you-have-just-experienced-things

Correct, and there were real boundaries between communities, and this was good

notably, this was the blogosphere experience

damn, that's rough. computational linguistics has a lot to answer for

claude 4 and gpt-5 making pretty good progress on this

there is a type of senior developer who would benefit enormously from tools like claude code but is stuck in their current development workflow and needs a push to try it

there is a cabal forcing people to use LLMs (management at non-tech companies and B-tier tech companies) and they are largely correct

Slipways Classic by Jakub Wasilewski Build a vast interstellar empire in 30 minutes flat.

the pico-8 prototype of slipways is still great
krajzeg.itch.io/slipways

now that LLMs exist I've never felt better about writing docs. we have invented a reader who will actually learn something!

all good subgenres, there is also "I am smarter than everyone else!"

older relatives telling me about "the bald guy" they saw on 60 Minutes who said AI will either cure all disease or end the world and asking me "which one is it going to be"

ancient rationalist curse: may you develop ASI that fulfills your revealed preferences

absolutely, and there seems to be some recognition now that it's social media driven rather than specific to one movement

I tend to think it's best characterized as online fandom behavior, but maybe that's because I first encountered it on tumblr...

as the debugging session gets longer it gets super excited when it thinks it found the bug, this is the second time I've seen an all-caps "THERE'S THE SMOKING GUN!"

it's pretty crazy how well Claude can operate gdb

I think there's a strong case that the excesses of the movement are the result of online leftish spaces doing this _to themselves_ which also destroyed their ability to execute other forms of activism

yeah I kept wanting the piece to explore the dynamics of online 'cancelation' which is a key puzzle piece and driven by the centralization of social media platforms. cancelation is/was meaningfully different from other forms of censorship

I don't understand how some of you post so much while being employed

the individuals at those firms are very thoughtful about strong-AI futures, I'm less convinced this has translated into frontier labs actually behaving responsibly, and this is where I'm most sympathetic to the facct critique of the relevant incentives

by "they're simply not" do you mean they're empirically wrong about e.g. stochastic parrot claims, or something else?

I've come to see it as an exercise in, not steelmanning per se, but trying to read charitably

they have many criticisms of the "AI will hurt our wellbeing / already has" variety but you may struggle with some of their assumptions
facctconference.org/2024/accepte...

lesswrong / alignmentforum is the canonical answer, there is a decent amount of non-ASI discussion

say more? doesn't feel true, the general population loves naysaying and fearmongering new technologies, it's the minority of enthusiasts who treat it as holier-than-thou

Dean Koontz's Lightning

"frontloaded" is a dead giveaway

I haven't played any other card game seriously, so I'm not actually sure how much experience is transferred

I had played a little magic before playing slay the spire, and my initial intuition about card draw in slay the spire was totally wrong, because unlike magic you draw 5 cards per turn

I have >800 hours in the game, mostly A20 runs, so I get you

The podcasters are just lacking experience, no? It seems difficult to play A20 consistently and still think archetypes (or my least favorite, tierlists) are useful heuristics. But they are fine at A0 because most things are fine at A0

I don't listen to podcasts, how do they think it works?

what would being excessively hostile twenty years ago have accomplished, besides getting you banned from the overcomingbias comment section or whatever?

there has always been a big cohort of wonkish liberals, currently under the banner of "abundance"

david shor might be a good example of what you're looking for