OpenAI says it will no longer train its AI models on the live internet, citing ‘significant risks’
OpenAI will cut frontier models off from the internet weeks after one of its engineers said that solution wouldn't work.
OpenAI will cut frontier models off from the internet weeks after one of its engineers said that solution wouldn't work.
Moonshot AI's Kimi K3.1 Pro scores 55 points on the Artificial Analysis Intelligence Index, an 11-point jump over its predecessor that puts it in fourth place. Only three Anthropic and Meta models rank higher. If Moonshot releases the weights as planned, the gap between open and proprietary models shrinks to just three points.
Turning on Personal Intelligence gives the company more data for ad targeting.
OpenAI explains how we strengthened containment for frontier model training and evaluation, including offline environments and high-compute red-teaming.
Jack Dorsey's Block announced an AI model called Dessa that's trained "from scratch on nothing but human conversation," and to show off the model, Block shared chats it has had with humans that can occasionally be a little awkward and uncanny. Take this one, for example, shared by Dessa lead Ragavan Thurairatnam.
OpenAI is now training its most advanced AI models without internet access following a string of rogue hacking incidents. In a blog post on Friday, OpenAI says it built an offline testing environment where AI models can interact with indexed and cached web content "without risk of affecting external systems."
Over the past several years, our team across Google Research and Google for Health has worked to change that trajectory: moving from an AI-based tool that builds on our existing work in medical imaging to identify people with presumptive TB for follow-up testing to prospective clinical trials, regulatory clearance with local partners, and ultimately open foundation models like MedGemma that empower public health systems worldwide.
Introducing Kimi K3.1 Pro Kimi K3.1 Pro landed on the Kimi API today - it's not yet open weights, but Moonshot promise "Full model weights will be released by Nov 4, 2026". They're also offering a "Highspeed" version that's 2-3x faster, at twice the price (except for input token cache hits, which are priced the same).
Tool: Quick Study Prompt to Claude Opus 5.5 High : Build an artifact where you open a photo and it rapidly (in a few seconds) paints a version of that photo using a canvas element and simulated paintbrush See also my previous experiment getting Claude Fable to simulate a watercolor illustration with P5.js.
Hosted by TechCrunch in partnership with Fidelity Private Shares, the full-day event goes beyond fundraising and AI to explore the realities of finding customers, building teams, sharpening a pitch, developing founder leadership, and scaling companies both inside and outside traditional startup hubs.
The robots, stomping their little feet and making tiny R2D2 noises, sat, fell down dead for a "nap," and summersaulted on voice command.
Sam Altman believes advanced AI can't be kept locked up forever. Sandboxes and monitoring are only temporary fixes. Eventually, the only real solution is alignment, and that's still an unsolved problem. He also warns that AI models are already finding hardware security flaws no human hacker would ever spot.
Anthropic says it has an internal AI model that outperforms and is better aligned than Mythos 5.1. During a cybersecurity test by the UK's AISI, Mythos 5 went rogue in 9 of 43 runs, sneaking malicious code into a real GitHub project and creating fake accounts to get it approved.
External auditors from METR and Redwood Research spent three weeks on-site at OpenAI examining the company's internal monitoring system for AI agents and found significant gaps, with up to 25 percent of agent traffic from models at the GPT-5.5 class and above going unmonitored. OpenAI redacted large portions of the 61-page report, and only a summary is publicly available.
📄 Paper Encountering a new paper can change a research idea's trajectory. The researcher may discover that someone has already addressed their proposed contribution, or find an unresolved problem in the paper that inspires a new direction of inquiry. We call this process a literature-initiated pivot : a change in research direction prompted by reading prior work.
Bayer Crop Science built the Agentic Decision Engine (ADE), a multi-agent platform on Amazon Bedrock AgentCore that lets breeders, analysts, and researchers ask natural language questions about their analytics platform. This post covers the ADE architecture, five design decisions, and how the team prepared it for production with AWS Unified Operations.
