The AI Singularity: How Artificial Intelligence Could Transform the Economy

In a recent episode of the Acquirers Podcast, Travis and Carlisle discussed the AI singularity and how artificial intelligence could change the economy. Here is part of their conversation:

[Tobias] On my other podcast, I talked about something that has made me curious about the singularity. No one really knows exactly what the singularity means, but I have been interested in the idea and have done some research. There is a short science fiction story from 1965 written by a statistician named Good.

The story was called "The Ultra-Intelligent Machine." It is interesting how much that name sounds like today's idea of superintelligence. The idea was that once a machine could build a machine smarter than itself, the process could keep repeating. Each new machine could build an even smarter one. This could happen at machine speed instead of human speed. At some point, machines would become smarter than humans and continue getting smarter. Based on what OpenAI and Anthropic have been saying, they may already be doing this, or they may be very close. By very close, I mean this year or very soon. They may soon build a machine that can repeatedly improve its own machine-learning model.

The machine would teach itself how to improve its model. To me, that means we have reached the singularity. Another person who discussed this idea was someone whose name I am not sure how to pronounce. It was Wernher Vinge, or something like that.

He believed that once this process of repeated improvement began speeding up, we would reach the singularity. He said we would not be able to understand or predict what comes next because humans have never faced an intelligence smarter than all of us together. One possible result would be more scientific discoveries happening faster and faster. I think we have already started seeing this, including in mathematics, where some problems have remained unsolved for a very long time.

I have a hobby-level interest in mathematics. About 20 years ago, I read a book called Fermat's Last Theorem, which explained how the problem had been solved. Mathematics is a very unusual field.

Mathematicians use clever and sometimes unusual methods to create proofs and solve problems. OpenAI has solved several mathematical problems that had remained unsolved for a long time. The company said it solved them in unusual ways that people might not have expected.

The AI made connections that people might not have made on their own. So we are reaching a point where we are seeing scientific breakthroughs. Researchers are also working out the genetic code. Anything that depends heavily on mathematics seems especially suited to this kind of AI.

We are seeing more discoveries, which may be signs of the singularity. I think we are getting close to a point where we may already be in the singularity. These things are happening now. I am not sure that it actually makes much difference.

We have been through major changes before. Technology has changed our lives in many ways, including through the internet and computers. At one time, hard drive storage capacity was growing very quickly. We have always come through these changes, and eventually the new technology becomes completely normal.

Electricity was once an amazing new technology. Now, hardly anyone thinks twice about turning on a light. I think AI will become normal in the same way. Eventually, probably everything will have intelligence built into it.

[Tim] That is certainly the optimistic view. Honestly, I think that is the way to approach it. I think that is how you embrace it.