Post

🧵 IN THE WINTER OF 2022, a few months after ChatGPT began to scramble everything, I was part of a small group of staffers who met with Sam Altman at the White House. A few reflections on that meeting, and what it tells us about this most recent hype/panic cycle 🧵🧵🧵

Remember the feel of that moment: Across continents, people were panicking about the new technology, and the OpenAI CEO was on a charm offensive, aiming to soothe nerves from Washington to Brussels.

The meeting, held in an ornate conference room on the fourth floor of the Eisenhower Executive Office Building, lasted a little over an hour. Altman politely shook everyone’s hand before we settled around the wooden table.

One scene from the meeting was particularly memorable: Altman had promised publicly that his technology would become the most powerful in the history of the world. But for several minutes, we watched in awkward silence as his assistant fidgeted with the HDMI cable, unable to project onto

the screen the “demo” of the new model the CEO was there to show us. I found this very funny. So did many of my colleagues, who giggled quietly. Unfortunately, the humor seemed lost on the CEO, who betrayed no visible emotion during the meeting.

When the screen was finally working, Altman pulled up a lengthy YouTube video—Vladimir Putin speaking Russian at a press conference (I also thought this was an amusing choice, given the setting)—and showed us how the website could instantly translate the speech into English and summarize

its contents. It was impressive. Then he told a story about his business partner’s wife: she had spent months talking to stumped doctors about her symptoms before one query in the app diagnosed her mystery illness.

The intent of these parlor tricks was to make us feel special and to inspire confidence, or perhaps even pride—to welcome us into an exclusive club, the privileged few being afforded a special peak at a marvelous future, one that Altman and his buddies were cooking up in their offices in

San Fransisco. This would soon be our reality, and he hoped we would help ease the path.

When the demonstration wrapped and Altman had finished giving detailed answers to a half dozen or so technical questions, a senior White House official asked him how he intended to make sure this technology served all segments of society.

Altman seemed a bit perplexed by the question, but he reverted quickly to a talking point that has become familiar to AI watchers these last few years: the ultimate vision, the CEO said, was to turn this technology over to society. We asked follow-ups: What did he mean by that?

I’VE BEEN THINKING ABOUT THIS memory these last two weeks as The Discourse has fallen into yet another hype/panic cycle prompted by AI fears. Altman and his counterpart at Anthropic, Dario Amodei, abruptly called for a slowdown in the technology’s development. The U.S. president has cast

those warnings—as well as the public’s overwhelming rejection of the water-guzzling, electricity intensive supercomputing structures used to power AI—as a “SICK conspiracy.” A young AI researcher quit Anthropic citing worries that “we could all die in the immediate future."

Why is this all happening now? What do they know that we don’t? There are a few forces at play: The AI companies may be warning about danger now partly to delay their expected IPOs—not yet ready for investors to ask whether all their spending and hype will actually pay off.