Hacker News Reader

Comment by pitchups | Hacker News Reader

Completely blown away after watching that video by the sheer range of tasks this AI can accomplish on its own. Learning to walk, balance a stick, using tools, co-operating, playing pong, and...even buying stocks low and selling high! And all of this without being programmed or given a goal to do so. If true this could be a huge breakthrough indeed. Would really like to learn more about the core technology and will be following this more closely.

View on HN

I'm far from being blown away. Researchers in AI have demonstrated AI systems with a "blank mind" that can learn by interacting with its environment given some reinforcement learning. They often operate in simple toy examples, but once confronted with the real, complex, messy reality, they fail. I found mostly two reasons: Either their mechanism is too simple and can't solve other than simple problems or it's (computationally) too complex and doesn't scale.

So I'm not saying this approach is bad, but we have yet to see whether it does more than simple examples.

From what I can gather the difference here seems to be the claim that a simpler, more fundamental principle may explain the motivation behind all intelligent behavior. A tall claim no doubt - so I agree that it needs to be validated with more complex examples.

Where do you see it learning to walk? You can't equate the inverted pendulum problem with upright walking. It is neat if it can choose it's own reference level ("goal") for a classical control problem -- but that's a long way from "learning to walk".

Granted that their model for upright walking is very basic. But if their basic premise is sound and valid and taking their claims at face value, it could potentially explain how more complex behaviors like walking may emerge from the simpler examples in their demo.

Reply on news.ycombinator.com