There was an AI winter in the 80s but the resurgence seems to be correlated with the increase in available computing power. Now that Moore's law no longer holds, is AI hitting a saturation point? We're developing ASICS for DL but they can only go so far. Do we need a new computing paradigm to truly make new leaps? Do new maths need to be discovered?
I'm not sure that quantum computing is all it's cracked up to be
I work in deep learning acceleration and have talked with the Microsoft quantum team about this... we currently have NO idea whether we'll get speed up with quantum computing. QC is only useful for a very narrow set of tasks
Have any of you done this Future Of Work In The Age Of AI study. They release results every 3 months about the state of things. It would be good to lend your voice. https://research.cusjo.com/tfow
There's a limit to how far one can go without explanatory power. Definitely a step forward in the field but it's not what's happening in our head
Are you saying there's a computing power issue, an algorithmic issue, or something else? If a computing power issue (I'm assuming that's what you mean since you referenced Moore's law), my response to that is that we have some promising efforts going on with quantum computing right now.
Both. Yea quantum might work well for scaling since you get the product space. I don't think DL is the answer though.
Quantum is the only way forward