rvk @rccelera

how viral will this post go bc i mention AI

ai slop

me seeing all of the nyc sunsets online & regretting not leaving my apartment all day 🫣

maybe apple is the only (big tech) company that doesn’t need to pace alignment bc it never had good AI to begin with. @Apple you need to step up your game.

writing and asking questions are powerful predictors of how you think.

i don't think grades will continue to exist in the coming years. grades are not an accurate indicator of performance in the classroom anymore.

ai is taking over our jobs

Kalshi @Kalshi

a month ago

JUST IN: Nearly 17% of Americans aged 16 to 29 neither working nor studying

259 362 5K 5.1M 665

i’m going to say something everyone else in silicon valley is too scared to say: harvard is a giant ripoff. it’s absurd to me that naive 18 year olds give away 4 years of their life & several hundred thousand dollars in this day & age. it’s a super boomer offering: - apply, wait, hope you get in - get thrown into a 7,000-person cohort where you are one tiny cog in their prestige portfolio - 4 years of “education” you can now get completely for free online - office hours with a guy whose actual job is writing papers nobody will read was harvard worth it 50 years ago when knowledge wasn’t freely available & you couldn’t build distribution for yourself via the internet? yes. is it worth FOUR years of your life in today’s age when you can go from $0 to insufferable linkedin thought leader in approximately six weeks with chatgpt? no.

you need people in the same curiosity bracket as you

reading is cool silence is cool thinking is cool writing is cool boring is cool awareness is cool happiness is cool frolicking is cool exercise is cool

chase discomfort i wish we all put more emphasis on experiencing discomfort. we win as humanity when people take changes & do crazy things. when someone quits their safe job. when someone ships the thing that isn’t quite ready. when someone says yes to the opportunitiy that terrifies them. tldr; you’re probably way too comfortable rn. chase discomfort & suprising outcomes will follow.

once in a while, you should do incredibly inconvenient things for a very select number of people that you love a lot, simply because you love them a lot

the best way to be likeable is to ask someone about themself. everybody loves talking about themselves. you already know the story of your own life. you’re far more likely to learn something from hearing someone else’s

Especially in this world of AI, where everything is mass-produced, it is increasingly rare to see people fixate on an idea for a very long time until it is perfect. I think that kind of obsession is what creates world-class people in their fields, whether they are engineers, scientists, designers, or mathematicians. It is revisiting a piece until the water flows laminarly, reworking every stitch in a dress, or fixating on a piece of code until there is nothing left to fix. I think this is partly why I’m sad to see that AI has solved the Millennium Prize. Part of me wants to pluck the brain of the next person who would have solved it. The way you pick the brain of an AI is fundamentally different from the way you pluck the brain of a person. If you ask an AI a question, it gives you an answer. If you ask a human a question, they will give you an answer, but probably not in a linear way. Humans go on tangents, and those tangents go on to reproduce more ideas. One thought reminds them of another thought, which reminds them of something they read ten years ago, which turns into a completely different question. Humans go on tangents. AIs do not go on tangents. And I think this is part of what makes humans human. Humans can become obsessed in ways AIs, which are mass-produced, never will. There is something so beautiful about a person spending years of their life thinking about one problem, returning to it again and again, and letting that problem reshape how they think, what they write about, and how they shape future generations.

i want to hear terrence tao’s thoughts on ai solving a millennium problem

no one knows what they're doing it's our first time living & we're figuring it out as we go

i think all of the Millennium problems will be solved by the end of this year, except P = NP

rvk @rccelera

a month ago

i think all of the Millennium problems will be solved by the end of this decade...maybe even sooner than that. 2027 is a good estimate

0 0 2 71 0

@tszzl which model did you use?

if all of the millennium prize problems get solved, what is the next northstar of mathematics?

bubble boi @bubbleboi

a month ago

Rumor: Another Millennium prize problem just got cracked today using AI. It’s being reviewed right now in secret … I can’t say which one as I don’t want anyone to scoop the team working on it. I don’t have the proof myself but if I get it will post here and credit the authors.

85 66 2K 334K 286

one of the reasons why AI labs are speaking out about superalignment is because their employees are speaking out about it if the proletariat are speaking against the bourgeoisie, the bourgeoisie will eventually have to do something there will always be more proletarians than bourgeois