Home

2026-02-10

update 2026-03-26

I seem to increasingly not trust base rates

Disclaimer

• Quick Note
• Contains politically sensitive info

Main

• If you want to understand literally anything, a common method is to identify a suitable reference class, collect lots of historical evidence, and calculate a base rate.
• This is popular in academia in general, but it is especially popular in EA/LW circles.
• I personally last encountered this when trying to evaluate base rates of whether consumerist expenses make people happy, and when trying to evaluate base rates of my plans, such as base rates of software developers who turned political activists and succeeded
• I do think it is still a good idea to calculate a base rate, and for lots of important phenomena in society, nobody has tried to calculate one.
• I do now seem increasingly sympathetic to Paul Graham's heuristic though, that once you are the frontiers of knowledge in a field, you are going to notice things that are missing or off about the consensus of the field. I have noticed many such things.
• It does seem totally possible for you to become one of the first people, world historically, to find ways to break that base rate

2026-03-26

Corrolary to the above

• If you plan to work on something where you have or will be reaching human frontier of knowledge, then do not make this decision solely based on success probability calculated from lots of historical data of similar plans. People like Ray Dalio and Paul Graham already agree with me that EV calculations of this sort are a bad reason to prematurely terminate your curiosity.

Subscribe

Enter "Subscribe: " followed by Signal username, email address or whatsapp phone number. You will receive atmost one update per month. View previous updates here.

Alternatively, you can use RSS, or a third-party service like blogtrottr.

Comment

Enter "Comment: " followed by comment. Alternatively, contact me