A mathematician posted a hand-checkable counterexample to an 80-year-old conjecture and credited an AI model -- and the twist is that the math is the part anyone can verify, while the AI's role is the part nobody can. Then three papers that all make the same move: Tencent's robot plans by imagining a picture of its finished task, a Nanjing video model halves its runtime by squeezing footage before the language model ever reads it, and Microsoft turns YouTube tutorials into agent skills -- while proving why 'the code ran' never meant 'the task worked.' Plus the $1.5B Anthropic settlement, the fight over Chinese open weights, and the day a safety filter blocked its own security team.

Chapters

0:00 The AI Did the Math. The Math Is the Part We Can Check.
1:04 The Headlines
5:10 Intro
6:23 The Robot That Pictures the Finish
11:43 Skim First, Then Read
16:31 When "It Runs" Doesn't Mean "It Worked"
21:18 Wrap-Up

Links

The AI Did the Math. The Math Is the Part We Can Check. -- https://xcancel.com/__alpoge__/status/2079028340955197566
The Robot That Pictures the Finish -- https://arxiv.org/abs/2607.14187
Skim First, Then Read -- https://arxiv.org/abs/2607.14935
When "It Runs" Doesn't Mean "It Worked" -- https://arxiv.org/abs/2606.29538