2024-12-20 12:02:53
hmm it does not pass my Fortune's Foundation step 1 test, only o1 does (4 correct steps) so far
2024-12-13 13:19:56
Re @yicaichina @Baidu_Inc @GeelyGroup This dude was my manager about 10 yrs ago, kind of funny that the mismanagement rumors match my impression of him
2024-12-11 08:05:47
I hope the person who made this won't be fired, they've been through a lot since hardcore took over
2024-12-06 05:06:39
Re it's actually a perfect setup for MCTS if that's what it really uses, down a few steps, evaluate and go back
2024-12-06 05:04:17
tried this on o1, it went as far as 4 steps:
"Move 11 from Queue 7 to Queue 3 after 12
Move 10 from Queue 11 to Queue 3 after 11
Move 5/L from Queue 5 to Queue 6
Move 17 from Queue 5 to Queue 8 after 18"
after that it hallucinates a bit and gives up
2024-11-25 04:53:53
https://www.josherich.me/short/fortune-foundation-test Fortune's Foundation is hard but getting the first step right should be easy, which is not something LLM can do
