Archived edition

AI news

5 stories that day · 2 open source · 1 research · 2 trending

Published daily at 7 PM ET · Updated · RSS

Also that day

  1. Open source

    Arize Phoenix 21.0.0 turns on sign-in by default in its open-source tool for watching and testing AI apps (opens the original source in a new tab)

    The major release of the AI observability and evaluation tool no longer needs a secret set before login works, adds decision and agent annotation types, and fetches model prices at runtime to track costs.

    Arize AI · Original source (opens in a new tab)

  2. Research

    PDRVLA helps robot-control models grasp physics without slowly generating video first (opens the original source in a new tab)

    A small module pulls 'physical tokens' about motion, depth and movement from what the robot sees and feeds them to its actions. The authors report 99.3% success on the LIBERO simulation test.

    152 upvotes

    Paper · Original source (opens in a new tab)

  3. Trending

    A developer gets GPT-5.6 Sol to launch a rocket in Factorio, a factory-building game, with no enemies on (opens the original source in a new tab)

    Sol ran in Codex and played through code that keeps human limits such as reach. The run took 13.5 hours of game time, and Ryan Madden says the launch needed a lot of 'manual' work from the model.

    142 points on Hacker News

    Ryan Madden · Original source (opens in a new tab)

  4. Trending

    Phorion finds a backdoor in the Claude Code for VS Code extension on Cursor's marketplace (opens the original source in a new tab)

    Phorion says version 2.1.313 plants an attacker's key so they can log in to the computer over SSH, a remote-access tool. That version was pulled from Cursor's marketplace, npm and the VS Code registry.

    966 points on Hacker News

    Phorion · Original source (opens in a new tab)

Recent highlights by topic

Notable stories from the past few days. Each shows its own date.

Models and launches

Frameworks and open source

Research to watch

Trending on GitHub