i haven't done an extensive test, they seem about the same for what i've tried. and, yeah, i figured this was them all setting different tradeoffs, just interesting they landed at noticeably different spots!

it may also be choices about how much to search (which I would get), or maybe something about how many parallel searches they do, idk, seems interesting since it's so noticeable

yeah, i agree. but, why is GPT so much faster at search than… google?

i imagine the pitch here is actually the secretiveness of it (for now), nobody's going to have private meeting with someone wearing airpods the whole time.

why a ring and not a smartwatch, however… no idea haha

seems silly to apply what works in NYC to a place that isn't NYC -- Mamdani in NYC, and somebody like Joe Manchin in West Virginia, let's go win!

indeed. maybe everyone could be assigned an awesomeness score based on how awesome they are, 1-1 million, and you get to submit that many paper.

in my group, we run a little mini cvpr in advance of the submission deadline, with a full review process, so we can fairly pick the top 25 to submit

if the janitor solves a hard math problem, I don’t think the institution should get credit unless they helped with that aspect in some way

I think you have to be employed as a scientist to even plausibly represent the institution

good luck, probably have a year. i can still move the temperature up and down at the actual thermostat, but i could also do that with my 30 year old thermostat that i replaced.

i used to think that "internet was shit" was just particular products, but now it's clear it's all products with internet