Good Start Labs trained an AI on a railroad game — and one version improved at financial research. The difference was the training design.

2 points•james_mir•10 days ago•1 comment•

1 comment

james_mir10 days ago
I think they kind of cheat within the given example (the model performed well at stock market operations, but the game they used to train it actually kind of simulates stock market operations ). However, it may be interesting or curious for board game lovers who works with AI.

Read the full thread on Hacker News →

Related stories