takedown of OpenAI’s apparent success at ‘Dota 2’https://apple.news/A45CaowNkR7GQ8x92qTIbdg …
-
The first, that it should use vision rather than an API, is essentially irrelevant imo. What the API provides could be produced by a vision model, the primary reason to avoid it is that it'd be a waste of training time _plus_ rendering the DotA client would be insane overhead.
-
-
They've noted this forever and have an actual interactive display of the information the bot receives from the API on their blog so I'm just confused as to what surprised you on it? Expand on why this is huge to you?
-
Honestly I hadn’t followed, but it does seem like a step back relative to DQN, ducking some of the hardest problems like knowing what entities exist and how to segment them etc.
- 1 more reply
New conversation -
-
-
Shuldn't such algorithms compete with pre-programmed bots where rules of the game are hardcoded. That wud be a fair play. But yes, that wud not make a hyping headline. Hype around intelligence without understanding really amazes me.
-
Hardcoded/pre-programmed bots were thoroughly defeated by
@OpenAI Five. One of the DotA2 players in the most recent test match had joked for years that they'd never be beaten by a DotA2 bot. This is an advance even if it has "shortcuts" (which I don't agree are) due to that. - 7 more replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.