Sam Altman will meet with the Trump administration and senators this week to preview AI models that don't exist yet and field questions about cybersecurity from people who still use "password123."
The OpenAI CEO plans to demonstrate the capabilities of upcoming models. Not current models. Upcoming ones. He will show the government what his company might build if everything goes according to plan and nobody important quits.
Senators will ask about open-weight models, a term they learned fifteen minutes before the meeting started. Altman will explain why some models should be released publicly and others shouldn't, then watch as every word gets ignored in favor of whatever sounds best in a campaign ad.
Cybersecurity concerns will dominate the discussion. Elected officials who forward phishing emails to their entire staff will demand answers about AI safety. Altman will nod sympathetically while calculating how many billions in government contracts he can extract from this performance.
Retail traders have already begun positioning themselves based on rumors of rumors about what Altman might say. They're buying call options on companies that don't make AI chips, selling puts on startups that don't have products, and somehow losing money in both directions.
One trader mortgaged his house to buy shares in a company because its name contains the letters A and I. Different letters. Not even next to each other.
The Trump administration will take extensive notes during the presentation, then immediately leak conflicting versions to three different media outlets. By Thursday, nobody will remember what was actually said, but everyone will have an opinion about it.
Altman returns to San Francisco on Friday with eight new regulatory headaches and a fresh understanding that explaining technology to Congress is identical to explaining it to a dog, except the dog seems more interested.
Photo by History in HD on Unsplash
Leave a Comment