Jason Antic

Exactly. To me and I’d bet most others in the field machine learning is just a powerful tool to enable new things that weren’t previously doable. Acting like it’s a wasted effort because you don’t think it’ll lead to AGI is just silly. It’s not just about AGI.https://twitter.com/tdietterich/status/1203189592712568832 …

Ok, but I can't really see any reason why deep learning won't (some day in the "distant" future) lead to something that can be called human-like intelligence. CNN already approximate human visual cortex, reinforcement learning already produces human-like strategies to win games.

And deep learning is just 7 years old, imagine the breadth it'll have when it will become a "mature" discipline with solid software engineering principles embedded into it and universally recognized.

I don’t think that it is a proper characterization to say DL is around 7 years old. Some of the core ideas are at least 25-30 years old. However lack of data, computing power and organized, large-scale sharing (GitHub) among practitioners restricted its scope earlier.