From a Teams bug alert to a pull request: GitHub unifies its app
GitHub’s October 23 update combines repository notifications and Copilot in one Teams app.
Tools
Apps, services and new features built on AI models.
Find land like this: Ai2 opens its satellite AI explorer
Ai2 released OlmoEarth v1.3 and a no-login map explorer on October 22.
From a quick question to a finished report: Claude unifies its enterprise app
Anthropic made unified chat and Cowork generally available for Claude Enterprise on October 23.
Claude’s newest features, billed by Google: a new deployment option
Claude Models on Google Cloud now has a generally available Anthropic-managed option.
AI waits, its sandbox pauses: Google announces Agent Compute Engine
Google announced Agent Compute Engine, with a virtual execution environment for each agent.
AI logs in without seeing your password: 1Password expands agent access
1Password announced a platform that fills sign-in credentials after a person approves access.
Beyond finding bugs: Theorem offers free proofs for open source
Theorem offers free AI-driven formal verification to qualifying open-source projects.
Many AI agents, shared company knowledge: Sakana introduces Aquarium
Sakana AI announced Aquarium, a platform for agents sharing company knowledge and rules.
Close your laptop, keep work going: Claude Duet enters research preview
Anthropic opened a Claude Duet waitlist for Pro and Max subscribers in the US.
Why did AI training slow down? NVIDIA’s Minerva connects the clues
NVIDIA announced Minerva, an AI infrastructure operations platform, on October 22.
NVIDIA expands Cosmos with five training-data blueprints
NVIDIA expanded Cosmos data, training and evaluation tools at GTC Berlin.
OpenAI launches Decisions API with up to 400 questions per request
OpenAI made its Decisions API generally available on October 22.
