Yes, what if it is combined with quantum computing?

[shapanwwuopi]

I think with this stuff you really need to be very careful about setting guardrails. But I also think that, at least for the next phase, it will take a lot of computing power to run these models and do many of the operations that they're talking about.

So I don't know. I think an interesting question is how much of this you can actually do on a pair of glasses or a phone versus a government or a company that owns an entire data center.

And, I mean, it can always get more efficient. Like you can start doing something and then maybe the next year you can do it 10 times more efficiently. But it’s definitely something that needs to be addressed in the next phase of the industry to make sure things run smoothly.

Host:

Mark Zuckerberg: Well, I'm not an expert in quantum student data computing. My understanding is that it's still a long way from being a very practical paradigm. I think Google just made some breakthroughs, but I think most people still think that's more than a decade away. So my guess is that we'll have very smart AI before then .

But yeah, I mean, look, I think these things have to be developed very thoughtfully, right? But I don't know; I still think we'd be better off overall in a world where this technology was evenly deployed. You know, it's... I guess there's another analogy that I've thought about: basically every piece of software that everybody uses has bugs and security holes. If you could go back a few years and know the security vulnerabilities we know about now, you as an individual could basically break into any system.

Artificial intelligence can do the same. It can detect and find vulnerabilities. So, how do we prevent artificial intelligence from getting out of control? I think part of it is the widespread deployment of artificial intelligence. That way, one system’s AI can defend against potentially problematic AI in another system. I think it's like... I don't think it's an AI war. That's not the case.