Sarah Walker
speaker
510 appearances
1 recordings
1 series
first heard Jun 2024
last heard Jun 2024
Sarah Walker’s voice in public audio — every appearance, attributed to the second.
Trend
recordings per month · last 12 monthsNo recordings in the last 12 months.Older appearances are listed below; set an alert to hear about the next one.
Appearances
That's a highly dynamic crystal because it's a crystallization in time of this massive abstract structure that's evolved over human history and is now put into a small device.
I think there's not, I mean it very purposefully because a particular instantiation of a language model trained on a particular data set becomes a crystal of the language at that time it was trained. But obviously we're iterating with the technology and evolving it.
That's right.
Right. It's a societal level technology, right? We've actually put collective intelligence in a box.
I actually, I don't like the sort of language we use around that. And I think the language really matters. So I don't know how to talk about how much smarter one human is than another, right? Like usually we talk about abilities or particular talents someone has.
And, you know, going back to, you know, David Joyce's idea of universal explainers, like, you know, adopting the view that, you know, we're the first, you know, kinds of structures our biosphere has built that can understand the rest of reality. We have this universal comprehension capability. You know, he makes an argument that, you
Basically, we're the first things that actually are capable of understanding anything. It doesn't matter. It doesn't mean an individual understands everything, but like we have that capability. And so there's not a difference between that and what people talk about with AGI.
In some sense, AGI is a universal explainer, but it might be that a computer is much more efficient at doing, I don't know, prime factorization or something than a human is, but it doesn't mean that it's necessarily smarter or has a broader reach of the kind of things that can understand than a human does. And so I think we really have to think about, is it a level shift
Or is it we're enhancing certain kinds of capabilities humans have in the same way that we can enhance eyesight by making telescopes and microscopes? Are we enhancing capabilities we have into technologies and the entire global ecosystem is getting more intelligent? Or is it really that we're building some super machine in a box that's going to be smart and kill everybody?
That sounds like a science, it's not even a science fiction narrative, it's a bad science fiction narrative. I just don't think it's actually accurate to any of the technologies we're building or the way that we should be describing them. It's not even how we should be describing ourselves.
Well, these are human questions, right? I don't think they're necessarily questions that we're going to outsource to an artificial intelligence. I think what is happening and will continue to happen is there's a co-evolution between humans and technology that's happening. And we're coexisting in this ecosystem right now, and we're maintaining a lot of the balance. And for
The balance to shift to the technology would require some very bad human actors, which is a real risk, or some sort of... I don't know, some sort of dynamic that favors... Like, I just don't know how that plays out without human agency actually trying to put it in that direction.
So, like, I think the things that are, you know, terrifying are, you know, the ideas of deep fakes or, you know, like... You know, all the kinds of issues that become legal issues about artificial intelligence technologies and using them to control weapons or using them for, you know, like child pornography or. Or like faking out that someone's loved one was kidnapped or killed.
There's all kinds of things that are super scary in this landscape and all kinds of new legislation needs to be built and all kinds of guardrails on the technology to make sure that people don't abuse it need to be built. And that needs to happen. And I think one function of sort of the artificial intelligence doomsday space
sort of part of our culture right now is it's sort of our immune response to knowing that's coming. And we're overscaring ourselves. So we try to act more quickly, which is good. But I just, you know, it's about the words that we use versus the actual things happening behind the words.
I think one thing that's good is when people are talking about things different ways, it makes us think about them. And also when things are existentially threatening and We want to pay attention to those.
But the ways that they're existentially threatening and the ways that we're experiencing existential trauma, I don't think that we're really going to understand for another century or two, if ever. And I certainly think they're not the way that we're describing them now.
Yeah, it's just what we do to ourselves. Yeah.
Yeah, for sure.
Well, I think people are going to have romantic relationships with them. And I also think that some people will be convinced already that they're conscious. But I think in order... What does it take to convince... convince people that something is conscious, I think that we actually have to have an idea of what we're talking about.
Showing 401–420 of 510 · page 21 of 26
← Previous
Next →