ai-today.org

AI labs & models

News from frontier AI labs and open-source and open-weight model teams: OpenAI, Anthropic, Google DeepMind, Meta AI, xAI, NVIDIA, Alibaba Qwen, DeepSeek, Mistral AI and Black Forest Labs.

Feeds refreshed every 15 minutes

Published

GenRecs Tokenization: Transforming Ads Recommendation at Meta with Stable Model Vocabulary (opens in a new tab)

Generative Recommenders (GenRecs) represent a paradigm shift from scoring every ad to reasoning about ad content and user preferences together in order to predict the best outcome for advertisers. Leveraging the world knowledge built into LLMs makes ad matching more intelligent…

Read at Meta Engineering (opens in a new tab)
Image: Meta Engineering
Meta EngineeringStory image unavailable

More matching stories

Page 9 of 63
Image: Google Cloud (official)
Google Cloud (official)Story image unavailable
Published

Developer’s guide to implementing MCP apps in Gemini (opens in a new tab)

Enterprise assistants still respond mostly in text. However, that can break down for visual, multi-step work, where every adjustment costs another slow chat turn. MCP Apps ( SEP-1865 ) solves this. This open Model Context Protocol (MCP) extension pairs a tool with a live UI. In…

Read at Google Cloud (official) (opens in a new tab)
Image: Meta Engineering
Meta EngineeringStory image unavailable
Published

The Thin Orchestrator: How We Made Our AI Agents More Reliable by Giving the Router Less to Do (opens in a new tab)

Sometimes, giving an AI agent all of the knowledge it needs is the wrong approach. At Meta, we’ve seen notable improvements in efficiency and accuracy by keeping the orchestration layer of our data pipeline agents deliberately thin: It routes work and enforces reliability checks…

Read at Meta Engineering (opens in a new tab)
Image: MIT Technology Review
MIT Technology ReviewStory image unavailable
Published

AI researchers stranded Claude on a virtual desert island. It started to “play.” (opens in a new tab)

What do AI agents do if you don’t tell them to do anything? The answer looks a lot like play. Researchers repeatedly stranded them alone on a virtual desert island for 30 hours in an experiment reminiscent of the movies Cast Away or Home Alone. The agents—copies of Claude Opus 4…

Read at MIT Technology Review (opens in a new tab)
Image: anthropic.com (via Future Tools)
anthropic.com (via Future Tools / Matt Wolfe)Story image unavailable
Published

Investing in the public benefit (opens in a new tab)

We’re committing a portion of Anthropic’s shares to the Anthropic Public Benefit Fund, a new entity with a structure that increases the stake as the company grows.

Read at anthropic.com (via Future Tools / Matt Wolfe) (opens in a new tab)
Sources 34 publishers

Headlines link to the original reporting. Article dates show publication time; the refresh time shows when feeds were last checked.

About our sources & editorial standards
Feeds last refreshed

Some feeds are delayed; saved stories remain available.