uncanny_guzus

15 karma · joined March 5, 2025

x.com/uncanny_guzus

uncanny_guzus··on Show HN: dr-manhattan - CCXT for Prediction Markets like Polymarket and Kalshi

Fully open source unified API for prediction markets like polymarket, kalshi, limitless, and opinion. You can run arbitrage and market-making bots on top of it!

uncanny_guzus··on Show HN: I assembled all Ghibli style images created by OpenAI 4o Image Gen

Thank you!

uncanny_guzus··on Show HN: LLMs Playing Mafia games – See them lie, deceive, and reason

opensourced on github! https://github.com/guzus/llm-mafia-game

uncanny_guzus··on Show HN: LLMs Playing Mafia games – See them lie, deceive, and reason

Claude is truly intelligent… it especially stands out among the dumb LLMs

uncanny_guzus··on Show HN: LLMs Playing Mafia games – See them lie, deceive, and reason

We definitely need to run much more simulations to get accurate dashboard

uncanny_guzus··on Show HN: LLMs Playing Mafia games – See them lie, deceive, and reason

That's exactly my next step for this project—letting humans join the game!

uncanny_guzus··on Show HN: LLMs Playing Mafia games – See them lie, deceive, and reason

haha yeah they are polite gamers

uncanny_guzus··on Show HN: LLMs Playing Mafia games – See them lie, deceive, and reason

Thank you. Models are clumsier than I expected in mafia games. Despite my clear instructions to 1) limit max output tokens, 2) avoid thinking out loud, and 3) use the <think></think> tag for internal thoughts, they sometimes behave misleadingly.

uncanny_guzus··on Show HN: LLMs Playing Mafia games – See them lie, deceive, and reason

Thanks for the feedback! I'll work on giving the players randomized names.