People often make decisions through "satisficing," gathering just enough information to make a satisfactory prediction of a likely outcome. A series of experimental games shows that people also employ satisficing to learn social rules and conventions. This finding offers new insight into social learning and reveals a key difference between how humans and LLMs make predictions. The premise of AI large language models is that any problem can be solved by vacuuming up as much information as possible, running it through probability models, and performing complex calculations to make predictions and come up with the optimal solution. Another premise behind LLMs is that they emulate the way human brains operate.

The economist Bryan Caplan was sure the artificial intelligence baked into ChatGPT wasn't as smart as it was cracked up to be. Caplan, of George Mason University in Virginia, seemed in a good position to judge. He has made a name for himself by placing bets on a range of newsworthy topics, from Donald Trump's electoral chances in 2016 to future US college attendance rates. And he nearly always wins, often by betting against predictions he views as hyperbolic. That was the case with wild claims about ChatGPT, the AI chatbot that's become a worldwide phenomenon.

No matter what companies say, AI is not going to solve the problem of content moderation online. It's a promise we've heard many times before, particularly from Facebook CEO Mark Zuckerberg, but experts say the technology is just not there -- and, in fact, may never be. Most social networks keep unwanted content off their platforms using a combination of automated filtering and human moderators. As The Verge revealed in a recent investigation, human moderators often work in highly stressful conditions. Employees have to click through hundreds of items of flagged content every day -- everything from murder to sexual abuse -- and then decide whether or not it violates a platform's rules, often working on tightly-controlled schedules and without adequate training or support.

The pace of content creation has never been faster. Each day, the world generates 2.5 quintillion bytes of data, and more than 90 percent of all data in existence has been produced since 2016. In the process, it is being hidden from search engines, locked away in multimedia formats that cannot be catalogued. A gold mine of information is waiting to be tapped if only the spoken word could be easily converted to text. Doing so by hand is time consuming and, often, prohibitively expensive.