Tim Sweeney
speaker
640 appearances
1 recordings
1 series
first heard Apr 2025
last heard Apr 2025
Tim Sweeney’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
When we have much, much, much more nuanced control of it, and artists can go in and just, you know, like, you know, let's enhance this object, do this, do that, do that, they'll feel it's, you know, like some of the tools that exist in Photoshop, which are in some ways compared to a paintbrush are superpowers already.
AI will come to feel like that too, and will increasingly serve creators creating and enhancing a work in a way that feels just a natural extension of their own, you know, their own bodies and minds.
Yeah, there's a lot of complicated trends underway. It can be hard to break them down and distinguish them. I think a lot of people, like the theories that get the biggest traction on social media often don't capture the real underlying motive forces at play there. I think AI involved in code production will probably create a net benefit for the need for humanity to be involved in coding.
It may change parts of jobs. I don't think it's going to obsolete anybody who's willing to learn new ways of doing things. And it's always been this way. And I think that there's also a lot of overhype in AI. AI is really great at spewing out code that does something that a million GitHub repositories already do because it's kind of learned the underlying pattern.
It's notoriously hard to get to do something new that hasn't been done before, especially when it's a complex task. And the bigger amount of code you ask AI for, the more it leaves you with just a mess of code that sort of works, right? And that's the problem with code. It's like 99% works, but the 1% might be harder to get to 100% with AI than with hand coding.
And everybody who's looking at this topic should actually try using the coding assistance on hard problems and see how they do there.
The topic of boilerplate code is an interesting one because the mere existence of boilerplate code is a failure of programming language and of the idea of creating software modules, right? You ask AI to create a sorting function, great. Now you have another sorting function that might be buggy alongside the million others that different people have written.
It would be better to have a sorting function that's been written and tested and optimized and everybody relies on it and More modular software, I think, will actually reduce the opportunity of AI because people doing programming work will largely be solving unique problems. They're actually hard problems in themselves and not just connecting other widgets.
We're going to get shockingly close over the coming years, certainly less than 20 years. If you look at the progress, what areas where we have achieved total photorealism and what areas where we fall short, we're getting very close in all non-human interactions you see in the world, walking through a jungle or a city. All the lighting, it's very close, and that might be just a few years away.
But then all the problems that involve humans, human dialogue and intent, have a much, much, much higher bar that they need to meet to satisfy our brains and convince us that they're realistic or real. And I think that's going to be the primary challenge of graphics development and simulation development over the coming decade.
Yeah. I was asked about this about 10 years ago, and I said that even if you gave us an infinite amount of computing power, we couldn't simulate realistic humans because we simply don't have the algorithms. We have no idea how to simulate human intelligence. And that was absolutely the case then, but... it's not really true anymore.
You know, what we're seeing with generative text AI is not only at a level that you could say that it's actually doing a pretty good job of simulating a human, at least humans at the text level, not at the emotional level yet, but at least at the level of words spoken and find more and more ways of training on more and more scenarios that, you know, you might have a very, very compelling human simulation going on in the next, um,
five years even. I'm not saying it's a good idea, but I think the arc of the technology is inextricably heading in that way, and it's heading at a shocking, shocking rate.
Yeah, absolutely.
Well, I don't think that these questions are necessarily unanswerable. I think I'd like to see more actual effort to ascertain what is the underlying mechanism of the universe. And I don't think we're here for no reason at all. I think the world's a pretty cool place.
And the fact that we can exist and, you know, the laws of physics and especially the standard model of physics and all the parameters that lead to these atoms and life evolving in the presence of thermodynamic gradients, that's really cool. And I think it's a worthy field to study more about that holistically.
The question of are we living in a simulation ourselves always boils down to, well, if we are living in a simulation, what are they living in? Because at some point there has to be some base reality. Or one of the philosophical theories that was put forth seriously was that There is no physical reality.
If you have a system of equations, such as the laws of physics, then all possible evolutions of dynamical systems under those equations kind of have a physical reality. So we just are kind of a manifestation of laws of math rather than needing an actual universe around us. I don't know. I like dabbling in that philosophy.
And as we get CAI becoming smarter and smarter and we get closer and closer to really capturing the full laws of physics, these questions become quite a lot more compelling.
So the whole idea of nested simulations, perhaps... Given sufficiently advanced technology, it's kind of mooted such that if you wanted to simulate another reality, you're kind of just actually creating the reality. You're doing quantum mechanical operations that would produce the same result anyway, and you're running them at full performance. So it's not really a nested simulation.
Showing 301–320 of 640 · page 16 of 32
← Previous
Next →