Categoría
🗞
NoticiasTranscripción
00:01And joining us to moderate the discussion, the founder and managing partner of Conviction
00:07and a good friend, Saragot.
00:09As we started hitting record speeds this year.
00:13So how are you continuing to drive this innovation forward?
00:16And I think we'll be at the point next year where you can not only use a system to sort
00:21of automate some business processes or build these new products and services, but you can
00:26really say, I have this hugely important problem in my business.
00:30I will throw a ton of compute at it if you can solve it.
00:34And the models will be able to go figure out things that teams of people on their own can't do.
00:39And the companies that have gotten experience with these models are well positioned for a
00:44world where they can say, okay, you know, AI system, whatever.
00:48Go, you know, like redo my most critical project.
00:52And here's a ton of compute, think really hard just to figure out the answer.
00:55People who are ready for that, I think we'll have another big step change next year.
00:59What advice would you have for people from financial services or healthcare that are working their way through?
01:08Um, I would bet next year that in some limited cases, at least in some small ways, we start to see agents
01:17that can help us discover new knowledge or can figure out solutions to business problems that are kind of very non-trivial.
01:24Um, right now it's, it's very much in the category of, okay, if you've got some like repetitive cognitive work, we can automate it at a kind of a low level on a short time horizon.
01:35And as that expands to longer time horizons and higher and higher levels, you know, at some point you can add scientists, uh, an AI agent that can go discover new science.
01:45And that will be kind of a significant moment in the world.
01:55The thing that matters is the rate of progress that we have seen year over year over the last five years
02:01should continue for at least the next five, probably well beyond that, but hard to say.
02:06And whether you declare the AGI victory in 24 or 26 or 28, um, and whether you declare the super intelligence victory in 28 or 30 or 32,
02:18is way less important than this one long, beautiful, shockingly smooth exponential.
02:24Um, all of that said, to me, a system that can either autonomously discover new science
02:34or be such an incredible tool for people that are right of scientific discovery in the world like quadruples or something.
02:41Um, that would, that would satisfy any test I had imagined for an AGI.
02:46But already you could see greatness for me when you saw this problem called, um, abstractive summarization.
02:54The models over the next year or two years are going to be quite breathtaking.
02:59Um, really, there's a lot of progress ahead of us, a lot of improvement to come.
03:06And like we have seen in the previous big jumps, you know, from GPT-3 to GPT-4,
03:11businesses can just do things that totally were impossible with the previous generation of models.
03:17And, and so what an enterprise will be able to do, we talked about this a little bit, but just like,
03:23give it your hardest problem.
03:25If you're a chip design company, say, go design me a better chip than I could have possibly had before.
03:30Um, if you're a biotech company trying to cure some disease, say, just go work on this for me.
03:36Like, that's not so far away.
03:38I could be more energized about this.
04:01So, if you're a biotech company, do you want to join the light and throw it in here?
04:03Are you going to have a biotech company that has the diabetes system?
04:04Are you going to be able to drop out with you?
04:05Um, that's not so far away.
04:06Don't be able to get any anything.
04:07Okay?
04:08Okay.
04:09Yeah.
04:10Okay, let's see.
04:11You might have a biotech company in just straight ahead of the first time.
04:13But we're gonna, you know, don't worry.
04:14We need to be able to get a biotech company.
04:15Let's see.
04:16Okay.
04:17Just like, we've been able to get a biotech company.
04:21In some time, please get any new people to find a biotech.