It's all deep learning, making it the exact opposite of a Cambrian event.<p>Has anyone has tried searching for new basic operations, below the level of neural networks? We've been using these methods for years, and I doubt the first major breakthrough in ML is the most optimal method possible.<p>Consider the extreme case of searching over all mathematical and logical operations to see if something really novel can be discovered.<p>How feasible would this be?<p>>Tying it together, a big diversification is likely coming<p>I'd be interested in knowing what signs, if any, currently point to that.<p>What I see is a coalescing around the current paradigm. And considering that the hardware side led by NVIDIA is catering to this paradigm, I don't see us breaking out of this local optimum any time soon.
Hey hn, this is my first blog post. The community here inspired me to finally try it out.<p>I would really appreciate any feedback on clarity, or anything that causes good or bad gut reactions.
I understand it seems clever to frame AI trends in an evolutionary setting.<p>But no one was there in the Cambrian period. We have no idea what the circumstances really were. I'm not dismissing historical sources. Just saying the timescale and comparative lack of sources equate to a lot of unreliability.<p>The real question is how much AI is a step up in tooling. The fact that humans develop tools using their intelligence is well documented. And doesn't require a detour into evolutionary theory.<p>The society with the best tools for the problems it faces has a competitive advantage.
I think big jump in ai 'evolution' will be the competition part. Like DeepMind with alphastar. They got results by having different models compete, in multiple rounds to more 'evolve' towards a winning strategy. So far it seems like the news is with these single big models, but evolution kicks in if they start doing what is needed to survive.
IMO the only problem we should be focusing AI on right now is having it figure out how to make more advanced and powerful AI. Alas, this is beyond AI’s potential right now. Humans are still very much needed.