Notes

A question for the chatbot. I expect you remember first hearing about ChatGPT-3 in late 2022. You might also remember all the attention to AI and its potential, wildly differing future paths. The sensation was followed by a flood of money into the AI industry, which is on track to invest more than US$1 trillion in the technology over the next couple of years. Now, though, some investors, and others who follow the sector, are starting to have doubts about the business of artificial intelligence—and the direction it’s heading in. In late August, the prices of AI firms’ shares began to fall sharply. OpenAI, the maker of ChatGPT, will likely lose around $5 billion for the year. What’s happened?

Today, Daron Acemoglu explores the growing doubts about AI—and whether, with the architecture it has, the technology can ever learn to think and do the way humans can.

—Michael Bluhm