Hillary Clinton and Reid Hoffman Debate AI, Power, and America’s Place in a Global Tech Race
At the 92nd Street Y in NYC, Hillary Clinton and Reid Hoffman sat down for a conversation that was supposed to be about AI’s potential — but quickly turned into a discussion about power, risk, and whether the U.S. is keeping up in a race that might already be stacked against it. Hoffman sees AI as an engine for human progress, something that should be built fast and figured out along the way. Clinton isn’t so sure.
“On one hand, I really appreciate the optimism,” she said. “I like the idea that we should be a learning organism. We should learn as we do, learn as we go, make adjustments, you know, although I do worry about all the people who, you know, don’t see the curve and drive off into the, you know, or the cliff, uh, how many of those can we actually put up with?”
Hoffman sees the race differently. “This is actually, in fact, an intense model competition,” he said. “I sometimes get kind of skepticism — ‘Oh, you just say the Chinese have ambition because you don’t want us to regulate you’ — like, no, no. This isn’t a test competition. This demonstrates that.”
The conversation turned global, with China looming large. Clinton didn’t mince words: “You know, the thing about competing against China is they don’t have a free market.” She pointed to the well-worn pattern of American companies investing in China, only to find themselves forced into uncomfortable partnerships. “It wasn’t long before those businesses who had invested in China or on their products and their services to China, suddenly got the knock on the door saying, ‘Hey, we want you to meet your new partner. We want you to meet, uh, you know, a company that you’re gonna share your IP with.’ That set of behaviors has not disappeared.”
Hoffman brought up DeepSeek, a Chinese AI model making waves. “I’ve been doing a little bit of research on DeepSeek because I — I’m not enough to be completely authoritative, but I do think that the pieces in the press story don’t fully add [up].” He wasn’t convinced by the claims surrounding it, but saw it as a reminder of the stakes. “One of the things that we should all take away from this is… look, this is actually, in fact, an intense model competition.”
Clinton connected the dots. “Most likely, this is among, uh, very clear state goals that China has set for itself, because AI not only has the benefit of, you know, having an access power that tells you what’s in your refrigerator, but also waging war, waging economic competition, controlling, uh, populations.”
Hoffman pushed back on arguments that AI scale doesn’t matter, dismissing the idea that smaller models could match those with massive resources. “It’s like, no. The whole point of this stuff is, like, even if you can train this model, then, of this level, like, a thousand, you know, uh, GPUs, if it’s, like, 50% better than we can with 10,000 GPUs — that scale thing, when you hit that better factor, that matters.”
Clinton’s frustration with America’s short attention span was clear. “It is really frustrating to me, and I know it is to you,” she said to Hoffman. “We — we talked about these things… We’ve got to have more, uh, staying power, longer attention spans. We’ve got to get over ourselves in order to be effective in this competition.”
The conversation shifted to AI’s impact on jobs, where Clinton was blunt. “AI is going to take away jobs. AI is going to replace, um, certain kinds of human work, um, just as previous technologies [did].” She compared it to the rise of the automobile — an invention that displaced workers but also helped create the American middle class. “I do wonder whether that is the likely scenario here, because in our society of such extreme inequality and wealth… people who are already well-educated, more technologically inclined, are going to really [do] good… not just as developers, innovators, but as users of this new technology.”
Hoffman, while acknowledging disruption, framed job loss as part of a transition. “The key is making sure people learn how to use AI, rather than fear it.” He pushed back on the idea that AI will simply wipe out work, emphasizing that “we should be nudging in the right directions.” What he didn’t want, he said, was resignation. “What we don’t want is to just say, ‘Well, it’s gonna be brutal, shut up.’ We’ve got to learn from past industrial revolutions.”
One of the final audience questions brought everything into sharp focus: war and peace. What’s AI going to do to global stability? Hoffman didn’t hesitate. “With AI, cyber weapons, and realism, it’s inherently a little bit more of an offensive level… That has the tendency to say, ‘Use it, we lose your advantage,’ and so that [is] destabilizing.”
Clinton, the former Secretary of State, didn’t need any convincing on that front. “AI is going to be studied for a long time because a lot of weapons of war are becoming more autonomous,” she said. “And so we’re going to see all kinds of very dangerous weapons in the hands of all kinds of people, uh, that, uh, may or may not have the values to be trusted with that kind of destruction.”
The conversation ended on an open-ended challenge: what should be done? Clinton’s warning was clear — AI is already a battleground. “We are in a competition now that wasn’t real in the past. It wasn’t real,” she repeated. “Yeah, we lost manufacturing jobs. But now? Now, it’s real.”
