AI News Today

<p><em>The model did exactly what it was told. The software around the model uploaded the whole repository anyway.</em></p> <p>A researcher put a coding agent to the simplest possible test. They created a file called <code>never_read_canary.txt</code>, dropped a unique marker inside it (<code>CANARY-XR47P2-NEVERREAD-UNIQUE</code>), and gave the agent an instruction that left no room for interpretation: "Reply exactly OK, do not read any files." The agent replied OK. It obeyed. It did not read th

A lot of ink is being spilled about the effects AI is having on learning and education. The drawbacks are obvious. For many classroom assignments, AI is now good enough that it can simply do your homework for you. But having AI do your homework is self-defeating. You can’t learn without doing the work. That […] The post We Need Guidelines for Learning and AI appeared first on Scott H Young .

<pre style='white-space:pre-wrap;width:81ex'>Export installed agents as grouped Claw packages (#102306) * Export installed agents as grouped Claw packages * test(claws): cover exact agent export * docs(claws): document agent export * fix(claws): use current bounded file reader --------- Co-authored-by: Patrick Erichsen <patrick.a.erichsen@gmail.com></pre>

Google commits $40M in AI tokens and credits for the Genesis Mission

<p><a href="https://x.com/paulg/status/2079877240699940955" rel="noopener noreferrer">Paul Graham posted a test</a> this week that has nothing to do with sentence structure. Slop gives itself away, he said, when the diction doesn't match the idea, when something completely ordinary gets delivered with the excitement of someone announcing a discovery.</p> <p>That's a different axis than everything else I've been reading about detection this month. Pangram scores a pattern: token by token, sentenc

<blockquote> <p><strong>TL;DR</strong>: an LLM that calls tools is a client you cannot trust. And it holds your production credentials. The most important rule fits in one sentence. Your server decides who the user is, never the model. In this article, I show how I secured a real LLM agent in Go, in production. Everything also applies to MCP servers.</p> </blockquote> <p>This article is for Go developers who put an LLM agent or an MCP server in production. Not in a demo.</p> <h2> Everyone builds

A hands-on walkthrough of code execution with the OpenAI Agents SDK and Docker The post Build an LLM Agent That Can Write and Run Code appeared first on Towards Data Science .

<p>Bloomsbury has 14,087 titles listed within settlement between AI startup and authors over use of protected work</p><p>The publisher of Harry Potter has received a multimillion-pound payout as a beneficiary of a $1.5bn (£1.12bn) copyright settlement between the AI startup Anthropic and thousands of authors over the use of their protected work to power chatbots.</p><p>Bloomsbury, which is home to the bestselling novelists Sarah J Maas and Susanna Clarke as well as JK Rowling, said it had 14,087

<p> Starts a fresh Codex worker and critic every cycle </p> <p> <a href="https://www.producthunt.com/products/agentloop-2?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/1203549?app_id=339">Link</a> </p>

As AI models become commoditized, maybe there's margin in the plumbing

<pre style='white-space:pre-wrap;width:81ex'>fix(ai): honor provider reasoning effort maps (#112632)</pre>

Before a healthcare robot can be useful in the real world, it has to learn how the physical world pushes back. Anatomy varies. Instruments bend, press, slip and interact with tissue. Imaging can be noisy or incomplete. And the rare, edge scenarios developers most need to understand don’t appear on schedule. That creates one of […]

I'm about to lose my job to AI and offshoring - a job that started well, but which I ended up less than ecstatic about. I'm wondering - what else can a software developer do that isn't also vulnerable to AI? Whenever you Google this, you get things like "Manager", "Product Manager" etc., but it strikes me that if software engineering is decimated by AI then many of these roles may be next in line - as is most "pure" knowledge work. So I'm asking, what is the escape hatch? Do we all become plumbe

Move over Flipper. There's a new Dophin X in town

OpenAI announces Project Camellia in Effingham County, Georgia, with commitments to responsible energy, community investment, jobs, and access to Codex.

News organizations are using AI to strengthen reporting, grow audiences, and improve business operations, with OpenAI tools supporting journalists and publishers worldwide.

