August is often dismissed as a lost month in the business world – nothing gets done because everyone is on vacation. But where I live, August is the apex of the year. A heady mix of artists, financiers, philanthropists, business leaders and politicians converge on our little island for the month. The ensuing deluge of lecture series, film screenings, impromptu house concerts, and remarkable dinner conversations can be overwhelming. Four weeks into the month, those of us who live here are a bit strung out, ready for the tide of intellectual and cultural offerings to recede. Fall becomes a time to repair.
As might be expected, one topic dominated the conversations I’ve had at countless dinners, soirees, and fundraisers this summer. Yes, it’s “AI,” but that was true last summer, and even the summer of 2024. This year, however, the discussions turned from wonder and anticipation to something…darker. A central question came up over and over again: Is this technology – and the leadership behind it – fundamentally out of our control?
It’s not hard to understand why that question has become so dominant: The leaders of nearly all the frontier AI labs already have their answer. Here’s Elon Musk in a staff all hands this week, as reported by The Information yesterday:
“On the call, Musk said it is inevitable that AI models will become so advanced they’ll be impossible for humans to control. For that reason, he told employees … they needed to help SpaceX’s AI division build the technology before other companies did, the people said.”
Let that sink in. If you believe that AI will ultimately be out of control, then the best thing to do is to run the company that builds it first. That’s quite a performative, self-serving contradiction!
But self-serving statements are the coin of the realm at the highest levels of power in the Valley these days. And that’s what has my dinner companions in a dark mood. Here’s OpenAI CEO Sam Altman, who is neck-and-neck with Mark Zuckerberg in the race to say whatever helps him win: “The two biggest risks in AI today …. are loss of control, where AI becomes too powerful in a way that we can’t control, and the other is power gets too centralized.” This from the CEO of the second-largest pure play AI company in the world (OpenAI was #1 until they lost their lead to Anthropic a few months back. I guess by “centralized” he means… not him?)
But it’s not just the Musks and Altmans of the world that are driving my summer colleagues into a deep-seated pessimism. It’s also the inability of our current administration to step into what has always been government’s role when it comes to public threats (and public good). You might say regulation is fundamentally a bad thing, and often you’d be right. But rampant laissez faire in the face of an AI industry that has captured one third of our entire economy? That’s just madness, and even my Republican friends in the Valley are getting uncomfortable with it.
The overwhelming political philosophy of the Silicon Valley elite is libertarianism – which conveniently preaches the gospel of free enterprise over government regulation. Here’s the Editor at Large of Reason, the bible of libertarian thought, on who should control AI:
“The last people you want safeguarding any industry are politicians. I think the major AI firms should be doing this on their own.”
OK, Jan. Actually, the last people we want safeguarding the AI industry is … the AI industry. Let’s start there, and get to work on what might come next. When the artists, VCs, and politicians return from their summer vacations next month, I have a feeling we’ll be hearing a lot more than platitudes about how we need to regulate an industry dangerously on the brink of being out of control. Buckle up – we’re going to have one hell of a Fall.
—