Taek

6 ms·

61.
by Taek 7mo ago
It's more likely your attitude rather than your quest for verification that gets you downvotes.
62.
by Taek 8mo ago
No not at all! You can find plenty of videos on YouTube of humans taking down 2015-era stockfish. Usually it involves exploiting specific weaknesses in the engine, for example bringing the game to a stalled position where the game nearly re
63.
by Taek 8mo ago
Out of sheer curiosity, I did a bunch of research to understand just how dramatic a 350 point rating gap is in real word chess. Magnus Carlson, for example, has a 98% win rate against players >350 rating points lower than his own, with z
64.
by Taek 8mo ago
Elsewhere in the thread he revealed that he achieved these results around the year 2015, which means we was playing against Stockfish 6 or earlier, estimated to have about 400 less ELO than today's Stockfish 18. Stockfish 6 didn't
65.
by Taek 8mo ago
2015 stockfish is quite a different beast from 2026 stockfish. Stockfish didn't even add NNUE until 2020. Based on what data I can find, it's estimated that the difference between the 2025 stockfish (stockfish 6) and today's
66.
by Taek 8mo ago
He declined to defend it because he disagreed with the way FIDE was organizing and managing the tournament. I believe this is around the time they threw him out of a tournament for wearing jeans, when he was not the only competitor present
67.
by Taek 8mo ago
My understanding is that this is purely a strategic choice by the bigger labs. When OpenAI released Whisper, it was by far best-in-class, and they haven't released any major upgrades since then. It's been 3.5 years... Whisper is o
68.
by Taek 8mo ago
Yes modern AIs have an entire opening database and generally have cached the first 20+ moves of the game (for most common openings) from a database of very deep searches identifying the best move. This is absolutely a form of opening prep f
69.
by Taek 8mo ago
This is some fascinating data, thanks for pulling it together.
70.
by Taek 8mo ago
It scales quadratically with speed* Those two things very different.
71.
by Taek 8mo ago
I'm late to this thread but I've manufactured about a dozen SKUs and I think you could have saved an enormous amount of grief with one simple step: If you are American, and you are manufacturing 500 units, do it in America. Yes, i
72.
by Taek 9mo ago
Wow, this article is a tour de force in menu and icon design, I learned a lot. For anyone who wants to understand the design world better, or wants a glimpse into the brain of a design-minded person, this is a great article. Incredibly acce
73.
by Taek 9mo ago
The arrow indicates where the hole is.
74.
by Taek 9mo ago
Childhood is only half of subjective life if you stagnate as an adult. If you keep changing your routine, you'll find each year has just as much subjective experience as the previous. Start a company in an industry where you have no ex
75.
by Taek 9mo ago
Wasn't 4.5 before 4o?
76.
by Taek 9mo ago
Oh man, DB has one of the worst experiences in my well traveled history. If you don't buy a seat, you don't get a seat. I was taking the 4am train 8 hours from Brussels to Berlin, and I bought seats for both legs of the trip. To s
77.
by Taek 9mo ago
I have close acquaintances who will take that $50, spend it on drugs, and then starve to death. If you want them to stay alive, you have to give them non-money. If everyone spent money like a rational, 100 IQ individual with a moderate amou
78.
by Taek 10mo ago
> The AI systems that do most of these things are not "LLMs". Uh, every single example that I listed except for the 'playing video games' example is something that I regularly use frontier models to do for myself. I h
79.
by Taek 10mo ago
We seem to be moving the goalposts on AGI, are we not? 5 years ago, the argument that AGI wasn't here yet was that you couldn't take something like AlphaGo and use it to play chess. If you wanted that, you had to do a new training
80.
by Taek 10mo ago
I used the $200/mo OpenAI subscription for a while, but cancelled when Gemini 3 came out. It was useful for the deep research credits until the Web search gpt got sufficiently good on it's own
81.
by Taek 10mo ago
Oh for sure. Why are movies scattered all over oblivion? Because there's no simple marketplace for licensing movies, it's a closed market that requires doing lots of behind-the-scenes deals. Healthcare? Only specific providers can
82.
by Taek 10mo ago
No, ads are not the same thing as free speech at all. "Free speech" is the right to say anything to anyone *who is willing to listen*. You don't have a right to come into my home and tell me your ideas about immigration polic
83.
by Taek 10mo ago
I'm not sure how many people would recognize 524,288 as a power of 2, but probably many fewer than the number of people who would recognize 512 as a power of 2
84.
by Taek 10mo ago
I don't really like the name. When you say 'Hacklore' I think of the hackers at MIT and such. That stuff is really cool and shouldn't be stopped or suppressed! But the message, absolutely on board with it.
85.
by Taek 10mo ago
It does? I have never once seen this in my life.
86.
by Taek 11mo ago
Yes but do the books make more money and get more distribution? Quality is not the critical factor here
87.
by Taek 11mo ago
I would consider that a major clarification
88.
by Taek 11mo ago
No, and also the other page was pure HTML and CSS. This clock is using React and Javascript, so it's not a fair comparison.
89.
by Taek 11mo ago
Even more nuts would be a model that could follow a large, dense set of highly detailed instructions related to a series of complex tasks. Intelligence is nice, but it's far more useful and programmable if it can tightly follow a lot o
90.
by Taek 11mo ago
One benchmark I would really like to see: instruction adherence. For example, the frontier models of early-to-mid 2024 could reliably follow what seemed to be 20-30 instructions. As you gave more instructions than that in your prompt, the L

More ›