Menu
Sign In Search Podcasts Charts People & Topics Add Podcast API Pricing

Sean Carroll

👤 Person
10759 total appearances

Appearances Over Time

Podcast Appearances

But when we see individual behavior, we find that very judgeable in a way that the invisible corporate or systemic or government, for that matter, behavior behind the scenes does. tends to be a little bit more invisible, going back to the shooting the United CEO kind of thing, right?

But when we see individual behavior, we find that very judgeable in a way that the invisible corporate or systemic or government, for that matter, behavior behind the scenes does. tends to be a little bit more invisible, going back to the shooting the United CEO kind of thing, right?

I mean, how do you relate, how do you compare the crime of shooting an individual person on the street to the crime of denying a whole bunch of people health care, right? One is much more visceral. and visible.

I mean, how do you relate, how do you compare the crime of shooting an individual person on the street to the crime of denying a whole bunch of people health care, right? One is much more visceral. and visible.

In the case of AIs and large language models, I do think that there is some tribalism going on, some choosing of sides, because there's a whole bunch of reasons why you might be either skeptical or outright hostile to AIs. One reason is the way that they have been deployed by corporations, right? Google searches or Microsoft products are now full of AI when nobody asked for it.

In the case of AIs and large language models, I do think that there is some tribalism going on, some choosing of sides, because there's a whole bunch of reasons why you might be either skeptical or outright hostile to AIs. One reason is the way that they have been deployed by corporations, right? Google searches or Microsoft products are now full of AI when nobody asked for it.

Another reason is if you're an artist or a writer or for that matter, any other number of professions that have their livelihoods in danger of being stolen away by AI. Or like you say, there's the worry about resource usage in various ways. And these are very different objections, right? Like these are not the same objection, but we get aligned.

Another reason is if you're an artist or a writer or for that matter, any other number of professions that have their livelihoods in danger of being stolen away by AI. Or like you say, there's the worry about resource usage in various ways. And these are very different objections, right? Like these are not the same objection, but we get aligned.

We care about one of these objections and we start relying on other ones just to build our case because we've already decided what side we're on. And then we just collect evidence to make our side happy. I think that the question of using up resources is a very tricky one. I mean, it is clear that As a whole, on the aggregate, AI and other – the label AI is a little not exactly appropriate here.

We care about one of these objections and we start relying on other ones just to build our case because we've already decided what side we're on. And then we just collect evidence to make our side happy. I think that the question of using up resources is a very tricky one. I mean, it is clear that As a whole, on the aggregate, AI and other – the label AI is a little not exactly appropriate here.

People are using computing power for all sorts of things. But computing power broadly construed uses up an enormous amount of resources, uses up an enormous amount of electricity and water and things like that. I've seen plenty of people complain about this.

People are using computing power for all sorts of things. But computing power broadly construed uses up an enormous amount of resources, uses up an enormous amount of electricity and water and things like that. I've seen plenty of people complain about this.

I have not seen much of a sensible breakdown of how much an individual attempt to query ChatGPT about what is the best way to make a noodle dish with ground pork or something like that. Does that really... use up an enormous amount of electricity compared to something else? I honestly don't know. I'm saying that I haven't seen it carefully compared to other things that we do, right?

I have not seen much of a sensible breakdown of how much an individual attempt to query ChatGPT about what is the best way to make a noodle dish with ground pork or something like that. Does that really... use up an enormous amount of electricity compared to something else? I honestly don't know. I'm saying that I haven't seen it carefully compared to other things that we do, right?

This is just, again, another very basic human flaw. We like to think about total amounts rather than rates, but the rate is what matters. What is the rate of an individual person's use of AI in terms of electricity and water consumption versus other uses of AI?

This is just, again, another very basic human flaw. We like to think about total amounts rather than rates, but the rate is what matters. What is the rate of an individual person's use of AI in terms of electricity and water consumption versus other uses of AI?

You know, people in California work very hard to not run their faucets too much because there's a water shortage, but the enormous fraction of water is used by agriculture, not by people running their faucets.

You know, people in California work very hard to not run their faucets too much because there's a water shortage, but the enormous fraction of water is used by agriculture, not by people running their faucets.

That's not to say that people should waste water running their faucets, but it's just very hard to get accurate information about the actual scope of the problem versus the thing that you feel you personally can have a handle over. I think that the real danger of AI is neither one of these things, is neither power usage or AGI taking over and ruling us.

That's not to say that people should waste water running their faucets, but it's just very hard to get accurate information about the actual scope of the problem versus the thing that you feel you personally can have a handle over. I think that the real danger of AI is neither one of these things, is neither power usage or AGI taking over and ruling us.