Jason Cohen

asmartbear.com

20 December 2025 20:19

It is the tricky part! All the more reason not to spent precious time on brand, when you still have the hardest part to figure out.

Jason Cohen

asmartbear.com

20 December 2025 20:19

I wouldn't even trust AI to book a flight, much less set a strategy. Definitely quit, if your CEO is setting strategy by ChatGPT!

Carlos Eagle

mexicoseagle.bsky.social

26 October 2025 15:51

each day on this planet is worse than the last

Sam Power

spmontecarlo.bsky.social

26 October 2025 15:09

Oh okay, sure; quite a broad question! I guess that if I wanted to narrow the answer to something like "research directions which I can't really pursue now, but would like to be able to", then some examples would be things like (next post)

Sam Power

spmontecarlo.bsky.social

26 October 2025 15:08

1) 'full-pipeline' theory which combines statistical theory w/ computational theory, 2) understanding the utility of 'purely predictive' methods for statistical inference and estimation 3) using scaling limits to reason usefully about large-scale problems and I may be missing others.