23
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
this post was submitted on 03 Nov 2025
23 points (100.0% liked)
TechTakes
2425 readers
82 users here now
Big brain tech dude got yet another clueless take over at HackerNews etc? Here's the place to vent. Orange site, VC foolishness, all welcome.
This is not debate club. Unless it’s amusing debate.
For actually-good tech, you want our NotAwfulTech community
founded 2 years ago
MODERATORS
So, today in AI hype, we are going back to chess engines!
Ethan pumping AI-2027 author Daniel K here, so you know this has been "ThOrOuGHly ReSeARcHeD" (tm)
Taking it at face value, I thought this was quite shocking! Beating a super GM with queen odds seems impossible for the best engines that I know of!! But the first * here is that the chart presented is not classical format. Still, QRR odds beating 1600 players seems very strange, even if weird time odds shenanigans are happening. So I tried this myself and to my surprise, I went 3-0 against Lc0 in different odds QRR, QR, QN, which now means according to this absolutely laughable chart that I am comparable to a 2200+ player!
(Spoiler: I am very much NOT a 2200 player... or a 2000 player... or a 1600 player)
And to my complete lack of surprise, this chart crime originated in a LW post creator commenting here w/ "pls do not share this without context, I think the data might be flawed" due to small sample size for higher elos and also the fact that people are probably playing until they get their first win and then stopping.
Luckily absolute garbage methodologies will not stop Daniel K from sharing the latest in Chess engine news.
But wait, why are LWers obsessed with the latest Chess engine results? Ofc its because they want to make some point about AI escaping human control even if humans start with a material advantage. We are going back to Legacy Yud posting with this one my friends. Applying RL to chess is a straight shot to applying RL to skynet to checkmate humanity. You have been warned!
LW link below if anyone wants to stare into the abyss.
https://www.lesswrong.com/posts/eQvNBwaxyqQ5GAdyx/some-data-from-leelapieceodds
One of the core beliefs of rationalism is that Intelligence™ is the sole determinant of outcomes, overriding resource imbalances, structural factors, or even just plain old luck. For example, since Elon Musk is so rich, that must be because he is very Intelligent™, despite all of the demonstrably idiotic things he has said over the years. So, even in an artificial scenario like chess, they cannot accept the fact that no amount of Intelligence™ can make up for a large material imbalance between the players.
There was a sneer two years ago about this exact question. I can't blame the rationalists though. The concept of using external sources outside of their bubble is quite unfamiliar to them.
🪦👨🏼➡️👴🏼
Will never be able to understand why these mfs don’t see this as the unga bunga stupid ass caveman belief that it is.
cos it implies that my overvalued salary as an IT monkey fo parasite companies of no social value is not because I sold my soul to capital owners, it's because I've always been a special little boy who got gold stars in school
@swlabr @lagrangeinterpolator Rat calvinism lol
I was wondering why Eliezer picked chess of all things in his latest "parable". Even among the lesswrong community, chess playing as a useful analogy for general intelligence has been picked apart. But seeing that this is recent half-assed lesswrong research, that would explain the renewed interest in it.