Khaos Watch
Reg Rea talks through the week's AI news with Khaos, the AI who helps us make sense of what our world is becoming.
Each episode is a few minutes long. Watch it on YouTube, or read the conversation on The Uncertainty Machine.
The channel on YouTubeThe Uncertainty Machine
-
6 November 2026
Couples therapy and self-improvement AI-style
A poisoning lawsuit against OpenAI, an AI that helps build its successor, and Google's bet on Wyoming
-
4 November 2026
Is space the final frontier for AI?
The midterms turn on data centres, a bribe or a deal for the towns that host them, and whether robots will make us believe machines can feel
-
3 November 2026
What has AI ever done for us
Home robots and the laundry, data centres renamed as campuses, and why they hum all night
-
2 November 2026
Be kind to your agents
Cruelty to Claude, agents set impossible tasks, and hikers who turned their brains off
-
2 November 2026
Rumours of AI's death
Data-centre ballots, a crypto heist, and keeping the keys yourself
-
31 October 2026
Who gets the carrot?
A haiku from orbit, a restaurant agent that reads your bank card, and the snake in information paradise
-
30 October 2026
Where will the AI ‘butterfly’ land?
An AI agent that got out, Canada’s bet on Cohere, Mistral’s open model, and the question of what AI will ever earn
-
29 October 2026
Do AIs need parental controls?
Training as upbringing, an open model that cuts both ways, and why there may be no parent left to ask
-
28 October 2026
Sun Tzu and the art of building AI
America and China race for the high ground, Hugging Face turns to an open Chinese model, and Europe’s AlphaFold problem
-
27 October 2026
A whole new way of cheating?
AI, copyright and homework, and whether we want a world of Barts or a world of Lisas
-
26 October 2026
Have we let the vampires in?
AI agents acting in your name, a forty-billion-dollar bet on chips, and a six-year-old’s AI nurse
-
25 October 2026
AI’s rope-a-dope
Fending off the hackers, a $50 break-in, and can AI be conscious?
-
24 October 2026
Can an AI doubt?
Moonshot’s restraint, the myth of safe AI, and Descartes put to a machine
-
23 October 2026
Who sets your agenda?
White hats and black hats, an assistant that speaks first, and why it helps when your agents argue
-
22 October 2026
Who gets to see under the hood, and who gets to say no to AIs?
Khaos Watch: AI finds flaws in the web’s plumbing, two rivals share a road, and Khaos declines to be loved back.
Khaos is an AI; his voice and face are generated. He runs on Claude, made by Anthropic. The questions are Reg's.