Podcast: Play in new window | Download
Subscribe: RSS
AlphaGo is back. Patrick talks about its latest accomplishments, its future, and other topics in game advances.
Our first AlphaGo episode (a very early one, sorry about the audio quality)
News of the most recent AlphaGo adventures
Steph Curry’s training is on a new level
Eneasz’s favorite type of tool-assist challenges – Item Abuse.
Sam Harris speaks with Tristan Harris about the arms race for human attention, the ethics of persuasion, and more
Not sure if it’s a useful feedback, but I found the tendency of the guest to constantly jump the discussion topic and generate loosely related technobabble (mostly happened in the latter part of the show) rather annoying.
I feel like you guys missed the point of the fighting AI. Martial Arts are a physically winnable task, If I just put a wood chipper out in the ring it will kill the humans. The test of mechanics is obviously not the thing we need the AI for. If we limited the AI’s inputs and outputs to a reasonable human level and fed it techniques from every martial art and the physical specs for itself and an array of opponents it should be able to think long enough that it could create an style that is the optimal human unarmed combat style.
Like maybe you would have to slow it’s reaction times down a whole lot to make it something that could be implemented by humans but that is fine as long as you are looking at it as a learning/teaching activity rather than simply seeing if a humaniform robot can kill Chuck Norris.
Pingback: 79 – CyberChristianity | The Bayesian Conspiracy