></strong></p> Excellent piece of work by Dylan Castillo, who took a deep-dive into the frequently pondered question of whether the AI labs have been deliberately training models to draw pelicans riding bicycles in response to my <a href="https://simonwillison.net/tags/pelican-riding-a-bicycle/">deeply unscientific benchmark</a>.</p> <p>I've been randomly spot-checking this in the past by testin

Article URL: https://www.roleframe.ai/blog/why-ai-resume-sounds-generic Comments URL: https://news.ycombinator.com/item?id=49013985 Points: 2 # Comments: 0

Article URL: https://twitter.com/joefioti/status/2077430481809989662 Comments URL: https://news.ycombinator.com/item?id=49013890 Points: 1 # Comments: 0

Article URL: https://github.com/ykushch/agsig Comments URL: https://news.ycombinator.com/item?id=49013862 Points: 1 # Comments: 0

Article URL: https://www.rnz.co.nz/news/business/773456/ai-is-changing-the-way-consumers-shop-and-businesses-need-to-get-ready Comments URL: https://news.ycombinator.com/item?id=49013796 Points: 4 # Comments: 1

Build a website for AI pics for dating apps in 20 hours, launched ads, failed I had some free google cloud credits (included nano banana for AI gen photos) Build a simple website - user uploads a selfie, receiving cool pics for dating apps with his face + he can edit photos with promts. 3 photos with watermarks for free, after that pack with 25 photos for 9,99$, In the backend its prefiltered cool photos + face recognition in order to return only photos with a good simularity. End result - $202

Article URL: https://www.reuters.com/business/google-quarterly-cloud-revenue-growth-beats-expectations-2026-07-22/ Comments URL: https://news.ycombinator.com/item?id=49013423 Points: 2 # Comments: 2

Article URL: https://github.com/PJHkorea/Continuous_Wave_Field_LLM_Brain Comments URL: https://news.ycombinator.com/item?id=49013403 Points: 2 # Comments: 1

Article URL: https://manish.sh/writings/ai-tools/inside-character-ai-the-technical-story-of-what-keeps-users-hooked Comments URL: https://news.ycombinator.com/item?id=49013296 Points: 1 # Comments: 0

Article URL: https://github.com/forgesynapseltd/ClawLite Comments URL: https://news.ycombinator.com/item?id=49013033 Points: 1 # Comments: 0

Article URL: https://www.techdirt.com/2026/07/22/doj-now-citing-fake-ai-generated-cases-to-keep-ice-detainees-locked-up/ Comments URL: https://news.ycombinator.com/item?id=49013031 Points: 12 # Comments: 0

Article URL: https://github.com/Role1776/netmon Comments URL: https://news.ycombinator.com/item?id=49012930 Points: 1 # Comments: 0

Article URL: https://www.ft.com/content/b02f972c-c764-4006-9377-42563d9d5530 Comments URL: https://news.ycombinator.com/item?id=49012855 Points: 8 # Comments: 5

Article URL: https://vester.si/ai-impact/ Comments URL: https://news.ycombinator.com/item?id=49012843 Points: 4 # Comments: 2

Article URL: https://www.wsj.com/cio-journal/openais-high-stakes-high-touch-push-to-make-ai-work-for-business-7f08b7d4 Comments URL: https://news.ycombinator.com/item?id=49012822 Points: 1 # Comments: 0

Article URL: https://loadbearingtech.substack.com/p/self-improving-agent-loops-verifier Comments URL: https://news.ycombinator.com/item?id=49012760 Points: 3 # Comments: 0

Article URL: https://llmgateway.io/blog/openrouter-vs-vercel-vs-llmgateway-performance Comments URL: https://news.ycombinator.com/item?id=49012741 Points: 3 # Comments: 0

Article URL: https://www.ft.com/content/3fcda833-80d2-4e13-86f4-29b479b79adf Comments URL: https://news.ycombinator.com/item?id=49012685 Points: 6 # Comments: 0

Article URL: https://github.com/lidge-jun/opencodex Comments URL: https://news.ycombinator.com/item?id=49012330 Points: 6 # Comments: 1

I built this as i thought this model could better to what other tools are doing. What do you guys think ? Harbinger is an mTLS proxy that sits in front of an AI agent's outbound traffic. Every AI agent, bot, and service account gets a cryptographic identity. Every request it makes is checked against policy before it's allowed, and any real secret it needs is pulled from your own vault and swapped in only when the request is verified to be going where it claims. The agent itself never holds a sta