AI beats world's best Dota player

Question

https://www.theverge.com/2017/8/11/16137388/dota-2-dendi-open-ai-elon-musk

This reminds me of the old Quake 3 aim bots that were made using neural nets. The bots eventually learned after a few months training that not playing at all was the best option to survive. Maybe they learned something that humans didn't - that peace is more viable than war?

Answer 1

Thanks

Answer 2

Interesting. Bots learning from history and experience will not repeat similar patterns to trigger mistakes. They stick to statistical fact at decision-making. Humans on the other hand often cloud themselves with other delusions and take unnecessary risks at attempting the harmful.

Answer 3

*people cried when deep blue beat the then #1 in chess*

Answer 4

https://code.sololearn.com/WEfflJx3Sus8/?ref=app

Answer 5

amazing sf on 1v1 mode