00:04:15 · Ty Morse, Eric Jorgenson
[00:04:15] S0: point where basically Martin knew the cost of a Roadster, and Elon went to, like, the engineers that were on the line or, you know, were working on it and basically asked for a line item, you know, breakdown of exactly what the costs were. And I think Martin Eberhard said something on the order of the Roadster cost, like, 120,000 to make, and the actual number was closer to $1.50. Yeah. And Elon knew the number, and Elon knew that Martin knew the number. [00:04:40] S1: Yeah. [00:04:40] S0: And he thought that basically there was like a permanent break in trust.
00:04:40 · Ty Morse, Eric Jorgenson
[00:04:40] S0: And he thought that basically there was like a permanent break in trust. [00:04:44] S1: Yeah. And and I think that was after a period of realizing that they didn't actually know the number. They they were like, what is this thing that we are selling cost? And, like, Martin and I didn't know. So, alright. They both went and tried to figure it out. And, yeah, it it was... That was a very ugly period. That was at the end of a whole bunch of kind of disagreements and challenges that they'd been through, But it did push Elon into that position of, like, I have to be CEO. I don't wanna be. It comes with a lot of chores. It comes with a lot of responsibilities. You're at the nexus of everything. But if I want this company to be successful, you know, I have to fully internalize and take responsibility for it, and that's what he did even though he was also in the same position at at SpaceX. And the fact that both of these companies are improbable, iconic, and that he funded and led them both in parallel is like... I I wrote in the introduction of the book. It's like running a two minute mile twice at the same time. It is just so improbable and so, it is such a higher degree of difficulty than most people can conceive of, and the fact that it worked is still unbelievable. Biology, his blurb on the book is, like, the better you are, the better you realize how good Elon is. And that is, like... Which is why Mark Andreessen and Brian Armstrong and you and me and so many people are like, how does he do it? What is he doing differently?
00:06:15 · Ty Morse, Eric Jorgenson
[00:06:15] S0: Do you think that he underestimated or miss estimated the likelihood of success with Tesla and SpaceX? I know he's been quoted as saying that he thought that each of them had a roughly 10% chance of success, whatever success is. [00:06:28] S1: Yeah. [00:06:29] S0: But then because both of them work, the probability of that one, like, in a 100 outcome actually occurring is, you know, asymptotically small. [00:06:36] S1: Yeah. I I think we we, you know, we're in a good, a very lucky universe in the multiverse, maybe, of, like, where both of them happen to work. Like, the timeline is is being good to us. [00:06:47] S0: Has there been any other examples where he overestimated likelihood of failure?
00:06:52 · Eric Jorgenson
[00:06:52] S1: He was probably being kind to himself, I think, in retrospect there of, like, being realistic that these companies have low odds of success. I mean, nobody had had started its car... A successful car startup in a hundred years in America. And space launch company is like an unprecedented category. So, like, 10% probably overestimating the odds of success, especially if you stack them on top of each other. And even, you know, the greatest venture capitalists in the world, like, new Elon saw him be successful with Zip two and PayPal, and then he came around and pitched this, and they were like, yeah. Like, Michael Moritz, the Sequoia, like, made a bunch of money with him in PayPal, and Elon comes around pitching Tesla, he's like, sorry, bro. Like, not gonna happen. And, you know, luckily, Elon was in a position to fund these companies until they were, still not quite undeniable, but that he was able to get them to the point where it was clear that additional resources would create, you know, subsequent breakthroughs and make them successful. I mean, your question, he's he's off on estimates all the time. Like, overestimating success and underestimating success and compressing timelines. Like, I think one of the most interesting things that I I kind of see you probably do the same thing is, like, creating short timelines that he says have a 50% probability of being made. And that's such an interesting reframe from, like, we wanna prevent any chance of failure and choose timelines that we know are successful to, I want a short timeline that I think has a 50 chance of being successful. And, you know, we're gonna miss a lot of the time, but we're gonna achieve a lot more than we would if we gave ourselves a timeline that was very comfortable to beat.
00:08:28 · Ty Morse
[00:08:28] S0: It's also interesting he actually frames it that way because he's just by default saying that most of the time... Or like a lot of the timelines, we're just not gonna achieve in the first place.
00:08:38 · Eric Jorgenson
[00:08:38] S1: But I also know it's just this forcing function to make everyone move faster. The purposeful 50%, like, timeline thing is actually a really good example of a meta that goes across his whole operating philosophy, which is encouraging or allowing small failures in order to make the whole, like, more robust. He... He's really focused on increasing the learning rate of his companies and moving ever closer towards this more perfect product. And the the sort of quote that encapsulates this is failure is essentially irrelevant unless it's catastrophic. And having a high iteration rate and giving young engineers, like, huge responsibility and pushing program timelines and pushing iteration speed, all of these things sort of create these small failures that make the end product better because those failures are instructive. It shows you, you know, where something breaks. You call it building the box in the sort of rocket program. Right? So you wanna push each variable to failure so that you know where it fails, so that you know where the outer limit of performance of the material or the system is, and then you just wanna be just like a scooch inside of it to cover for some variability. But if you're in the very comfortable center of the box, you have no idea what your actual true range of performance is. And in the case of something like Starship, you're giving up huge amounts of of economics if you're over engineering for redundancy and safety to a degree that you don't even understand because you've never pushed the system to the limit.