TechTakes

"Yeah I thought about going into civil engineering but the department of hustling really spoke to me y'know?"

The word "deranged" is getting a workout lately, ain't it?

I'm just curious how many hits you would get if you searched for '4 hour work week', as iirc that is where all these people stole the idea from. (well, not totally, the idea they are stealing is selling others the idea of the 4 hour work week, but I hope you get what I mean, 4 hour work weeks all the way down).

note that cwebber wrote the ActivityPub spec, btw

When people start going on about having nothing to hide usually it helps to point out how there's currently no legal way to have a movie or a series episode saved to your hard drive.

I suspect great overlap between nothing-to-hide-people and the people who watch the worst porn imaginable but think incognito mode is magic.

the marketing fucks and executive ghouls who came up with this meme (that used to surface every time I talked about wanting to de-Google) are also the ones who make a fuckton of money off of having a real-time firehose of personal data straight from the source, cause that’s by far what’s most valuable to advertisers and surveillance firms (but I repeat myself)

You just know Netflix's inbox is getting flooded with the absolute worst shit League of Legends players can come up with right now

Particularly hilarious at how thoroughly they're missing the point. The fact that it suggests illegal moves at all means that no matter how good it's openings are the scaling laws and emergent behaviors haven't magicked up an internal model of the game of Chess or even the state of the chess board it's working with. I feel like playing games is a particularly powerful example of this because the game rules provide a very clear structure to model and it's very obvious when that model doesn't exist.

I remember when several months (a year ago?) when the news got out that gpt-3.5-turbo-papillion-grumpalumpgus could play chess around ~1600 elo. I was skeptical the apparent skill wasn't just a hacked-on patch to stop folks from clowning on their models on xitter. Like if an LLM had just read the instructions of chess and started playing like a competent player, that would be genuinely impressive. But if what happened is they generated 10^12 synthetic games of chess played by stonk fish and used that to train the model- that ain't an emergent ability, that's just brute forcing chess. The fact that larger, open-source models that perform better on other benchmarks, still flail at chess is just a glaring red flag that something funky was going on w/ gpt-3.5-turbo-instruct to drive home the "eMeRgEnCe" narrative. I'd bet decent odds if you played with modified rules, (knights move a one space longer L shape, you cannot move a pawn 2 moves after it last moved, etc), gpt-3.5 would fuckin suck.

Edit: the author asks "why skill go down tho" on later models. Like isn't it obvious? At that moment of time, chess skills weren't a priority so the trillions of synthetic games weren't included in the training? Like this isn't that big of a mystery...? It's not like other NN haven't been trained to play chess...

LLMs sometimes struggle to give legal moves. In these experiments, I try 10 times and if there’s still no legal move, I just pick one at random.

uhh

Here are the results of these three models against Stockfish—a standard chess AI—on level 1, with a maximum of 0.01 seconds to make each move

I'm not a Chess person or familiar with Stockfish so take this with a grain of salt, but I found a few interesting things perusing the code / docs which I think makes useful context.

Skill Level

I assume "level" refers to Stockfish's Skill Level option.

If I mathed right, Stockfish roughly estimates Skill Level 1 to be around 1445 ELO (source). However it says "This Elo rating has been calibrated at a time control of 60s+0.6s" so it may be significantly lower here.

Skill Level affects the search depth (appears to use depth of 1 at Skill Level 1). It also enables MultiPV 4 to compute the four best principle variations and randomly pick from them (more randomly at lower skill levels).

Move Time & Hardware

This is all independent of move time. This author used a move time of 10 milliseconds (for stockfish, no mention on how much time the LLMs got). ... or at least they did if they accounted for the "Move Overhead" option defaulting to 10 milliseconds. If they left that at it's default then 10ms - 10ms = 0ms so 🤷‍♀️.

There is also no information about the hardware or number of threads they ran this one, which I feel is important information.

Evaluation Function

After the game was over, I calculated the score after each turn in “centipawns” where a pawn is worth 100 points, and ±1500 indicates a win or loss.

Stockfish's FAQ mentions that they have gone beyond centipawns for evaluating positions, because it's strong enough that material advantage is much less relevant than it used to be. I assume it doesn't really matter at level 1 with ~0 seconds to produce moves though.

Still since the author has Stockfish handy anyway, it'd be interesting to use it in it's not handicapped form to evaluate who won.

@gerikson @BlueMonday1984 the only analysis of computer chess anybody needs https://youtu.be/DpXy041BIlA

When the reporter entered the confessional, AI Jesus warned, “Do not disclose personal information under any circumstances. Use this service at your own risk.

Do not worry my child, for everything you say in this hallowed chamber is between you, AI Jesus, and the army of contractors OpenAI hires to evaluate the quality of their LLM output.