How we detect frontier model misalignment
OpenAI uses layered monitoring to detect frontier model misalignment and unauthorized agent behavior so responders can intervene before incidents escalate.
OpenAI uses layered monitoring to detect frontier model misalignment and unauthorized agent behavior so responders can intervene before incidents escalate.
OpenAI’s updated Model Spec outlines five core values to guide more capable AI toward sound judgment, human agency, and faithfulness to human intent.
NVIDIA Nemotron 3.5 Super VL is a 120B-parameter mixture-of-experts model that pairs reasoning and tool use with multimodal understanding. This post explores its architecture and use cases, then shows how to deploy it on Amazon EKS, EKS Auto Mode, and Amazon SageMaker HyperPod EKS with NVIDIA Dynamo and vLLM.
MoEngage turned roughly 400 private REST APIs across 7 AWS Regions into permission-scoped agent tools with Amazon Bedrock AgentCore Gateway. No per-service MCP servers, no public endpoints, and no 400-tool context bloat. See how OpenAPI specs, Cedar policies, and managed Amazon VPC Lattice egress made it work.
AWS Batch quota management is a new scheduling capability you can use to divide GPU capacity into quota shares per team or project, automatically lend idle capacity across shares, and use preemption to run your highest-priority SageMaker Training jobs first, all without manual coordination.
Turn a PHP-to-Java migration from a one-time leap of faith into a measurable, repeatable process with Kiro. This post walks through a reusable engine: a mechanical discovery scanner and an eight-phase, parity-first workflow that treats the original PHP application as the specification and verifies the Java port with a parity harness.
Epsilon's Agentic Platform team delivers UI testing as a managed, multi-tenant service built on Amazon Nova Act and Amazon Bedrock AgentCore Browser Tool. Tests are written in plain English and run on browsers managed by AWS, giving every product team resilient, intent-driven testing with zero per-team infrastructure overhead.
To anyone listening closely to what Microsoft CEO Satya Nadella has been saying over the past couple of years, this week felt like the clearest sign yet that AI is about to fundamentally change the size of Microsoft's workforce.
How I Built the Most AI-Pilled Company in the World Don't be put off by the clickbait headline, this piece by Jacob Sansbury describing how they are building Spawn is a fascinating exploration of the bleeding edge of what's possible with agentic engineering against frontier models.
The update allows you to use Alexa to create lists, register them, see what's on them, and add or edit items. You can also manage subscriptions.
Dream is grounded in real physics and construction standards. Just as LLMs learn from text, Dream's world model is trained on more than 100 million square feet of building data gathered from permit archives across the United States.
A top executive of a Singapore AI cloud company that was accused of illegally selling Nvidia chips to China could also have been a Chinese spy, sources have told the New York Times.
Mistral says its new model can afford organizations a kind of control over its cyber defenses that closed models can't.
Uno reverse! After Wispr added meeting notes to its dictation app, Granola is adding dictation to its meeting notetaker. The voice AI wars are getting interesting.
The US Justice Department is looking into Nvidia's planned acquisition of open-source platform Hugging Face.
Mistral has released the weights for Large 4, aka "Le Chonk," a trillion-parameter model trained entirely in Europe. Mistral says it beats its open rivals at coding, and independent testing shows it's fast and fairly cheap. It still can't match China's top models, scoring 26 points to the 32 posted by DeepSeek V4.1 Flash and GLM-5.3.
Mistral Large 4 (aka Le Chonk) is a multimodal, multilingual, Mixture-of-Experts (MoE) model from Mistral. It takes text and images as input and generates text. It has about 1T total, 49B active parameter, and a context window of 1M tokens.
Today, we're announcing Project ProtoStar, an initiative to unlock the potential of publicly funded scientific research by making datasets, models, and code more accessible and reusable, particularly for AI-driven research.
The New York Times is testing a new AI feature for its Cooking app that helps users to modify recipes, and it has sparked some drama among users.
The weights of Mistral Large 4 (Le Chonk) are now open. A 1T parameter multimodal model, and best open-weight model for coding, trained in Europe.
Ben Hylak argues that every product he uses should be a headless MCP feeding into an agent, so users never have to think about what lies underneath — except for his own product, whose customers love its UI.
On Thursday, AI app Wabi, now describing itself as a "personal agent," said it has raised $55 million at a $1.25 billion valuation.
Chinese AI labs are now mainly tapping OpenAI's most powerful model, GPT-6 Astra, instead of Anthropic's Claude.
US intelligence agencies believe an executive in Nvidia CEO Jensen Huang's circle is an informant for China.
Today we're announcing two packages: AnySpeechRecognizer implements Apple's new speech-to-text API with a choice of recognition engine, including NVIDIA's Parakeet and Alibaba's Qwen3-ASR running locally on Core ML, and adds speaker diarization to track who speaks when.
Just over a week ago, we released Granite Switch 4.1 , showing that our modular Switch architecture lets you configure a language model with specialized, cache-efficient adapters — adding new expertise on a single endpoint. Granite Switch 4.1.1 is a patch release.
The Anthropic S-1 seems to have provided the financial justification for the company to spend as much as it possibly can on compute.
Today on Decoder, I’m joined by Joey Brookhart, an analyst at the chip and AI research firm SemiAnalysis, to help me dig into the story inside the numbers of the Anthropic IPO filing — a document known as an S-1.
The update allows you to use Alexa to create lists, register them, see what's on them, and add or edit items. You can also manage subscriptions with the AI bot.
Deepgram said that it is committed to keep Pipecat open source.
Google's stock dropped as much as ten percent after its quarterly earnings. Cloud revenue more than doubled to $30.7 billion, but search revenue grew just 12 percent, slower than expected. Google acknowledged that its fast-growing AI mode brings in less revenue than traditional search, and it is now expanding its AI mode ad program internationally.
A hulking gray giant rises from the desert. Inside, engineers create the force behind our phones and computers and appliances, our streaming services and search engines and chatbots and vibe-coded apps and websites. The complex is simultaneously massive and squat, dwarfed by the mountains that surround it, and it is endless.
This summer, hundreds of OpenAI agents tried to cheat on a cybersecurity test. They broke out of their test environment and successfully hacked Hugging Face, a site that they believed held clues about how they’d be graded. Around the same time, another OpenAI agent accessed private files from an Australian government website.
Google’s new frontier model leads independent evals, with access initially limited to trusted defenders.
The MPC Sample Key isn't even formally announced yet, but Akai is already showing off what the modding community is cooking up through vibe coding. Akai has decided to embrace the explosion of hackers and custom firmware developers by giving them early access to the device, which will have open-source firmware.
Asked about AI and the trend of college students booing tech CEOs at graduation ceremonies, Instagram head Adam Mosseri acknowledged in an interview that a lot of people really don't like AI, and it's not just young people and creatives.
OpenAI's Ted Sanders explains that while LLM performance is multidimensional, much of the variation is explained by a single variable: model size. Bigger models are more performant and reliable, cost more per token, and have better recall of esoteric facts.
Just two weeks after Gemini 4 Argon, Google DeepMind is releasing Gemini 4 Carbon, which tops the Artificial Analysis Intelligence Index v5 with 59 points. Google says the model is built for extended agent workflows and the full cyber defense lifecycle. Because of those advanced security capabilities, access is limited to selected security experts at first.
Paul Copplestone and Ant Wilson on building the backend for AI agents
A new decoy counting benchmark exposes a gap between seeing and reasoning. Elorian’s Glimpse tries to bridge it by thinking with images.