Halfway to Success by Thomas Kendall

I think what’s dead clear at this point is that half the takes about the Starship launch indicate that people don’t know what Starship is for, or how the program around it within SpaceX operates. In 2021, Everyday Astronaut did a tour of Starbase guided by Elon Musk, and interviewed him the whole time. Even if you’re not a fan of Musk, the whole interview should be required listening for people who are serious about understanding Starship. The second part especially is the juiciest in terms of understanding some key points that I will condense down here. The below quotes by Musk are pull-quotes from the video. I have in some cases used ellipses to condense down some beating around the bush, and used bold to emphasize what I feel are critical statements, but you are welcome to hear the un-retouched statements and form your own opinions.
First, Starship is designed to be a rapid prototyping program that throws things at the wall as quickly as possible. Musk sees it as explicitly a program designed around aggressive iteration to the exclusion of all else, describing it as the “polar opposite” of Dragon. He says:
“We have just a fundamentally different optimization for Starship versus say the polar extreme would be Dragon. Like Dragon there can be no failures ever, everything’s gotta be tested, to, you know, six ways to Sunday. (…) That’s like extreme conservatism. Then Falcon is a little less conservative, it is possible for us to have say a failure of the booster on landing, that’s not the end of the world. And then for Starship … is like the polar opposite of Dragon. We’re iterating rapidly to create the first ever fully reusable rocket… Orbital rocket. And fully and rapidly reusable, reusable in a way that is like an aircraft.” (5:34)
Rapid prototyping is not a new concept and Musk did not invent it. While it is unusual to see something the size of Starship run through rapid prototyping, it’s a mainstay of engineering, and practical work of all kinds. Even people who follow makers and engineering YouTubers should be familiar with the process (But if you aren’t, watch any video by Stuff Made Here, Collin Furze, or even I Did A Thing. Or for that matter, CodeBullet, for the equivalent in coding. Or NileRed, for the equivalent in chemistry. This is a common process, commonly employed by people trying to do difficult things that don’t come with an instruction book. It is also a fundamentally scientific and empirical process. Empirical translates roughly as “guess and check while paying attention”.)
Second, Musk knows that this will result in ships not coming back, and he is perfectly comfortable with that:
“But this is also a case where we’re intentionally iterating the design rapidly. And basically ships and boosters will either be amazing lawn ornaments which then have to be stored, and they look awesome, but you know, we don’t want 12 of them. It’s gonna look bizarre and where would we put them? So, since we’re making rapid iterations with each, almost with, basically every single ship and booster has had significant iterations. But you either want it to blow up or… The early ones… You want them to blow up or you’re going to have to find a place to store them. So we actually want to push the envelope. And frankly if you don’t push the envelope you cannot achieve the goal of a fully and rapidly reusable rocket. It’s not possible. You have to go close to the edge on margins.” (22:16)
This goes to the question of whether the launch was a success. While I have somewhat more muted opinions than LaughingWolf (which I will get to in a minute, and explain why I think what I think) I think this article is an excellent and informed one. One point made in the article is that Starship is testing to destruction, and they are getting good data in the process. I concur, and I think Musk’s critiques of the shuttle program are extremely poignant to the discussion of the value in testing to destruction.:
Musk: “Space shuttle had almost no room for iteration because there were people on board, so you couldn’t be blowing up shuttles.”
Interviewer: “Well and they did very, very little iteration.”
Musk:”Very little. In fact a lack of iteration was the problem. Because they, a lot of the issues they were aware of, but people were too afraid to make change.”
Interviewer: “Cause the design froze.”
