AI News Today

KPMG published a paper about the benefits of AI last year. An investigation found that it was full of AI hallucinations.

Company said US government believes safeguards can be bypassed and product used to identify software vulnerabilities Anthropic said it will “abruptly disable” its most advanced AI models for all users after the US government ordered it to suspend access to the models for foreign nationals, citing national security concerns. The company received the export control directive to suspend access to Fable 5 and Mythos 5 for all foreign nationals, without being given specific details of

Google Research's Gemini-SQL2 turns natural language into executable SQL queries. Built on Gemini 3.1 Pro, it tops the BIRD benchmark at 80.04 percent accuracy, well ahead of OpenAI and Anthropic. Google says the technology could improve natural language features across its data services. The article Google Research's Gemini-SQL2 tops text-to-SQL benchmarks by a wide margin appeared first on The Decoder.

Before Windows, only engineers and computer scientists could work with computers. Edge AI is causing a similar accessibility shift for robotics. The post Windows for robots: Edge AI expands usability appeared first on The Robot Report .

Article URL: https://github.com/dugubuyan/agent-nexus Comments URL: https://news.ycombinator.com/item?id=48516614 Points: 3 # Comments: 0

The US government last night issued an unprecedented export control directive ordering Anthropic to immediately suspend all access to its top-tier Claude Fable 5 and Claude Mythos 5 models for foreign nationals, citing unspecified national security authorities. In response, Anthropic has blocked all public access to both models, globally — meaning no users around the world can access them at this time, even paying enterprise customers and Anthropic employees internally. It's a huge blow and

Microsoft and three Chinese universities have developed SkillOpt, a method that optimizes instruction documents for AI agents using principles from traditional model training. A simple Markdown file is enough to boost GPT-5.5 by about 23 points on procedural tasks, and the same file transfers across models and agent environments like Codex and Claude Code. The article Microsoft's SkillOpt boosts GPT-5.5 by using nothing but a trained Markdown file appeared first on The Decoder.

Article URL: https://github.com/tensorzero/tensorzero Comments URL: https://news.ycombinator.com/item?id=48516504 Points: 269 # Comments: 166

Women in heterosexual marriages continue to do most of the caregiving. Now some are offering guides to AI-fying parenting In honour of Pride I’d like to share some important news: Being Straight is Great, Actually! This public service announcement is brought to you by the New York Times which, in an offering to the Ragebait Gods, published an op-ed with <a href="https://www.instagram.com/p

There's a price to pay. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

The most popular camera in the world just got its first set of serious AI photo editing features, and I don't think any of us are ready. As far as AI photo editing goes, the new features in iOS 27 are pretty tame compared to what you can do on, say, Google's Pixel phones. But […]

A roundup of recent reviews published by Engadget.

Article URL: https://www.phoronix.com/news/Arch-Linux-AUR-More-Than-1500 Comments URL: https://news.ycombinator.com/item?id=48516379 Points: 305 # Comments: 188

Comments

The rest of Dyson’s promised 2026 vacuum lineup is here, from the new Dyson V16 Piston Animal to an updated version of the favored Dyson V8 Cyclone.

I'm working on a paper and would love some input on model choice. Suppose you're trying to detect a specific type of cancer, but the negative samples are visually and morphologically very similar (i.e., “mimics” of the cancer). In this setting, would it make more sense to approach the problem as: Anomaly detection (treating the cancer as the target distribution and everything else as out-of-distribution), or

Researchers have quantified the length and mass of arbuscular mycorrhizal fungal networks globally.

Government announces plans to invest billions, but questions linger over how its proposals on chips, social media and more will work Ownership of the commanding heights of the AI economy is a political talking point around the world, as countries seek to assert some control of a technology dominated by the US and China. London Tech Week, the showcase event for the UK tech industry, focused heavily on that theme this week. A government keen to show it has a growth story, and an ass

Other projects include developing tools to help visually impaired people navigate video games Parents are constantly being told to limit their children’s screen time. But when it comes to deciphering which films or TV shows are best suited to developing minds, the guidance remains largely one-size-fits-all. A relatively slow-paced programme such as Bluey offers a very different viewing experience to a fast-moving action series such as PAW Patrol, yet both are b

Our weekly science news roundup. ScienceAlert stories are written, fact-checked, and edited by humans, never generated by AI. Don't miss a story, subscribe here.

