I love how some folks are like "Elon isn't a real engineer!" and then he starts going on a tangent in the middle of a presser about how flanges are a PITA to seal once you exceed a few hundred bar (just over a few thousand PSI), and how welding improves the plumbing significantly.
Like, what!? That's not how C-suite bigwigs talk. That's primo enginerd banter.
One of the secrets about Musk is that he is actually a much better speaker when he just goes of on technical stuff like that.
The problem with Musk is that when you ask him an even slightly open question, he always basically goes to the what he considers the fundamental answer. That's why so many of his answer are the same and people who know him well can already tell what most answers to most question will be.
However good interviewer can actually lead him into explaining into going into actual detail and ask smart follow up question and his answers are usually very good. The problem is, far to often he is either doing presentations like this with Q/A or he does podcast aimed at general audiences.
The best interviews with him are with people like Everyday Austronaut where he gets to follow up multible time to actually get out really good information.
I wish he would do more interviews like that, where people can prepare questions and be more direct at what they want to learn. Musk comes of much better in interviews like that and its where you actually learn more of the technical information. We understand that reusable rockets are important, but not why the engine doesn't blow up after 5 runs.
Yep, I can really feel the friction from the gears in his head when he's trying to explain something at a general level, or express broad empty platitudes about space, electric cars, AI, or whatnot. That's something I've definitely struggled with, and also why I dig it when he shuts off the filter and starts talking about leaky flanges or deleting some crufty noise-absorbing padding that was the engineering equivalent of the parable of monkeys with a ladder. Like if I wanted to hear billionaires bloviating boring bullet points about business buzzwords, I'd listen to Bezos.
I have no opinion to express on Elon but I will say that I’ve heard non-technical managers lap up stuff like that in meetings with the engineers and then regurgitate it elsewhere.
If I had to guess his technical competence is somewhere in the middle of the distribution of opinions on the matter. I’m sure he’s fine.
Sure but if you ever listen to his longer interviews, in particular those with Tim Dodd (Everyday Astronaut), it's very clear he has a deep understanding of rocket mechanics, production, logistics, launch pad development, etc. It's well beyond manager-just-regurgitating-stuff they hear. One interview he stops briefly and gets into some nitty gritty about hold-downs on the starship pad or something.
looks more like the case of the manager repeating stuff heard from the engineers discussion.
Lex Fridman asked the questions, was the one that clearly layout the challenges, and kindly enough repeated the answer that Elon Musk was not providing. Meaning its not a problem of "vector space". Humans interpret the world and give semantic meaning to the scene in front of them.
He repeated it several times as Elon Musk answers completely missed the critical challenges, until he was perceptive enough to pick and regurgitate back the examples Lex Fridman was providing. Like the one where a human would not miss the fact that although some children temporarily disappear from your field a vision as you overtake a school bus for example, they are likely to appear in front of you as they will try to cross from the front of the bus to the other side of the road.
The interview is actually concerning, as its basically Lex Fridman explaining to Elon Musk the challenges and points he is supposed to address, and throwing him a soft ball while he was running around as this is simple, easy problem of creating a "vector space".
Curiously he does pick up on hints a bit later, but still concludes and reluctantly goes on to declare the problem will be fully solved by middle 2022... At 1:26:48: https://youtu.be/DxREm3s1scA?t=5206
He also has a deep understanding of how important government assistance was to getting SpaceX where it is now but that hasn't stopped him from trying to stop others from benefiting from the same.
He shows a lack of rational thought in regards too SpaceX that is sometimes rather worrying. Take, for example, the time he smoked a blunt on camera which isn't exactly what one of his largest customers likes to see in their executives they are entrusting with launching one of a kind national security assets.
I'm glad SpaceX has an actual professional management team to run most day to day operations.
> but that hasn't stopped him from trying to stop others from benefiting from the same.
That's complete nonsense. Unlike others SpaceX has never really attack other space companies. Musk has been supportive of other space companies.
SpaceX busted open the establishment in both DoD and NASA so the contracts are fairly competent, and everybody coming in the industry benefits from that.
SpaceX just recently lost a contract about Commercial Leo Destination but unlike BlueOrigin they didn't release a pamphlet, an army of lobbyist and a bunch of lawyers.
Can you be specific of what actions SpaceX has taken to prevent others space company from gaining government contracts (other then competing for them)?
> Take, for example, the time he smoked a blunt on camera which isn't exactly what one of his largest customers likes to see in their executives they are entrusting with launching one of a kind national security assets.
Something that had basically no impact other then some not really serious slap on the wrist because the military is federal and not state law and they had to pretend to give a shit.
It wasn't the smartest thing to do but to hold that up as some great example about how he is a bad manager is ridiculous for a company that has been as successful as SpaceX.
> I'm glad SpaceX has an actual professional management team to run most day to day operations.
A lot of those manager are engineers that came up in SpaceX. Early on it was basically Elon and his secretary and a group of senior technical team lead that did everything. Musk was heavily in involved in recruitment, supply chain and other typical management things.
He is very capable of doing management work for small or large company but whenever he can avoid it, he doesn't do it, because its not where he adds most value beyond overall strategy.
In a recent WSJ he, Musk, spoke out against government subsidies. I don't mind such subsidies but I do mind when people who benefit from government help want to stop others from doing the same.
