I hope those steering wheels are for software developers, not remote support for cybercab. Actually remotely driving what's supposed to be an autonomous vehicle is dangerous. Waymo support never actually drives a Waymo. They pick from a menu if possible actions, or they leave the vehicle where it is and dispatch on-site help.
> all while consuming very little capital or engineering resources.
... If you ignore Starship. SpaceX is operating on a $20 billion bridge loan which, if the IPO struggles at all, adds to the risk. Starlink depends on becoming a telco scale mass market ISP, but it's operating in a shrinking TAM, squeezed by increasingly cheap terrestrial wireless infrastructure, and the fact that high income customers are a minority in areas rural enough for terrestrial wireless to be an uneconomical.
Tesla is spending more than the market cap of Rivian to become the Allbirds of the car industry. They're gonna blow their capex on robots and AI without refreshing the car line.
Another commenter remarked that Elon's fame is a failure of skepticism. I've been a skeptic for a few years now, starting with skepticism about agile rocket development, which sounded like bullshit to me.
Well, Starship is going to add another order of magnitude to an already overwhelming domination (SpaceX launches 80+% of all payloads by weight, even corrected for orbital energy). This isn't cheap and it did a lot of things that seemed impossible or unlikely already.
There is no reason to say that its development is slow. Falcon 9 took 8 years from concept to first flight and another 8 years to high cadence, reliable reusable flights. Starship is now only 10 years in development and already went through several iterations.
Rocket companies aren't high value. The vast majority of Falcon 9 lunches are for Starlink. They're not bringing in outside money. There are lots of financial shenanigans in the capitalization of satellites with short lifespans, and hiding the football on Falcon 9 refurbishment costs. "Dominating" the launch of your own payloads makes AI circular market participants blush.
Well, but at least it shows his engineering and management skills. Both US incumbents and Russians struggled to create new viable launch vehicles and Russians even lost capability to produce old ones (Zenit, then Proton). SpaceX seemed to have no problem doing it.
Falcon 9 refurbishment costs can't be high because whole thing takes under 2 weeks with less than a week spent in hangar for any refurbishment work... There may be no refurbishment at all on many flights, just some checks.
And yes, they ate entire worldwide commercial launch market and it wasn't enough... So there's Starlink. They appear to have no war around creating demand for themselves with their own payloads because others are not up to that task - there is too much rot and dysfunction everywhere in the space business worldwide except SpaceX, so they are kinda forced to vertically integrate.
The claim was that all Teslas with version three of their onboard computing would become fully self driving. That's now definitively untrue. But, it isn't as if upgrading to the version four hardware will guarantee that you can get fully unsupervised self driving. No Teslas no matter what hardware in them is ready for fully unsupervised self driving. It's just a new gamble. You pay your money, you take your chances.
Some systems, like lawful intercept, are designed to be hidden even from telco network management systems. The LI console that set up a wire tap might log activity at that particular console at that particular law-enforcement agency. But if you don't know where to look exactly, good luck.
This is why the Chinese picked lawful intercept as a hacking target for the salt typhoon exploit. It's almost impossible to know whether that exploit is continuing or when exactly it began.
Especially for a end user product, this is the biggest example of don't sell upgrade-ability, and don't buy it if it's being sold to you. It's not an absolute rule. But it applies to about 98% of the market.
reply