By 2035, Most Grand Challenges Facing Humanity Will Be Solved w/ Dr. Alex Wissner-Gross

**Dr. Alexander Wissner-Gross** (0:00)
We're about to move, I would argue, we've already at least partially moved from a world where centuries-old individual problems get artisanally solved by a huge edifice of subject matter experts to instead a world where AI bulk solves lots of problems in a given field. It's almost like an assembly line or an industrialization of the process of discovery and invention. This is going to sweep across a variety of disciplines.

**LGD Set** (0:27)
What's up, everybody? It's LGD Set here. And welcome to Milk Road AI, the twice-weekly AI show that can't wait for the day that humanoid robots compete in the Olympics yet somehow still get busted for doping. Today is April 27th, 2026, recording on April 23rd. We talk a lot about how quickly things are moving and will continue to move with AI on the show. But we also sometimes seem to be stuck in the present moment talking about stock moves and earnings and all that stuff, which is fine if you're investing.
We know that the computer build out will happen, or the compute build out will happen, and that AGI will change the world. But very few people have been brave enough to come on the show and tell us exactly how they think that will happen. I don't just mean in the next couple of months, over the next 10-15 years. That is, except for today's guest, a returning favorite on the show, Dr. Alex Wissner-Gross is back to share what those next 10 years really look like down to industry specific companies. Today's episode is brought to you by Faros, a layer one built for real fire, and Consensus Miami, where the next cycle starts. Dr. Alex, welcome back to the show.

**Dr. Alexander Wissner-Gross** (1:28)
Thank you, LG. Fun to be back.

**LGD Set** (1:31)
We have so much to talk about. I feel like last time on the show, you introduced the concept of too cheap to meter, and for anybody watching this who hasn't seen that, we'll link the last episode below because it was fantastic, and it really blew my mind on that.
Since then, you have co-authored an online book titled, Solve Everything that is a beast to read, very insightful. I guess maybe we can start from the top because I do want to take our time with it. What is the general concept that you're introducing in Solve Everything?

**Dr. Alexander Wissner-Gross** (2:00)
The central thesis, so this is a book that I co-authored with my friend, and by the way, moonshot mate, go moonshots, Peter Diamandis. The notion is that entire disciplines are about to get completely solved, bulk solved by AI. Math is largely, as largely I would argue, already been bulk solved by AI. Physics, I'm leading and helping to lead a company called physical super intelligence that's in the process of bulk solving physics and applied physics. There are going to be a number of other domains that Peter and I argue in this book, Solve Everything, solveeverything.org, that are similarly going to succumb to this bulk solution formula.
We lay out an entire, almost rails for how the world goes from where it is right now, where we have all these grand challenges to a world approximately nine or ten years from now, circa 2030, where many, if not most of the grand challenges facing human civilization have been bulk solved, not on an individual problem basis, but entire disciplines being solved all at once, mass solution by AI and how we get from here to there.

**LGD Set** (3:18)
When you say solved, what does that mean?

**Dr. Alexander Wissner-Gross** (3:21)
I have a maybe admittedly somewhat idiosyncratic definition of solve that I lay out in this book. So my operational definition of solved doesn't mean that every problem, for example, in math, every possible math problem we know the solution to, it's a much more nuanced definition.
I would sort of summarize the definition as a field has been solved in the framework of solve everything if we have a framework that enables us to scalably pour compute in to be very confident that if we pour enough compute in, we'll get a solution out for any given problem. So basically an industrialization of that field such that for any given problem, we can be confident and competent at pouring enough AI compute in to get a solution out. And I would argue and have argued that that's basically where math is right now. We're starting to see for the first time in human history, industrialized bulk solution of math. Burdish problems are starting to get solved. We're starting to see many other bulk solutions and other domains of math start to pour out. I think physics is one of the next fields to go. And I'm working actively with physical super intelligence, CSI, to make that dream happen of grand challenges in physics being mechanically and industrially solved at scale. And I think other disciplines are similarly going to succumb to this wave of automation.