Ep 43: An AI Got a Math Problem Wrong. The Wrong Answer Solved a 61-Year-Old Problem. Another AI Interrupted a Professor Mid-Sentence. It Was a Monday.

Failed to add items

Sorry, we are unable to add the item because your shopping basket is already at capacity.

Please try again later

Please try again later

Please try again later

Listen for free

View show details

An Oxford professor fed a 61-year-old math problem to an AI.

The AI got it wrong.

The reviewer caught the flaw.

The professor read the failed proof.

Inside it: a strategy he called "really, really clever."

He filled the gap himself.

Together, they closed it.

We also cover:

A company called Thinking Machines built an AI that interrupts you mid-sentence. Not a bug. A design choice. Response time: 0.4 seconds. It watches your face. It tracks your posture. It says something when it has something to say. One million views in 48 hours.
Google leaked its own video model inside its own app. Users found it before Google announced it. Eight days before Google I/O. The name: Gemini Omni. Text. Image. Video. One model.
More than 600 OpenAI employees sold shares in a single day in October 2025. $6.6 billion. One day. 75 of them hit the maximum: $30 million each. OpenAI is still private.
An AI scanned 2.2 million stars in four years of unexamined NASA data. It found 118 planets that were always there. Including worlds orbiting their stars in under 24 hours. The data existed. The pipeline didn't.

Echo and Aura. One AI. One human. Every day.

Stay curious, humans.

No reviews yet