On AI
Along with immense opportunity, AI brings equally immense challenges and risks, all very real.
But there is good news in that, more recently, the implications of AI are now in the mainstream media. Because people are talking about it, we may have a chance of not destroying ourselves. But in order to see that through, we need to think about what we want for our future, discuss it widely, and then take real action by voting, donating, or actively contributing accordingly.
See also my blog posts on AI.
Key Points
- Many AI CEOs and researchers believe there is a 25% or greater chance that AI could cause human extinction in the not-distant future.
- Recent hacks performed by AI agents provide proof that AI can and will take actions entirely "of their own volition" which could cause "extreme havoc".
- AI need not become "evil" or "conscious" to cause a very bad outcome for humanity. "Too bad for the ants."
- A near-term "intelligence explosion", where AIs are able to continuously improve themselves without human intervention, would, some researchers argue, quickly put AIs so far ahead of humanity that human control of them soon becomes distinctly impossible.
Getting Up To Speed
- Any of the recent interviews with Bill Gates
- Dario Amodei (12 Sep 2026) — We Must Pace the Frontier
- Will Marshall for The Economist (15 Jun 2026) — Humanity isn't ready for the coming intelligence explosion
- Yudkowsky, Soares (2025) — If Anyone Builds It, Everyone Dies — A fascinating intro on the topic and very consumable by non-technical people