Why should I have opinions about AI timelines?
substack.com · 1,143 words · saved by 1 readers
Thoughts about deference
August 2026 update: The original version below was written in March 2026. The piece was originally written about AI timelines, as it felt related to impact calculations & career decisions I was trying to make around that time. I now think the piece is less related to why one should have opinions about timelines in particular, but rather an illustrative snapshot of some reasoning pattern I followed that led me to pretty confusing conclusions, that is analogous to many topics that are not timelines. Also, I recently realized that the confusions I describe below has been well discussed in Toby…
saved by
related reading
- The case for multi-decade AI timelinesepochai.substack.com
- How do we (more) safely defer to AIs? — LessWronglesswrong.com
- Consider chilling out in 2028 — LessWronglesswrong.com
- Deferral is a skilllawsen.substack.com
- Broad Timelines - Toby Ordtobyord.com
- Short AI Timelines Aren’t Always Higher-Leverageforethought.org
- Deep Deceptiveness — LessWronglesswrong.com
- How to keep thinkingseangoedecke.com
- Why You Should Almost Never Use AI to Write Anything Substantiveerichgrunewald.com
- I Would Have Solved Alignment, But I Was Worried That Would Advance Timelines — LessWronglesswrong.com
- Thoughts on AI in academiatheinfinitesimal.substack.com
- Questions about the Future of AI - by Dwarkesh Pateldwarkesh.com