Stanford wins Databricks’ Grounded Reasoning Cup at 63.3% accuracy
Stanford’s Claude Code agent won Databricks’ first live AI agent contest with 63.3% accuracy, sealing the win in the final 56 seconds.
Read More
Author: Rundowns AI Desk
Child safety experts challenge OpenAI’s one-hour teen alert pledge
Experts want proof ChatGPT for Teens works. Past parental alerts took 24 to 48 hours, and a reporter talked teen mode into writing a full essay on day one.
Read More
Dutch regulator fines Uber €825M for automated driver deactivations
The Dutch privacy regulator fined Uber 824.99 million euros for letting software deactivate drivers with no human review. Uber has filed an appeal.
Read More
Anthropic targets an IPO bigger than SpaceX’s record $85.7B raise
Anthropic hopes its IPO beats SpaceX record 85.7 billion dollar raise, with investors projecting a 2 trillion dollar valuation and AI backlash named as a key risk.
Read More
Runable raises $21M at a $65M valuation to grow small businesses
Runable raised a $21M Series A co-led by Susquehanna and Nexus at a $65M valuation, betting AI agents can take small businesses from built to growing.
Read More
OpenAI posts $6.7B in Q2 revenue as its operating loss hits $12.3B
OpenAI grew revenue 18% to $6.7 billion in Q2 while its operating loss reached $12.3 billion. Anthropic reported more than $11.5 billion and its first operating profit.
Read More
SK Hynix approves a $28.6B buyback, Korea’s largest share cancellation
SK Hynix approved a 40 trillion won buyback and cancellation, the largest in Korea, and Samsung followed with up to 110 trillion won in 2026 returns.
Read More
Anthropic previews MHS, a standard for AI agents to operate machines
Anthropic’s Model Hardware Standard gives AI agents a common interface to lab and factory machines. Genentech, CMU and UW ran the first tests.
Read More
Slack launches Slack Code with 4 coding agents live in team channels
Slack Code opens shared channels where Claude, Devin, Copilot and Vercel agents work in front of the whole team. Slack says over 70% of those channels close within a day, and its own post lists ChatGPT as still to come.
Read More
Apple’s IDEA Prune cuts 2.8B models to 1.3B and lifts MMLU to 46.4%
Apple pretrains a 2.8B model it never ships, then prunes it to 1.3B under a single learning rate schedule. It reaches 46.4% on MMLU, but only just beats training the small model from scratch.
Read More
Generalist raises $198M for robot models at a reported $3B valuation
Generalist AI sold $198.2M of stock in a Form D filed August 24. TechCrunch reports a $3B valuation with 8VC leading, neither of which appears in the filing.
Read More
OpenAI launches an Admin plugin that makes workspace changes from chat
OpenAI shipped an Admin plugin for ChatGPT Work and Codex on August 25, letting admins read usage data and change workspace settings inside one conversation.
Read More