If they can convince the money hose that they just need one more OOM of compute bro, they can keep vacuuming infinity dollars. The incentive is obviously there for any amount of lying, but at some point, I assume even the most braindead investors will start asking around if this really is the only game in town.
Or maybe they wonāt, which would be an admission against the core tenet of capitalism, but this has been a crazy year, and itās only january. :/
https://www.wheresyoured.at/deep-impact/ And the companion piece on his blog.
What I didnāt wager was that, potentially, nobody was trying. My mistake was ā if you can believe this ā being too generous to the AI companies, assuming that they didnāt pursue efficiency because they couldnāt, and not because they couldnāt be bothered.
This isnāt about China ā itās so much fucking easier if we let it be about China ā itās about how the American tech industry is incurious, lazy, entitled, directionless and irresponsible. OpenAi and Anthropic are the antithesis of Silicon Valley. They are incumbents, public companies wearing startup suits, unwilling to take on real challenges, more focused on optics and marketing than they are on solving problems, even the problems that they themselves created with their large language models.
I mustāve read the words āJevonās Paradoxā a hundred times today, I didnāt realize that many people had their livelihood predicated on NVidia going to the moon forevermore.
Iāve been struggling with what the appropriate level of engagement for all the tech shit is.
I can stick to making fun of the AI crap and whatever else the tech people shit out because itās tangible for me, and I can more or less be an effective gatekeeper for my community, but the problems go beyond just a bunch of rich tech weirdos floating bad ideas, itās what theyāre trying to paper over. The fact that theyāre incompetent at it is very funny, but Iāve been laughing with gritted teeth for too long.
They understand that all of the major model providers is doing it, but since the major model providers are richer than they are, they canāt possibly ask OpenAI and friends to stop, so in their heads, it is what it is and therefore must be allowed to continue.
Or at least, thatās my face value read of it, I certainly hope Iām simplifying things too much.
The conspiracy theorist who lives in my brain wants to say its intentional to make us more open to blatant cheating as something thatās just a ācost of doing business.ā (I swear I saw this phrase a half dozen times in the orange site thread about this)
The earnest part of me tells me no, these guys are just clowns, but I dunno, they canāt all be this dumb right?
Weāre like 3 steps removed from the iq dickmeasurers moving from measuring skulls to comparing PoE builds.
I donāt even know if Iām joking here, what the fuck are these people doing.
The computation seems to be generating a uniformly random set and picking a sample of it. I can buy that itād be insanely expensive to do this on a classical computer, since thereās no reasonable way to generate a truly random set. Feels kinda like an unfair benchmark as this wouldnāt be something youād actually point a classical computer at, but then again, thatās how benchmarks work.
Iām not big in quantum, so I canāt say if thatās something a quantum computer can do, but I can accept the math, if not the marketing.
tfw the Holy Book (Atlas Shrugged) was misinterpreted.
Itās also just less effective this time around. Everyoneās afraid of being poor, capitalism as it exists functions based on that, so thereās a strong emotion that gets people wrapped up with the NFT hype train.
Itās a lot less convincing to go āHave fun being as productive as you are right now, loser.ā, at least without also going mask off and sounding like a supervillain.
Chat-GPT-TFSD-21guns can have a little anthropomorphism, as a treat.
This is somehow even sadder than āI have depicted you as the wojakā.
āI contracted the least efficient computer possible to depict myself as the chad, therefore I win.ā
I believe its because they A-B tested his thumbnails, editors found that they vacuumed slightly more eyeballs when he was smiling, and updated the thumbnail images of him not smiling with photoshop.
The fault isnāt with the photoshop though, he just doesnāt have a soul so it actually looks like that when he smiles.
Itās the imbalance of power. No one wants to be the gatekeeper at the local game store telling the 12 year old kid āSorry, you canāt play your Spongebob deck here tonightā, whereas as the ones printing the cards, they can be gatekeepers of āSorry, you canāt play your deck without the One Ring here tonightā off of the basis of ānormalā power creep.
Skimming the reddit thread in search of general public sentiment about this, but unfortunately mostly just found a greatest hits compilation of very gross comments.
According to these very smart people, parents should expect your teenager to die as an outcome of not being perfect people 24/7, technology can never be at fault even when it literally tells you to commit suicide in coded language, and itās actually impossible to understand which parts of society are causing kids to be depressed, so we must take it as a given that we canāt do anything about it. I regret having done this to myself.
Now I wonder if Musk calling the diver āpedo guyā was him thinking āyeah, this is exactly what Tony Stark would say to the bad guyā because he doesnāt understand media literacy even a little bit. (More likely, itās just him being the regular type of weird weāve come to expect from him)
https://github.com/SatoshiNT0 Satoshi really got into emojis, I guess?
š± Iām currently learning ā¦ Quantum Cryptography.
šļø Iām looking to collaborate on ā¦ our world!
Cool.
Metaās happy to give away models for free, so models are evidently worth $0.
Youāre correct that I canāt stop them from making pants on head stupid decisions, but Iām not going to stop making fun of them.
very likely making it economically viableā¦
Theyāre going to fund currently economically nonviable nuclear plants to power their currently economically nonviable genAI schemes? Over the time horizon of 25 years a decade (edit: misread the article) before they scale up energy capacity at all past the rnd stage? Maybe pants on head is too generous.
There are companies doing ācool-soundingā things with AI like Waymo. āGoodā would require more definition.