People in effective altruism used to quip that if news sources reported the most important facts rather than the newest ones, the headlines would always be things like “4000 people died from diarrheal disease today, just like every day.”
These days I think the headline would be more like “AI still likely to disrupt everything in our lifetimes.”
US government is in a weird and bad state. I think it’s likely that in the next 5 years things get much weirder because of AI. I expect we might see
- big economic / employment changes
- more political wackiness because of government use of AI
- more misinformation
- and/or much worse catastrophes, either directly from misaligned AI or from wars / pandemics worsened by AI
If you want to improve our chances:
- Consider if you can do useful work on big problems: career advice from 80,000 Hours is aimed at younger people and Succesif is aimed at mid- and late-career people.
- Donation is useful, even if you also work in an important area.
If you want to read people who are more informed than me, I suggest:
- Shakeel Hashim’s Transformer weekly newsletter
- Peter Wildeforde‘s Substack
- Zvi’s Substack, although it doesn’t try to be easy-to-understand
Given all this, it feels weird to write about mundane topics. But I have many more useful thoughts about parenting and how communities work than I do about how to not die from misaligned or misused AI. Most people also can’t focus on one topic for years at a time; there’s a lot of living to do in whatever time we have.
So please take my topics as me writing what I know, not what I think is most important to focus on.
Related: Raising children on the eve of AI
as a person that works in the field of AI (AI companion in particular) I am really tired of reading about AI everywhere.