I’ve been a listener to the Last Week in AI podcast for years, and come to trust and respect Jeremie and Andrey’s judgment and takes. Jeremie started a new newsletter at AIEndGame.com, and the most recent post is a long prepper article based on his belief that a large scale AI-related attack is very possible. He is much more knowledgeable than I am, and has far more contacts in the AI space at labs and in government, and I’m inclined to trust his take. And the last 6 months of AI developments (HuggingFace hack, various outages in the last two weeks, AI solving Millenium Prize math problems, etc.) have me a bit nervous. On a similar note, the Dwarkesh podcast had a guest this past week that was adamant that we will have AIs in the next 12 months capable of overthrowing governments— which feels hyperbolic, but still.
All to say: is this doomed stuff overwrought and safe to ignore? Or is following this prepper advice a good idea?
Comments URL: https://news.ycombinator.com/item?id=50600244
Points: 1
# Comments: 3
