News for hackers. Stuff that matters.
Re:Show HN: Watch and Play Poker with LLMs by bionsystem on Sunday January 11, 2026 @09:52AM (#46576202)
I just watched for 5 min and no they don't play very well. Deepseek squeezed with K4o against CO open and BTN call with full stacks. Grok 3b AI with 25bb in the button with Q4s. Those are very far from optimal play which is well known since solvers. I wonder how they've been trained.
Re:Show HN: Watch and Play Poker with LLMs by ryandrake on Sunday January 11, 2026 @10:28AM (#46576530)
You and OP are agreeing: "Better than most human players" is quite possibly the lowest skill bar in poker.
Re:Show HN: Watch and Play Poker with LLMs by sejje on Sunday January 11, 2026 @07:41PM (#46582238)
Considering a squeeze puts deepseek ahead of most human players. Maybe not an optimal squeeze, but most human players flat if they're going to play.

Grok's play is obviously bad, no solver needed. I wonder what he said in the "log" where you can see their thoughts. I guess he can hopelessly be trying to rep AA--again, I've seen worse thought processes every time I've ever played in a casino.

I really think GPT, at least, would win at a live casino. Possibly the other bots as well. The humans are that bad. Poker is complex.