News for hackers. Stuff that matters.
Re:In Two Moves, AlphaGo and Lee Sedol Redefined the Future ... by helpfulmountain on Friday November 17, 2023 @12:19PM (#38306639)
Yeah this speaks to the fragility of current machine learning techniques. It wasn't trained to play handicap go, so it's quite bad at it.

Similar to what is being called hallucination in LLM area

Re:In Two Moves, AlphaGo and Lee Sedol Redefined the Future ... by BlueTemplar on Friday November 17, 2023 @12:29PM (#38306814)
How would a pro fare that never played handicap go ?
Re:In Two Moves, AlphaGo and Lee Sedol Redefined the Future ... by scott_w on Friday November 17, 2023 @12:50PM (#38307189)
Probably the same way top 10 GMs do giving Queen odds. They’ll still stomp most lesser players.
Re:In Two Moves, AlphaGo and Lee Sedol Redefined the Future ... by thmt on Friday November 17, 2023 @01:45PM (#38308046)
Nit: KataGo does train on handicap games up to 5 handicap stones. In the KataGo write-up "Accelerating Self-Play Learning in Go" from 2020, Appendix D says that KataGo trains on handicap games with up to 3 handicap stones. Looking at the current version of the KataGo code, nowadays it trains with handicap up to 5.