Hacker News Reader

AI players, however strong, sometimes leave glaring (to humans, at least when told) loopholes exploitable by humans.

See for instance amateur Go player armed with loophole consistently beat top Go playing software which otherwise easily crushes world champions.[1]

[1] https://www.ft.com/content/175e5314-a7f7-4741-a786-273219f43...

This sort of stuff has been a thing for balancing AAA games for a very long time (it was in production over a decade ago), and is kept surprisingly secret by those doing it.

One of the shocks of the games industry is in many respects it doesn’t learn from the wider world but in others it is quietly so far ahead that you can throw certain ML papers at former game devs to be met by confused faces as to why what is claimed is considered surprising.

That sounds interesting, can you give some examples? Is this limited to classical RL Q-learning type stuff for turn based games, or full AlphaStar-level agents for RTS games?

would love to have AI be able to play MtG sensibly, key word being sensibly!

Some would say that Microprose's "Magic: The Gathering 1997" aka "Shandalar" or "Manalink", from Sid Meyer, could already play decently well.

God that game was good.