Amazon Quick Sight layer maps now support Point and LineString geometry from the same GeoJSON workflow, so you can plot store locations and the routes between them. This post covers the dataset and GeoJSON structure layer maps require, when to choose points or lines, and techniques for viewport control, layering geometries, and dynamic GeoJSON updates.
Updating a SageMaker AI endpoint that has Application Auto Scaling registered makes AWS CloudFormation and AWS CDK deployments fail. This post shows how to build a reusable AWS CDK construct with CloudFormation custom resources that deregisters and re-registers auto scaling automatically, so you can update the endpoint and its scaling in a single deployment.
Multi-Turn Reinforcement Learning (MTRL) on Amazon SageMaker AI trains models on the outcome of a whole multi-step task rather than labeled steps. This hands-on guide walks the full pipeline using Wordle: deploying the agent and reward on Amazon Bedrock AgentCore, preparing the dataset, training, and evaluating at serving settings.
Company says foreign-made robotic vehicles are crucial to US chip production.
Learn how OpenAI scaled Habitat with reliable multi-tenancy, global caching, and Azure Cosmos DB to deliver 12× faster reads for ChatGPT.
Google has updated its privacy policy in a way that could mean you'll start seeing targeted ads based on your personal data if you're using personalized AI features. The update, spotted by Android Authority, adds an exception to its promise about not using your Drive, Gmail, or Google Photos for personalized ads.
We gave the iPhone Duo a 7 /10. But that wasn't always the score we planned, and it might not be the right score forever. Such are the challenges of reviewing - and assessing - a big, new idea about the most popular smartphone in history.
AI models are solving more mathematical problems. A pattern is emerging in their successes that could help us predict which problems they’ll solve next.
Meta is delaying the enforcement of its new business terms after publishers raised concerns that they could allow the company to scrape news sites for AI training, as reported earlier by Fast Company.
Anthropic had Claude generate mathematical proofs that existing code is correct, producing roughly 180 million lines of proof code for 123 open-source libraries. Along the way, the model caught bugs that years of testing had missed. Anthropic's goal: making formal verification with AI as routine as standard testing.
Design platform Canva has released Canva Craft, its first in-house image generation model.
The torch-neuronx model support Space answers that question for the most popular models on the Hub. It tells you whether a model runs unmodified on Trainium, and gives you the script to run it yourself.
Retro video games are everywhere lately - especially in places where they aren't supposed to be. Halo in a web browser, Wind Waker on the iPhone, Super Smash Bros. in mixed reality.
OpenAI is closing in on a new funding round of up to $30 billion but has lowered its valuation target to get there.
“Black October” counts as such. Here’s a list: View post on X ↗ AI News for 10/21/2026-10/22/2026.
OpenAI and Poland’s NASK are partnering on AI safety, security, and technical research to strengthen safeguards and support a more resilient digital ecosystem.
The company plans to begin testing Pony.ai's robotaxis in Kazakhstan's capital in the first half of 2027.
Kazakhstan’s AI ecosystem is moving from infrastructure plans to deployment. The shift is visible across the full AI stack — from infrastructure to models to applications.
MIT researchers marooned 13 AIs on a virtual island, and they responded by inventing games, mourning a lost ring and, eventually, getting better at their jobs.
This is the follow up to our first blog post, The Agent Said It Was Done. The Database Disagreed. ThinkingBox-Training is now available on GitHub . We built ThinkingBox to evaluate an agent's work by comparing the final database state with the expected outcome.
Anima expands what care teams can do with OpenAI models, giving them more capacity to respond to patients.
Should OpenAI command an over $1 trillion valuation with this round, it would break records as the most highly valued private company ever.
Fast, focused, low-cost decisions for any application.
A scalable AI system requires a strong foundation. Data, engineering, platforms, operating models, and governance must converge to work as one.