Podcast: Retrofits
Reading of posts from August 17, 2026
AI StopWatch
A non-profit AI-focused newsroom by analysts and writers of the Machine Intelligence Research Institute. We help you keep up to help you speak out.
Overdue Last read · last published · next check
Last read 24 hours ago, longer than this feed's 3 hours schedule.
Latest episodes
Retrofits
Shrinking timelines, lawslop in Congress, and AI treaty verification technology
Verifiers verified
The tech needed to verify an AI pause or slowdown treaty is coming along, but is shamefully under-resourced
Timelines shrink in AI Futures quarterly update
Three new methods give the same unnerving answers
Congressional office drowning in lawslop
Congress's legislative proofreaders say AI-generated bills aren't ready for primetime
Podcast: An ounce of prevention
Reading of posts from August 16, 2026
An ounce of prevention
Insights from public health principles, and updates on White House policy conflicts
Updates on the White House's conflicted AI posture
How's that whole "supply chain risk" thing going?
When prevention works, nothing happens
What public health can teach us about acting on AI risk before disaster arrives
Podcast: Doot doot do-dee do-dee doot doot doooo doot
Reading of posts from August 15, 2026
Doot doot do-dee do-dee doot doot doooo doot
Clownery at Anthropic, and the limited value of watermarks
Who are AI watermarks for?
Anthropic's new watermark feature unlikely to change much on the ground
The clown college inside Anthropic
New risk report is commendably transparent, but well-meaning farce is still farce
Podcast: Coordination problems
Reading of posts from August 14, 2026
Coordination problems
Bipartisan AI, Chinese AI, global governance, and agents (not) collaborating
Ten thousand of you, all at once
What happens when frontier models are forced to work — or feud — together
To keep the world safe from AI
Writers and academics argue that world governments need to cooperate to govern AI
How much should we worry about Chinese open models?
Incentivizing open weight models is the wrong way to deal with China's AI strategy
AI sentiment crosses party lines
Ahead of the midterms, politicians seek AI policies that resonate with voters
Podcast: Heave to, me hearties
Reading of posts from August 13, 2026