Large language models (LLMs) have demonstrated remarkable capabilities in natural language understanding and generation across various domains, including medicine. We present a comprehensive evaluation of GPT-4, a state-of-the-art LLM, on medical competency examinations and benchmark datasets. GPT-4 is a general-purpose model that is not specialized for medical problems through training or engineered to solve clinical tasks. Our analysis covers two sets of official practice materials for the United States Medical Licensing Examination (USMLE), a three-step examination program used to assess clinical competency and grant licensure in the United States. We also evaluate performance on the MultiMedQA suite of benchmark datasets.

In recent years, artificial intelligence agents have succeeded in a range of complex game environments. For instance, AlphaZero beat world-champion programs in chess, shogi, and Go after starting out with knowing no more than the basic rules of how to play. But AlphaZero still trained separately on each game -- unable to simply learn another game or task without repeating the RL process from scratch. We created a vast game environment we call XLand, which includes many multiplayer games within consistent, human-relatable 3D worlds. This environment makes it possible to formulate new learning algorithms, which dynamically control how an agent trains and the games on which it trains.

Even though William Baumol didn't win the Nobel prize this year it got me to thinking about the cost disease, as did the death last week of William Bowen, the co-author of Performing Arts – The Economic Dilemma which brought the cost disease to public attention. The cost disease says that if two sectors have unequal levels of productivity growth then the sector with lower growth will increase in relative price. If in 1900, for example, it took 1 day of labor to produce one A good and 1 day of labor to produce one B good then the goods will trade 1:1. Now suppose that by 2000 1 unit of labor can produce 10 units of A but still only one unit of B. Now the goods trade 10:1. In other words, in 1900 the price or opportunity cost of one B was one A but in 2000 the to get one B you must give up 10 A. B goods have become much more expensive.

Popular Science: A pilot A.I. developed by a doctoral graduate from the University of Cincinnati has shown that it can not only beat other A.I.s, but also a professional fighter pilot with decades of experience. In a series of flight combat simulations, the A.I. successfully evaded retired U.S. Air Force Colonel Gene "Geno" Lee, and shot him down every time. In a statement, Lee called it "the most aggressive, responsive, dynamic and credible A.I. I've seen to date." The fact that an AI downed a professional fighter pilot? Or the fact that the AI was developed by a graduate student?

Here is how I think about these issues. The Artificial in AI can sometimes mislead so let's start by getting rid of the A and asking instead whether more NI, Natural Intelligence, will decimate the middle class. As I said in my TED talk, the brainpower of China and India in the 20th century was essentially "offline". Instead of contributing to the world technological frontier the people of China and India were just barely feeding themselves. China and India are now coming online and I see the increase in natural intelligence as one of the most hopeful facts for the future.

Here is how I think about these issues. The Artificial in AI can sometimes mislead so let's start by getting rid of the A and asking instead whether more NI, Natural Intelligence, will decimate the middle class. As I said in my TED talk, the brainpower of China and India in the 20th century was essentially "offline". Instead of contributing to the world technological frontier the people of China and India were just barely feeding themselves. China and India are now coming online and I see the increase in natural intelligence as one of the most hopeful facts for the future.