This was a case where Bezos wanted NASA to fund it's entire development program from scratch at a cost much higher than NASA ever funded SpaceX's development program. See my other comment here: https://news.ycombinator.com/item?id=30297198
You mean when Blue did everything they could to prevent SpaceX from gaining a contract. Blue complained to GAO and lost. Then Blue started a propaganda campaign propagating lies about SpaceX and tried to get is bought politicians to force NASA to change, then they sued NASA.
After all of that, with SpaceX never actually doing anything to Blue, all they did was offer a way better option to the government.
> He also has a deep understanding of how important government assistance was to getting SpaceX where it is now but that hasn't stopped him from trying to stop others from benefiting from the same.
That is false. The only companies he has objected to are ULA which was getting a 1 billion dollar a year retainer (effectively a subsidy) from the US government, which is rightfully bad, and Blue Origin which basically wanted NASA to bankroll it's entire development program with little of their own money spent. SpaceX spends a lot of it's own money on things and then is thankful for the little bit on top that NASA has provided to them for portions of development. In no case that I can remember has SpaceX investment on a project been less than 50%, and it is often 100% but with a guaranteed government customer.
Certainly plausible, though I've seen enough interviews with him talking to other engineers where he can pro-actively engage in a freeflowing conversation to believe that it's far more than mere regurgitation.
I feel like he has a much worse understanding of ML than other subjects. He's not nearly as mathematically inclined as he is engineering inclined and ML is heavy on math.
I think he's more in-depth about the rocket stuff than the Tesla stuff. He seems to exhibit more of the broad sentiments and more manager-esque talk when it comes to FSD.
I don't know how that plays into legal ramifications at all; I'm sure there will be the inevitable "well Elon said it can drive itself and make breakfast and solve P=NP!" but I have no clue if that carries any legal weight and frankly I don't care. He's a bit of a massive douche (no worse than bog-standard billionaires) but he has nerd chops and done a good job of hyping electric cars and space.
Honestly I kinda wish he'd just focus on the rocket stuff and lay off the FSD until it's a bit more mature exactly because of the media and lawyers chomping at the bit.
As someone who has been following him for a long time and as someone who likes him a lot, I feel his understanding of ML and technologies related to FSD to be the worst of his knowledge. He fails to answer interesting technical questions when they are posed to him by audiences in the past. His lack of understanding of ML and FSD make me highly doubt that Tesla will ever achieve it. I still like Tesla and they have revolutionized an industry, but they need to completely end the FSD program and move on with mass production of cool vehicles. It's a boondoggle.
Elon is wrong about FSD being easy and about his approach, and his timeline. If he can be wrong about those things with grace it’ll be fine. I don’t know if he can do that. His track record for admitting he was wrong hasn’t been great.
And any driver facing several years in jail for an accident caused while using the Beta FSD, will engage a proper legal team. Any legal team worth its salt, will immediately subpoena Tesla for documents.
They will present a legal case, that will include past YouTube videos
and Tesla company presentations. They will state that their client was
misled by the company statements, Twitter and YouTube endorsers,
and the contents of the car documentation were minor standard disclaimers.
Contrarily to what many think, most legal teams are actually pretty
smart and astute;-)
Tesla autopilot has already been driving during multiple accidents that killed people. Tesla the company isn’t at significant risk from these accidents or related lawsuits as the driver and thus their insurance is still considered legally responsible.
Further, car accidents in general have relatively low payouts even when someone is blatantly at fault. Even being held responsible for say 10 fatalities per year would be a rounding error on their balance sheet, which is consistent with other industries that have killed vastly more people via things like pollution.
It’s reasonable to suggest human lives should have more value, but when you look at an airplane crash that kills hundreds it’s mostly an issue for an airlines reputation.
Aka people did dumb thing in car, and then died. But again even if they are held either 100% responsible or say 50% the payouts on such lawsuits don’t amount to that much.
"...Thus, Ford knew that the Pinto represented a serious fire hazard when struck from the rear, even in low-speed collisions. Ford officials faced a decision. Should they go ahead with the existing design, thereby meeting the production timetable but possibly jeopardizing consumer safety? Or should they delay production of the Pinto by redesigning the gas tank to make it safer and thus concede another year of subcompact dominance to foreign companies? Ford not only pushed ahead with the original design but stuck to it for the next six years.
What explains Ford’s decision? The evidence suggests that Ford relied, at least in part, on cost-benefit reasoning, which is an analysis in monetary terms of the expected costs and benefits of doing something. There were various ways of making the Pinto’s gas tank safer. Although the estimated price of these safety improvements ranged from only $5 to $8 per vehicle, Ford evidently reasoned that the increased cost outweighed the benefits of a new tank design...
How exactly did Ford reach that conclusion? We don’t know for sure, but an internal report, “Fatalities Associated with Crash-Induced Fuel Leakage and Fires,” reveals the cost-benefit reasoning that the company used in cases like this. This report was not written with the pinto in mind; rather, it concerns fuel leakage in rollover accidents (not rear-end collisions), and its computations applied to all Ford vehicles, not just the Pinto. Nevertheless, it illustrates the type of reasoning that was probably used in the Pinto case.
In the “Fatalities” report, Ford engineers estimated the cost of technical improvements that would prevent gas tanks from leaking in rollover accidents to be $11 per vehicle. The authors go on to discuss various estimates of the number of people killed by fires from car rollovers before settling on the relatively low figure of 180 deaths per year. But given that number, how can the value of those individuals’ lives be gauged?..."
Like, what!? That's not how C-suite bigwigs talk. That's primo enginerd banter.