[yuhe00]

I don't know a lot about psychology. The only things I've read are pop-psych articles like this about "traits", "D-factors" and "personality types", which you got to admit do sound a bit much like pseudo-science, or at least not telling the full story. It's useful, for sure, but I'd like to know more about the formal methods employed in modern psychology today. I imagine statistics, surveys and telemetry perhaps being key, but like all methods based in data - You have to be very careful; testing methodology and interpretation biases can heavily influence the results.

I think psychology could bring valuable insight into the future of AI as well. It's a shame that most psychologists involved in tech are in HR/communications or at best tasked with keeping people addicted to whatever social platform they are building.

[Sharlin]

It is true that psychology as it is currently practiced in many places has one of the worst replication crises [1] of all sciences, with lots of p-hacking and other unethical practices going on as researches attempt to get publishable results in a field where most data you get is necessarily messy and difficult-to-control confounding variables abound. But that's more about the usual perverse incentives in the academia rather than a fundamental problem with the field itself.

[1] https://en.wikipedia.org/wiki/Replication_crisis

[yuhe00]

Oh, I was genuinely unaware that the replication crisis was such a big problem in psychology - It makes a lot of sense though, and it was a fascinating read :) And yes, I do agree that there's a problem with incentivization in all the sciences.

[analog31]

In my view, the impact of the replication crisis is to leave us with a science that embraces scientific methodology but has not produced a reliable scientific knowledge base and is burdened by an unscientific past. Digging out of this mess could take a long time, and is not guaranteed to produce results that are interesting to the general public.