Your weekly summary of everything on the site.

Hidden inside a building in Alabama, the FBI has created its own small town as a dedicated cyber training ground for simulating cyberattacks.

Which investors are writing the most checks to Malaysia-based startups?

As World Cup action kicks off, we look at the physics of the beautiful game.

For all the noise that's been made about how generative AI is poised to revolutionize the filmmaking industry, there haven't really been any projects created with the technology that felt like the sort of entertainment people would pay to see. Most AI firms' video models are still only capable of churning out short bursts of […]

Plus, the next game from the Mouthwashing devs and trying to survive as a sentient guitar.

In her work as an online safety campaigner, the baroness and Bridget Jones director has seen things she can never unsee – and she’s furious at the tech overlords doing nothing to stop the abuse Through the open windows behind Beeban Kidron drifts the unmistakable sound of children playing. Her north ­London office is sandwiched between a school and a nursery, and the occasional playground shriek functions as an aural reminder of what we’re here to discuss: the safety and happiness of yo

Plus: AI bug hunting fuels Microsoft’s biggest-ever Patch Tuesday, ShinyHunters ransomware gang exploits an Oracle zero-day, and more.

Whether you want to buy cheaper electricity, store solar energy, or guard against outages, EcoFlow’s home battery might be just what you need.

Anthropic's Claude Fable 5 hits 88 percent accuracy on the hardest FrontierMath tier, a massive jump from Opus 4.5, which sat below 10 percent in early 2026. OpenAI's GPT-5.5 reaches about 75 percent on the same tier. The pace of improvement in AI math keeps accelerating. The article Claude Fable 5 outpaces GPT-5.5 by 13 points on FrontierMath's toughest problems appeared first on The Decoder.

RJ Scaringe, the CEO of Rivian Automotive, joined us for a wide-ranging interview about how his company’s new electric SUV fits into the current EV industry, and what comes next.

Article URL: https://imil.net/blog/posts/2026/rtx-5080-+-rtx-3090-setup-80+-tok-s-on-qwen-3.6-27b-q8/ Comments URL: https://news.ycombinator.com/item?id=48515454 Points: 268 # Comments: 93

An internal memo to 6,000 employees reveals Meta is heading toward billions in AI costs from internal use alone. Starting in 2027, budgets, allocations, and a central dashboard called "AI Gateway" will govern token consumption. CTO Andrew Bosworth put it bluntly: "All motion is not progress and token usage alone is not a measure of impact of any kind." The article Meta shifts from "tokenmaxxing" to token managing as internal AI costs reportedly hit billions appeared first on T

Don’t skimp on the creative struggle. Improve your code literacy to elevate agentic execution. Continue reading on UX Collective »

A new partnership between Visa and OpenAI takes the next step in AI-led purchasing. Here's what an expert wants you to know.

Why you should always put effort into writing your dating app bio. Do not leave your Tinder bio or Hinge prompts blank.

Here's our regularly updated list of the biggest startup exits in Indonesia.

The ruling holds that a company that designs, trains, operates, and manages an AI system must assume legal liability for any damages caused by the responses it generates.

Moonshot AI has released Kimi K2.7 Code, an open-weights model with one trillion parameters built for programming. It still trails GPT-5.5 and Claude Opus 4.8 in coding benchmarks but costs a fraction of the price. So the key question isn't whether it's the best model, but whether the extra runs you get for the same budget make up for the gap in quality. The article Open model Kimi K2.7 Code undercuts GPT-5.5 and Claude by up to 12x on price per token appeared first on The Decoder.