Microsoft AI CEO Mustafa Suleyman on what's ahead for AI & humanity
Welcome back to Squat Box. We are here at the Aspen Ideas Festival, where yesterday I spoke with the head of Microsoft's Artificial Intelligence unit. Mustafa Sulliman is probably the most important person now in the AI world at Microsoft, of course, in major partnership with Open AI. We should mention that Sam Altman from Open AI is going to be here a little bit later today. We'll bring you part of that conversation a little bit, but I asked him about chat bot hallucinations because we talk about it all the time and whether he believes he could take us, he could ever really take us inside the black box to actually understand how AI works. I can't in the way that would satisfy you, but I also can't explain why you chose to wear a blue shirt this morning. I can explain very little about who you or I are. The the requirement for explanation is, I think, again a little bit of a human bias. When I ask you to explain why you had scrambled eggs for breakfast this morning, you will creatively imagine an explanation in hindsight. And depending on your mood and the context that you're in and the other like associated metadata of the explanation, you're probably going to say it slightly differently. And it's not clear that it was entirely causal. There is some basis in reality for it, but this one to one mapping is a very hyper rational imposition on the way that we think as humans. In fact, we just don't really operate like that. We operate, I think, far more by association. Should we worry about that though? I mean, should we worry that we it's one thing to not understand ourselves. It's another to not understand how the computer gets. You know, most people think that all of this is, it's a big mathematical problem, right? 2 + 2 should always equal 4. If it's not equal to four, we need to understand what's on the computer, right? So that that's sort of not how sort of human reasoning and culture actually works. In my opinion, we human reasoning works as a result of behavioural observations. When you consistently do the same thing over and over again, I gain trust. You become more reliable when you tell me that you are unsure about something. And then I look back and it turns out that your uncertainty was a correct assessment of that outcome. That uncertainty gives me reassurance, right? So I learn to trust you and to interact with you by observation. And I think that that's actually how we all operate in relation to one another is how we produce culture and society. And that's how we'll treat these models. So guys, we're going to have to start thinking about these models really very much in in human dimensions, not necessarily in mathematical or sort of thinking about them as a computer. I also asked him about the issue of intellectual property, a big one here at the Aspen Ideas Festival with so many authors and other creators, and especially given the reports that Open AI and others have trained off of transcripts from YouTube videos. Here's what he said about the economics of all of that around content. The economics of information are about to radically change because we're going to reduce the cost of production of knowledge to 0 marginal cost. And this is, this is just a, a very difficult thing for people to intuit. But in 15 or 20 years time we will be producing new scientific cultural knowledge at almost 0 marginal cost. It will be widely open sourced and available to everybody. And I think that is going to be, you know, a true inflection point in the history of our species because what are we collectively as an Organism of humans other than a knowledge, an intellectual production engine? We produce knowledge. Our science makes us better. And So what we really want in the world, in my opinion, are new engines that can turbocharge discovery and invention. We're going to have to get on the couch because if you really start to think about it, Joe and Melissa, I mean, who are we if, if we are content creators in many ways and if the the marginal cost of content is going to zero. Well, there's lots of industries, of course that are going to shape shift as a result. I want to show you one other piece of sound if I could. Yesterday, the European Union charged Microsoft with breaching antitrust rules by bundling its Teams and Office products. Meanwhile, Apple now saying that it's delaying the launch of AI features in Europe given the regulatory environment. So I asked Mustafa about the concentration of power in artificial intelligence among the big tech companies. Makes me very anxious. The the reality is everywhere you look, we've seen rapid concentrations, whether it's in news media and the power of the New York Times, the Financial Times, The Economist, the great news organisations, or whether it's in cities, the concentration of power around a few big metropolitan elite cities or whether it's in technology companies. And the practical fact is that over time, power compounds, power has a tendency to attract more power because it can generate the intellectual resources and the financial resources to be successful and out compete in open markets. So while on the one hand it feels like an incredibly competitive situation between Microsoft, Meta, Google and so on, you know, it clearly is also true that we're able to make investments that are just unprecedented in corporate history.