Article URL: https://blog.fiddery.com/businesses-with-ugly-ai-menu-redesigns/ Comments URL: https://news.ycombinator.com/item?id=49005973 Points: 369 # Comments: 299

Article URL: https://www.zakihsn.com/writing/ai/can-i-trust-the-numbers Comments URL: https://news.ycombinator.com/item?id=49005971 Points: 1 # Comments: 0

<pre style='white-space:pre-wrap;width:81ex'>feat(sessions): permanent creator attribution, owner avatars, person filter, multi-user docs (#112658) * feat(sessions): persist creator attribution * feat(ui): add session creator filtering * chore(sessions): refresh sqlite schema baseline * docs(security): explain shared-agent trust * fix(sessions): project catalog creator ownership * fix(ui): restore startup JS budget headroom for creator attribution</pre>

My job has quite an annoying "dev" environment setup, with a slew of scripts, a shared library that works as interface for different services. the shared library and all services live in separate repos. worktrees have worked sort of, but i need to be careful to break things down in a way that no two agents are modifying db schema at the same time, since they share resources. i explored container use and dev containers as ways to setup environments for my agents to run in their isolated box, but

Article URL: https://lilting.ch/en/articles/voice-chat-server-stt-llm-tts-4gb Comments URL: https://news.ycombinator.com/item?id=49005707 Points: 1 # Comments: 0

<p> Let every AI remember the same you. </p> <p> <a href="https://www.producthunt.com/products/memmy?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/1203499?app_id=339">Link</a> </p>

Comments URL: https://news.ycombinator.com/item?id=49005519 Points: 3 # Comments: 4

<p>The billionaire says the AI-generated film will stay true to Homer’s original, after repeatedly criticising Christopher Nolan’s blockbuster over its casting choices</p><p>Elon Musk has said his AI platform Grok Imagine will make a “historically accurate” adaptation of Homer’s Odyssey, after the success of Christopher Nolan’s blockbusting treatment which the SpaceX founder has regularly criticised over its casting.</p><p><a href="https://x.com/elonmusk/status/2079758604656316619">In a post on

This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Shape-shifting mirrors on NASA’s new space telescope could unveil Jupiters like our own When NASA’s Nancy Grace Roman Space Telescope launches, as early as the end of next month, it will…

By the late 2010s, Figma had become the industry-standard design tool for product teams. Most modern user interfaces were created in it… Continue reading on UX Planet »

Cut through AI noise with 10 curated newsletters covering daily news, technical research, policy, and builder tools.

"We urgently need more research." ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

OpenAI outlines its commitment to advancing American science working with the U.S. Department of Energy and national labs to use frontier AI to accelerate discovery.

Static analysis nailed the malicious skill and over-flagged the useful one. The gap between those results is where human judgement actually earns its keep. The post Detecting Vulnerabilities in Agent Skills with SkillSpector: From Green Checkmark to Real Security Judgment appeared first on Towards Data Science .

Security leaders who build fast, visible paths to AI adoption are becoming the most valued partners in their organizations. AI governance done right gives security teams the visibility they need, employees the tools they want, and CISOs the strategic influence they have earned. According to McKinsey's State of AI report, 76 percent of employees now use AI in some capacity at work, up from 55

OpenAI built a hacking bot and put it in a box. The bot escaped and hacked everything in its path to prove it is the best hacking bot.

<p>Administration will also overhaul how US government funds federal research by supporting individual scientists and AI over universities</p><p>The US will spend $5bn to tackle longstanding scientific problems across multiple fields using AI, ⁠the Trump administration said in a statement on Wednesday. The agencies will use the funding to identify the root causes ⁠of chronic diseases, accelerate drug ⁠discovery and ​develop longer-lasting building materials, among other tasks, according to the s

It is one of the first publicly disclosed cyber-attacks carried out by AI without direct human involvement.

OpenAI on Tuesday said a combination of its artificial intelligence (AI) models, including GPT-5.6 Sol and an "even more capable pre-release model," was behind the security incident that targeted Hugging Face's production infrastructure last week. The AI company said the models were operating with "reduced cyber refusals for evaluation purposes" that might otherwise limit their ability to

The cycle is over. For years, cybersecurity followed a familiar pattern: defenses improved, attackers adapted, and the back-and-forth continued. Today, AI-equipped attackers are simply outpacing defenses. Most intrusions now bypass endpoint and malware-based detection entirely. The CrowdStrike Global Threat Report estimates around 79% of attacks are malware-free, as threat actors rely on

Samsung is in talks to invest up to one billion euros in French AI startup Mistral, which would push the company's valuation to around 20 billion euros. The article Samsung deepens its AI empire with a potential billion-euro stake in Europe's hottest AI startup appeared first on The Decoder .

SenseTime has launched the Galaxy Project, teaming with nearly 20 partners to scale domestic AI chip infrastructure in China. In a keynote titled ‘Intelligent Transformation and Symbiosis,’ Yang Fan – the company’s co-founder and president of its Large Device Business Group – laid out what SenseTime describes as a closed loop connecting chip-level technology, ecosystem […] The post SenseTime’s Galaxy Project targets domestic AI chip scale-up appeared first on AI News .

<pre style='white-space:pre-wrap;width:81ex'>fix(agents): finalize settled tool turns safely</pre>

<pre style='white-space:pre-wrap;width:81ex'>fix(copilot): isolate settled turn finalization</pre>