Tim Scarfe: prediction

Transcript context

…So that brings in the idea of architectural innovation, architectural knowledge. So it's a very interesting concept that was introduced by Rebecca Henderson from HBS, and the idea is that when you innovate, often you have what would be called gradual innovation in which you are changing a component. So for instance, 1 of the classic examples in this literature is the manufacture of aircraft. So if you had propeller aircraft that had combustion engines, changing 1 engine for a more powerful engine or a newer engine model was something that you could do relatively easily because you didn't have to redesign the entire airframe. You just brought the new engine, replace old 1 with the new 1 and you were done. When jet engines were invented, you needed to redesign the entire airframe, you know, to be able to produce an aircraft. So the companies that were operating with combustion engines, they went bust, most of them, and there was a new wave of companies like Boeing, you know, that were newcomers at that time, that were specialized on jet engines because they were designing the entire airframe around the new engine. Now, in the case of Blockbuster and Amazon, those are very good examples of architectural innovation because you might think that, well, Barnes and Noble is able to ship millions of book to all of these stores, know. It has thousands, if not maybe tens of thousands of employees that are experts on the business of books and dealing with clients. And the idea of shipping the book directly to a consumer might look like a small incremental innovation. But in reality it was an architectural innovation, and when I do talks and I present this with slides, what I do is I show the picture of a Barnes and Noble, and then next to that I show a picture of an Amazon fulfillment center, which looks like kind of like this part of the airport that is you know sorting all of the different luggage, and that shows that no, that little idea of just shipping directly to consumer require a completely different organizational design, and the distance between the Barnes and Noble organization in this network that we were describing before in that model with the to the Amazon model was enormous in reality just because of that change. Exactly. And and this is the reason why in my opinion LLMs are not intelligent because they don't have this coarse grained dynamic adaptation of their architecture. But we're getting ahead of ourselves a little bit. So at the beginning of the book, you spoke about this concept of a person bite, which is roughly how much can 1 person know. And we're a collective intelligence. We work together. And you you spoke about this kind of power law learning curve, which is basically at what point does our learning asymptote? And and and we'll get to that as well. But there was 1 fascinating example you you gave you're talking about this this shipbuilding company. And over the course of, I think it was at the second world war or the first world war, they became much more efficient at building ships. And was that because of experience or was it because of process? The first law of knowledge, the law of time, is divided into several sub principles. And the first 1 is about the growth of knowledge in individuals and teams. That's a story that starts with Leon Thurston. He was the first 1 to, in my opinion, map like a really good learning curve in 1916. Funnily enough, he started as an engineer, and then you know, he actually produced a camera that got him an interview with Thomas Alba Edison. He decided not to work with Edison and go to teach at the University of Minnesota instead. He becomes frustrated that he's really good at math and engineering and it's hard to teach it to students, so he becomes interested in learning. He goes to Chicago, he enrolls in the PhD in education and after a year he switches to the program in psychology, and there he gets access to a dataset that was being collected at the Duff College of Business in Pittsburgh in which you had records of how well people learn how to type. So imagine you have a mechanography class, know, people are learning how to type. These are 18, 19 year olds that are typing on a typewriter for the first time, and you see every week how many words they're able to type per minute. Every 4 minutes actually. You know? And then you see how many pages they've written throughout the semester. And when you put those 2 things together, you get a very neat you know learning curve that follows this sort of power like imagine like a square root type of shape, you know, in which learning is really fast at the beginning and then it peters out. Then in 1936, you know, that's about 20 years later, Theodore Wright, is an aircraft engineer in The United States, he was actually important enough to be in charge of aircraft manufacturing for all of The United States at the end of second world war, publishes a paper in which he looks at the cost of producing an aircraft, you know. He's very smart, he looks at the cost of the last aircraft produced in a batch because aircrafts are producing batches, and he finds also that the number of man hours as a function of the number of aircrafts in the batch decreases as a power law. Okay. So it's the same result that Thurstone got. In 1 case, you can look at capacity, in another case, you can look at cost. And then in 1965, Leonard Rappin, an economist, grabs data from the liberty ships. The United States was producing during the second world war an insane amount of liberty ships in multiple shipyards. So he could use the fact that shipyards started at different times to have like a more causal story, you know, economists love kind of like having that extra little hint. And he was able to show that this learning that was observed, the fact that the man hours needed to complete a ship were decreasing over time, was not a consequence of changes in technology or increase in capital expenditure or increase in labor that basically more people was working on the ships, but it was a function of experience.…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.