vatsachak

4 ms·

61.
by vatsachak 1mo ago
Because it's beautiful. Because you love math for maths sake and not some weird egotistical game
62.
by vatsachak 1mo ago
Any serious mathematician would read the LLM output and rework their understanding. I read a bourgain paper a week in grad school and they're probably worse than an LLM generated paper. I still had to recreate the tricks in my own lang
63.
by vatsachak 1mo ago
You mean manufacturing
64.
by vatsachak 1mo ago
Exactly. Mathematicians are trying to cope that math wasn't slop this whole time. Grothendieck was anti-slop but most papers are slop. I don't think AI is going to rewrite bourbaki anytime soon
65.
by vatsachak 1mo ago
[flagged]
66.
by vatsachak 1mo ago
The irony of LLMs is that OpenAI provided software is still just a wrapper. The only way they can get more users is through exclusive discounts on their wrappers
67.
by vatsachak 1mo ago
Cool story. If it's true the company will be bought by open AI/Anthropic and Chinese labs will discover the trick and open source it by next quarter.
68.
by vatsachak 1mo ago
Nah the models are just that good at short scale math now. If your research program is close to it's goal the model can get it.
69.
by vatsachak 1mo ago
Very cool. I wonder what's the minimum amount of money you'd need to get it to score well on AIME
70.
by vatsachak 1mo ago
What was your prompt? I have gotten easy fixes with I tell if to do something.
71.
by vatsachak 1mo ago
The only show worth running for 20 years was Xavier Renegade Angel
72.
by vatsachak 1mo ago
How do you guys even use LLMs where you are finding Astra light is worse than Sol High?
73.
by vatsachak 1mo ago
Okay how are Add nodes created? Does an LLM come up with the name, guidance and edges for each node or is it a bespoke transformer model?
74.
by vatsachak 1mo ago
I have a math PhD and publications in top journals, I left math for programming because I hated academic politics
75.
by vatsachak 1mo ago
That's fair, but the proofs are different and there was no active perusing of Tristan's approach.
76.
by vatsachak 1mo ago
Totally. Anthropic is like a village cottage shop who was just like chilling until big bad OpenAI came in
77.
by vatsachak 1mo ago
Lol what? Everything is computation. The natural sciences will soon start breaking too. I will concede that AI seems likely to not invent a "research program" anytime soon. It has no taste
78.
by vatsachak 1mo ago
Brain has loops and parallel connections. Loops and parallel connections make transformer go brrr
79.
by vatsachak 1mo ago
45 pages only. God damn that internal model is crazy
80.
by vatsachak 1mo ago
Called it. AI wins a fields medal before managing a McDonald's
81.
by vatsachak 1mo ago
In math we have a thing called a "scoop"; another mathematician publishing a result that beats yours before you published it. The scooper hardly acknowledges the scoopee unless the methods used were orthogonal. The scooper gets th
82.
by vatsachak 1mo ago
Okay suppose that you have a trillion dollar competitor salivating at the mouth to ruin your business, which is based on user privacy. Why would you risk the trillions of dollars worth of business for the niche result of Navier-Stokes, whic
83.
by vatsachak 1mo ago
And everyone should be glad for that!
84.
by vatsachak 1mo ago
Winning a millennium prize is not worth the fallout of "we will steal your IP"
85.
by vatsachak 1mo ago
This is why I left math even after solving a 20 year old conjecture in grad school. Literally who cares who solved the problem just publish the results. Academia was always politics first results second and I AM GLAD that LLMs are becoming
86.
by vatsachak 1mo ago
There is a large market for artisan code. It's called open source. Good code is pretty much the solution with some boilerplate tacked on.
87.
by vatsachak 1mo ago
I read code for fun in my free time
88.
by vatsachak 1mo ago
RSI started when humans discovered tool use. I mean one could argue that RSI always begins in any physical environment. The book "What is intelligence?" by Blaise Aguera is great
89.
by vatsachak 1mo ago
So before I got a job this fall, I was working on a side project about compiling a particular language to SQL. To test Astra I pulled it off the shelf and asked it to take the grammar and then create a compiler to SQL. I've done this b
90.
by vatsachak 1mo ago
You can already sense that we've entered a new era. People are beginning to realize that in ten years humans will not be the main progenitor of technology

More ›