TIME100 AI Scientist: The Next Era of AI Has Already Started | Richard Socher

**Richard Socher** (0:00)
Every domain we can verify or simulate, AI will get superhuman in the next few years. It's just like no doubt.

**Marina Mogilko** (0:05)
This is Richard Socher, inventor of Prompt Engineering. Now he's the founder and CEO of Recursive Superintelligence, a company that just raised $650 million at a $4.65 billion valuation to chase one goal. Build Superintelligence, an AI that improves itself and pushes beyond human capabilities. How does my workflow change when you reach your goal with your company?

**Richard Socher** (0:30)
You could actually start to work less and less, and you could have much more abundance. We could live much longer.

**Marina Mogilko** (0:36)
If you're saying AI is almost here, what is your timeline for Superintelligence?

**Richard Socher** (0:40)
I think we will actually get to the loops of Recursive Self-Improving Superintelligence within two years.

**Marina Mogilko** (0:46)
So imagine we reach Superintelligence today. What would be your first question to that Superintelligence?

**Richard Socher** (0:52)
How to...

**Marina Mogilko** (0:54)
Now you're building this company that just raised 650 million at 4.65 billion valuation, building self-improving AI. I'm not a researcher. Can you explain what that means?

**Richard Socher** (1:05)
So right now, you can think about the scientific method. Like people having ideas, they're implementing those ideas and then they validate if they made any sense and if they are correct. We want to apply this scientific method to AI itself.
So allowing AI to understand its own shortcomings and then fix those shortcomings and hence do research on itself. And so when we talk about recursive self-improvement, we mean that the AI builds a new version. The output of that AI is a new version of itself that's different. And then you can loop that onto itself.

**Marina Mogilko** (1:42)
Does that mean you train it on very little data and then it acquires data that it needs for self-improvement? How does that initial stage work?

**Richard Socher** (1:50)
You kind of stand on the shoulders of giants, somewhat similar to evolution, where lots of species like our own species started from other apes and monkeys and other precursors to the Homo sapiens. And similarly, we will stand on the shoulders of the existing giants right now. You can use large language models, you can use world models. All of these are pieces to this overall intelligence.

**Marina Mogilko** (2:17)
Once you're on the market, can you explain to me as an end consumer, how would that change my process? So now I have like, okay, there's a chat bot that I can talk to, there are projects and skills I can build with Claude, there are agents I can build. How does my workload change when you reach your goal with your company?

**Richard Socher** (2:35)
So there will be different gradations of those goals over time, right? And when we have true super intelligence, all you will have to do is give it the right rewards, the right goals and it will automatically create a lot of the processes to achieve those goals. In the current state of the world, AI has sort of very spiky capabilities. It can be like extremely good at this one type of math, but then not very good still at some common sense reasoning and things like that. We believe that our approach of open-endedness where you allow the AI to kind of evolve in this very open-ended search process that is much more akin to sort of biological or technological evolution or even cultural evolution, the AI will become more smooth around its capabilities. But until that happens, what you see right now, if you give a reward is what we call reward hacking. And you can give you a concrete example. Let's say you're a company and you told this kind of very intelligent AI that it should improve your customer satisfaction scores, your CSAT scores in your service centers. The AI will say, easy, I'll just create a million bots.
They hammer my phone lines and then give a five out of five rating at the end. And you're like, no, that's not the reward I was thinking about when I told you that. It should be with real people. But then the AI says, easy, I'll just give everyone a thousand dollar gift certificate at the end of every call and I get a five out of five rating, even though I didn't solve anything, right? So these are all examples of reward hacks. And that will also be a new kind of job that we're going to see is like people as the AI doesn't have this sort of common sense understanding yet and find sort of almost in a weird way, like an autistic person, like just like, this is the thing you said you want it, but you didn't explicitly define all the edge cases that you didn't want. As we're getting closer and closer to that, we have to still think about these rewards and the reward engineering problems that may come.