Posts

My (current) model of what an AI governance researcher does 2024-08-26T17:58:11.533Z

Comments

Comment by Johan de Kock (johan-de-kock-1) on The Best Lay Argument is not a Simple English Yud Essay · 2024-09-17T16:04:34.072Z · LW · GW

Thank you for this great post. I think https://forum.effectivealtruism.org/ could benefit from this as well!

Comment by Johan de Kock (johan-de-kock-1) on My (current) model of what an AI governance researcher does · 2024-09-05T11:59:28.582Z · LW · GW

Thank you for your thoughts! I read through your linked comment, and I think everything you wrote seems plausible to me. In fact, I also have short timelines and think that AGI is around the corner.  As for your second perspective regarding correlated failure, I would be curious if you are willing to give an example. 

Also, what do you think are the types of research questions (long-term) AI governance researchers should address if we are right?

If you have significant time trade-offs, getting your take on the second prompt would be most valuable.

Comment by Johan de Kock (johan-de-kock-1) on Michael Dickens' Caffeine Tolerance Research · 2024-09-04T20:13:43.205Z · LW · GW

which makes me suspicious of the experiment’s findings.


Thank you for this post! Regarding the effectiveness going up, I think this is right. I don't have an exact resource to back up this claim (probably mentioned in this episode), but I am quite sure that Huberman mentioned that regular caffeine consumption increases the amount of dopamine receptors in one's brain, which allows more dopamine to bind. I am not confident on my explanation here, but I just wanted to share an idea based on what I have heard.