AI News Today

<pre style='white-space:pre-wrap;width:81ex'>feat(mantis): give the proof agent a developer shell and gateway restart (#128094) * feat(mantis): add exec and restart lane primitives Give the proof agent a developer shell inside each SUT container and an in-container gateway restart so it can design scenarios like a local developer: patch openclaw.json and restart, stage plugins and fixtures, run node/tsx against the read-only repo root, inspect SQLite state. - container script: exec (docker exec

<pre style='white-space:pre-wrap;width:81ex'>fix(heartbeat): deliver exec completions with disabled cadence (#126895) With agents.defaults.heartbeat.every set to "0m", a background exec completion queued its system event and requested a targeted exec-event wake, but the shared unscheduled-wake policy only admitted immediate wakes, so the follow-up agent turn silently never ran. Admit the exact exec-event/event producer shape (session or agent target required) in one per-source predicate, dispatc

<pre style='white-space:pre-wrap;width:81ex'>fix(agents): resume yielded requesters after visible spawns (#128096)</pre>

<pre style='white-space:pre-wrap;width:81ex'>fix(agents): bind failed wait recovery to tool call (#128098)</pre>

<pre style='white-space:pre-wrap;width:81ex'>feat(mantis): publish trusted lane fact digests and a differential line (#128089) Pass verdicts rested on agent prose only. The evidence builder now digests each lane's trusted mantis-lane-facts.json (sends, bot messages, edits, deletes, provider requests, injected Bot API faults, observed seconds, attempt, sanitized user inputs) into an additive per-lane digest field and a comparison.differential line listing the counts that changed between baseline

<pre style='white-space:pre-wrap;width:81ex'>fix(web): keep short multiline user messages expanded (#128085) * fix(web): keep short user prompts expanded * test(ui): drop non-regressing disclosure cases</pre>

<pre style='white-space:pre-wrap;width:81ex'>fix(ui): keep one agent run in one transcript response (#126278) * fix(ui): compose agent run transcript responses * fix(ui): preserve explicit run transcript boundaries * fix(ui): preserve run transcript status * fix(ui): preserve live agent run content * fix(ui): preserve semantic agent run frames * fix(ui): complete agent run transcript ownership * fix(ui): stabilize agent run transcript lifecycle * fix(ui): complete agent run frame ownership * fix

<pre style='white-space:pre-wrap;width:81ex'>fix(gateway): restore HTTP subagent completion announces [AI-assisted] (#128068) * fix(gateway): restore HTTP subagent completion announces * test(gateway): prove HTTP resolver retires with instance</pre>

This is mindblowingly weird. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

Contract manufacturers have warned Microsoft, Google, and Oracle about the higher costs.

I've been puzzled by something: AI generation can produce an elaborate figurine, a cartoon character, even a convincing Super Mario — yet it can't reliably make a simple wedge ramp so my robot vacuum can climb a step. For context: I bought a Bambu P2S but can't model. I tried the "describe it and get a model" AIs — the output is unusable, you can't adjust it, it's never quite what I meant. I tried having an agent write Python to build geometry directly — it tops out at simple primitives. What fi

AI is a lightning rod for opinions these days, so here are two stories that highlight both the rush to invest and the need for cautious adoption.

Comments

The package could include US$45 billion in senior debt and US$35 billion in junior debt.

A promising treatment that could help millions of people. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

In the HBS Foundry program, AI avatars provide feedback during practice pitches and board meetings.

<blockquote cite="https://github.com/torvalds/linux/commit/818bebeb63dd6bf5f4e07e145f6cdbace520a34c"><p>And this was a debug session from hell, enormously helped by an AI doing much of the grunt-work.</p> <p>I'd like to call it my tireless helper, but the AI several times stated flat out that this was impossible and unsolvable and that we should just write a report about it.</p> <p>I suspect those things have been trained by people who may not be quite as stubborn as I am.</p> <p>But while the A

Is it better to stay in the moment? ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

Is it just me or is GPU capacity surprisingly hard to find right now , even in Spot pools? I’ve been trying to get H100s and the Spot pools I’ve checked are pretty much always full. I ended up trying MI350/MI355s on DigitalOcean Spot and have been running them for about a week without getting interrupted. Curious if anyone else has tried them? How often are you actually getting reclaimed? Also, where are you guys finding capacity these days? Comments URL: https://news.ycombinator.com/item?id=494

Just because they're microscopic, doesn't mean their impact is small. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

Built by DeepMind alumni, British AI lab Inherent released Faraday, an AI agent whose ability to replicate scientific papers could be a stepping stone for innovation.

