When OpenAI published details of the stunningly capable AI language model GPT-4, which powers ChatGPT, in March, its researchers filled 100 pages. They also left out a few important details--like anything substantial about how it was actually built or how it works. That was no accidental oversight, of course. OpenAI and other big companies are keen to keep the workings of their most prized algorithms shrouded in mystery, in part out of fear the technology might be misused but also from worries about giving competitors a leg up. A study released by researchers at Stanford University this week shows just how deep--and potentially dangerous--the secrecy is around GPT-4 and other cutting-edge AI systems.

The hottest ticket in tech is an invitation to a banal South Florida business park, indistinguishable on the outside from countless other office buildings that dot America's suburban landscape. Humanoid robots walk down the halls, and green reptilian monsters hang out in the lounge. Cartoon fairies turn the lights on and off. Even the office equipment does the impossible. The high-definition television hanging on the wall seems perfectly normal. Incredibly, it is now levitating in midair. Get as close as you'd like, check it out from different angles.

The hottest ticket in tech is an invitation to a banal South Florida business park, indistinguishable on the outside from countless other office buildings that dot America's suburban landscape. Humanoid robots walk down the halls, and green reptilian monsters hang out in the lounge. Cartoon fairies turn the lights on and off. Even the office equipment does the impossible. The high-definition television hanging on the wall seems perfectly normal. Incredibly, it is now levitating in midair. Get as close as you'd like, check it out from different angles.