Altimeter Capital's Brad Gerstner will be at tonight's state dinner. He is on the north lawn of the White House for us this hour. It's nice to have you on. Welcome.
It's great to be here Scott on what is a really historic day here at the White House.
How are you thinking about what could transpire between these two countries related to AI knowing what the president has already put out on Truth Social? I want to leave it exactly where it is what he has said about what he thinks should come out regarding any sort of AI regulation. Do you think we need a greater platform for AI regulations and for safety?
Well, first let's start with this guy. We're here to talk about super intelligence. And we have this historic meeting between two of the most important economies in the world. It's the second of what looks like it's going to be four meetings this year. So that in and of itself is an extraordinary level of dialogue including on issues like super intelligence. I'm sure that conversation will come up today. But let's set the table a little bit Scott on what we are doing. This idea when the president truths this morning that he wants to continue the course. Remember the course that we are on is both safe and fast. The idea that we don't have any regulation is a little bit crazy. The Center for AI Standards and Innovation which is in the Commerce Department, it reviews and looks at every single frontier model before it's released. Remember and I think we're the only government on the planet to pull a model off the shelf when Mythos was suspended earlier this year in order to make it more cyber safe to be released. That was directed by the president and the White House. And so there is a cop on the beat. We are taking safety and these precautions seriously, but we're not doing it in a way that's slowing us down. And so I think when he says stay the course, the course is make sure that we're reviewing those models. Number two, the labs are reviewing the models. We've been red-teaming these models and increasing steps in compute that we're using to make sure that this is safe. And then we heard over the course of last week ideas such as independent monitors or out of Elon the idea of peer review. So we will continue to evolve those frameworks as the technology evolves. And then finally Scott this idea that we're going to shut down AI, have a pause, the moratorium on AI that is just unilaterally disarming both our economic security and our national security to others around the world who certainly are not going to slow down. So that's not on the agenda. So that's what... Go ahead, Scott.
It doesn't sound like you think that we need or that you're in favor of some globally accepted standard. You mentioned Elon. He's gone a step further in the last 24 hours where he had a conversation with China's state-run global television network on AI regulation and said the following. I'll quote and get your reaction to that. I think a really important thing that would be good for the safety of the whole world would be for China and the US to agree on a platform for AI regulation that ensures safety of AI for all of humanity. That seems to be going a step further than either what the president is in favor of and if we're hearing you correctly, what you would agree to.
Well, part of the reason that all the AI executives are here today. Part of the reason that I'm here is because these important conversations are happening between the United States and China. So, I'm not going to front-run the White House in terms of the things that are being discussed. But, you wouldn't be having these people convene at the White House if there weren't discussions about the frameworks that will guide us all as we come forward. What the president has made clear, he's not ceding any authority to some supernational entity like the UN for AI. He's going to maintain US sovereignty over AI but of course we're going to coordinate and cooperate where we can with countries like China that also have a vested interest in super intelligence. I think it's a great step forward. I think that the path that we're on right now is threading the needle on what is safe but what is also very fast. And so we're leading the world in AI today. I think that's the most important thing that we continue to keep innovating and that we evolve these frameworks as the technology evolves.
I understand. I mean, respectfully, you're not front-running anything if the president has put forth his point of view as saying, 'I want to leave it exactly where it is.' And then when Elon comes out in the last 24 hours talking about agreeing on what he terms a platform for AI regulation that ensures safety, are you in favor of that or not?
Well, let me ask you this, Scott. We have a model release framework in the United States. So no frontier model can be released in the United States unless it's reviewed by the Department of Commerce, like I mentioned, the Center for AI Standards and Innovation. So is that a regulatory framework? I think it is, and we're already doing it. So it wouldn't be surprising to me if we have similar frameworks that we can compare notes on and set those best standards with China on a go forward basis. So I don't know that there's a lot of daylight between Elon and the White House that you're suggesting. But I do know this. Everybody around the table has goodwill. There's nobody who wants to YOLO. There's nobody who thinks we should disregard safety. We know we got to bring this into the world in a way that is maximally beneficial to all citizens of the United States, and it is and will be. So everybody takes those issues seriously, but Washington has a way to do Washington. And what that means is they want to lead with regulation. They want to smother this opportunity before it even gets off the ground. We're not going to allow that to occur. The reason we haven't had incidents to date, and remember, we've been at this now for four years, is because we have a pragmatic and a sensible solution to regulation, and we need to continue pushing that forward.
I wasn't suggesting, I mean, I was merely reading the statements word for word of what was said by President Trump and Elon Musk in the last 24 hours. Let me ask you this. When you read the comments from both Sam Altman and Dario Amodei yesterday in front of the Security Council at the UN, what was your reaction to Sam Altman saying in our history there have been times where countries who compete and don't always like each other very much still come together for shared interest and the collective good in the face of a powerful new technology. We believe this must be one of those times.
Well, there's no better evidence of the president's leadership on this and the White House's leadership on it than the fact that President Xi is at the White House today having conversations on this very topic. And so, it's unprecedented. We've never had the US President and the President of China meet four times in a single year. This is the second of those historic meetings. Just behind me, those meetings are occurring right now. So, I think that he would agree that we need to coordinate. We need to have conversations, but that doesn't mean that we need to have moratoriums or we need to smother this. I don't think the Chinese want that. That's certainly not what they're doing. Remember, there's only one country on the planet that has actually suspended a model that has actually said a model's not safe enough to be released. That's not China. That's not Europe. It's the United States of America earlier this year with Mythos. So again I come back to this that I don't think there's a lot of daylight between the people around the table but everybody's going to express their opinion including Elon and that's how we move the conversation forward that the technology evolves quickly. We need to make sure that we're having sufficient conversations to evolve quickly with it.
We always hear this race between us and China in this great competition for the next frontier if you will. Do you think we have a good handle on where that race currently stands?
Well, it's very clear that the United States models are leading the world. I think it's a poorly kept secret that the Chinese models are distilling the US frontier models. And so in that way, they're derivative to those models. We have the best chips that are training these models and are running the inference on these models around the world. But we need to make sure that we continue to export the American AI stack to the world. There are some people in this country that would say we shouldn't be able to export any of that. And I'll tell you that that's one sure way to make sure that the Huawei stack, the Chinese stack runs the world of AI. The reason our values are embedded in the internet around the world is because we didn't regulate the internet to death when it started back in the late 90s and early 2000s. We allowed our companies to go compete. We need to do the same thing now. I think these are the conversations that are occurring inside that we all want to be safe, but we all want to move quickly to make sure that all these benefits are delivered to our economies and to our citizens.
So, you're in favor of greater chip exports between the US and China.
Well, I'm on record about this from the beginning. And I don't think we should be exporting our leading chips to China, but I would have loved if the United States had not unilaterally ceded that market to Huawei. There's no evidence that we actually slowed them down in that regard. And so now I think we have limited exports of H200s and H20s that are going to China. I support that. I would love to keep all those Chinese developers. Remember the vast majority of AI developers in the world are Chinese. We would love to keep those developers in the American ecosystem, in the Nvidia ecosystem, in the AMD ecosystem. Unfortunately, by pulling back as aggressively as we did, which really started under the Biden administration and then continued into the Trump administration. By pulling back that aggressively, we really ceded a lot of ground to the Chinese and around the world, not just in China, in terms of the global AI ecosystem.