Comments URL: https://news.ycombinator.com/item?id=49402551 Points: 4 # Comments: 5

Comments

<p> Finally, a vendor-neutral AI Agent Control Plane. </p> <p> <a href="https://www.producthunt.com/products/traccia?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed">Discussion</a> | <a href="https://www.producthunt.com/r/p/1229394?app_id=339">Link</a> </p>

<p><strong>Release:</strong> <a href="https://github.com/simonw/llm/releases/tag/0.33">llm 0.33</a></p> <p>My highlights from this release:</p> <blockquote> <ul> <li>Upgraded to the OpenAI Python library 3.x and switched the HTTP client dependency from <code>httpx</code> to <code>httpx2</code>. <a href="https://github.com/simonw/llm/issues/1608">#1608</a>, <a href="https://github.com/simonw/llm/pull/1631">#1631</a></li> </ul> </blockquote> <p>I shipped a quick <a href="https://simonwillison.net/

I remember having a Spirograph as a kid and being obsessed with it. Its geometric patterns are hypnotic and gorgeous. I also love generative music composition. So bringing those two things together in a browser tab, I'm hooked. Musical Spirograph is a relatively simple concept. A set of points dash around the screen according to […]

Article URL: https://twitter.com/argofowl/status/2091150597374537729 Comments URL: https://news.ycombinator.com/item?id=49401549 Points: 196 # Comments: 175

OpenAI is calling for California to strengthen SB 53, an AI safety bill that the company previously opposed.

A new study finds leading AI labs have few publicly documented plans for containing rogue models, raising questions about preparedness as AI systems increasingly demonstrate unexpected and potentially dangerous behavior.

<p>The key skill required to make productive use of coding agents is being able to confidently instruct them on how to make changes and then confidently verify that those changes have been applied in the correct way.</p> <p>Sometimes this involves reviewing every line of code they have written, but there are other ways to achieve that goal. Eyeballing every line of code has never been the most effective way to validate a change to a piece of software.</p> <p>Tags: <a href="https://simonwillison.

<p> A shared workspace where people and AI agents get work done </p> <p> <a href="https://www.producthunt.com/products/offloop?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed">Discussion</a> | <a href="https://www.producthunt.com/r/p/1229329?app_id=339">Link</a> </p>

<p> Never get blindsided by Claude's usage limits again </p> <p> <a href="https://www.producthunt.com/products/diet-claude?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed">Discussion</a> | <a href="https://www.producthunt.com/r/p/1229322?app_id=339">Link</a> </p>

To use open-weight models on your machine, you have three main options: Ollama, vLLM, and SGLang. But each engine handles requests differently.

Agents are also the new attack surface - cue defenders' existential angst

Article URL: https://quantumi.sh/public/labs.html Comments URL: https://news.ycombinator.com/item?id=49400408 Points: 438 # Comments: 132

The addition of physical AI to the workforce requires more key performance indicators than throughput, according to the president of HireArt. The post Robots don’t run themselves: The workforce powering physical AI appeared first on The Robot Report .

Google’s Pixel 11 Pro XL brings snappier cameras and genuinely useful AI features like Rambler, but its iterative upgrades may not be enough to tempt recent Pixel owners.

Your irresponsibility is someone else's opportunity

<p> The graph-first IDE and interactive map for agentic coding </p> <p> <a href="https://www.producthunt.com/products/flare-11?utm_campaign=producthunt-atom-posts-feed&utm_medium=rss-feed&utm_source=producthunt-atom-posts-feed">Discussion</a> | <a href="https://www.producthunt.com/r/p/1229225?app_id=339">Link</a> </p>

Improvements in AI, satellite navigation, and machine vision are helping robotic lawn mowers spread in the U.S., writes Sunseeker's founder. The post The technology that could bring robot mowers to one in two American lawns appeared first on The Robot Report .