Musk: “Yeah ’cause (…) There was a risk reward asymmetry. So big punishment for… If you make a change and something goes wrong, big punishment. If you make a change and it goes right, small reward. So the issues with the “o” ring and then with the insulation coming off and hitting the wing were… They had seen this before. They were known issues. But it… Because it had worked before they were like ‘well, it worked before’. Eh. Russian roulette works before. So. ‘Look, I’ve pulled the trigger so many times and… There must be no bullets in this gun’. Anyway, it’s hard to iterate though when people are on every mission. You can’t just be blowing stuff up ’cause you’re gonna kill people. Starship does not have anyone on board so we can blow things up. It’s really helpful.” (7:05)
Musk’s answer to the risk and difficulty inherent in making this rocket work is different from NASA’s. It is not to build in layers and layers of safety protocols. That was how Dragon was built, but Starship is a very different creature. Their plan is to fly the ship until they are familiar with every way it fails and have designed around it. This mindset is explained in more detail by Musk in a short conversation about launch escape and the lack thereof:
“Launch escape I think is… You basically just need to fly a lot and have a lot of redundancy. So if you lose an engine on the booster, it doesn’t matter, basically. If you lose multiple engines it shouldn’t matter. And you should be able to lose an engine on the ship and everything is okay. (…[substantial amount of technical discussion of launch escape] … ) you can’t have an escape system on the Moon or on Mars. You can’t have something pop off and then have chutes drop… there’s no atmosphere. And then Mars has a very low density atmosphere so it’ll just hit the ground supersonic. You know, it’s not going to save you. So the ship has to be safe enough for people without an escape system, because otherwise you can’t go to the moon, you can’t go to Mars. So it’s kinda pointless to do it on Earth. Just fly it a lot.” (8:51)
To summarize all of the above—Starship is a program meant to rapidly prototype and iterate rockets. They are operating the rockets with thinner margin for error and as a result they consider rockets failing, and in particular, exploding, to be a necessary part of this work. In a sense they even consider it desirable, because the purpose of the testing is in large part specifically to determine how the rockets fail, since they are deploying them in circumstances where escape simply will not be an option and figuring out how to engineer around the ways they commonly fail is the only way to make them safe enough for use.
There is an interesting and related side digression here I think is worth mentioning. The rocket that exploded doesn’t appear to be the most recent version. According to the wiki maintained by the SpaceX superfans who spend their lives staring at Starbase, the stack that blew up consisted of Ship 24 and Booster 7. For simplicity’s sake (and because I find it funny), let’s dub the now-vaporized test ship 24/7 for short (Which reverses the nomenclature actual SpaceX fans use. Then again, given its definitely-not-intentional launch date, and fate, I would accept “420 Blazin’ ” as an alternative name. Incidentally, and apropos nothing, the first ship ever stacked by SpaceX—and fueled, but not launched— was reportedly Ship 4 on Booster 20 according to the same wiki).
The presumptive successor to Ship 24, Ship 25, is built. In fact it looks like it has been sitting on the test site since January. Technically even Ship 26 has been built (though it looks so different that there is still speculation on what it’s actually meant for. Starship has many planned variants.).
Booster 7 was also a few iterations back. SpaceX is currently working on building Booster 10. For complicated reasons Booster 8 was scrapped before completion, so Booster 7’s successor is the already-constructed Booster 9, which is, even from external inspection, heavily redesigned from Booster 7. Also— semi-famously in some internet circles, there is at least one substantial and poignant difference between the Boosters—Booster 9 uses an electric servo to direct its engines, while Booster 7 was the last to be produced with hydraulics. Why is that relevant? Because there is speculation that a small explosion seen on the side of Booster 7 during thrust was the hydraulic system.
It is possible that the use of the older hardware is intentional, to try to suss out any fixable problems with the ship and booster waiting in the wings before deploying what would theoretically be the 25/9 rocket (Which I would like to vote, if it ever get stacked, be referred to as “Odds Squared”.).
In light of that possibility it’s worth noting that Musk describes SpaceX as intentionally running old hardware on missions where they might be expended:
“For Falcon Nine and even like the block 5… So called block 5, which is more like version 7 really… But we don’t even wanna use the early block 5s. So, like even those are a pain in the ass and we would prefer to retire them. So like when we have like a mission that requires an expendable booster we’ll put an early block 5. Because the early block 5s are not as good as the later block 5s, and they’re more of a pain in the ass to get ready for flight.”(28:34)
And as a digression within this digression, Everyday Astronaut—who you might have guessed follows SpaceX closely— speculated during the livestream of the rocket launch that the prototyping and variation extends down to individual engines. Which is to say, it was his opinion that even the individual engines on the booster may have differed from one another, just based on how SpaceX operates. Of course, that is pure speculation, but especially if you view launches as an opportunity to investigate a variety of failure modes rapidly, would make a strange kind of sense, and allow you to gather much more data, albeit at the cost of increasing the risk to the overall launch. It would also recontextualize the engine failures on 24/7 if the engines had subtle variations.
So now that we have this context, the obvious question is—why do I have a more muted take than LaughingWolf? Well, while I agree that SpaceX did say on the day of the flight that it was a victory if the thing launched rather than blew up, Musk had also previously addressed the question in detail during the above interview. And, yes, this was remote from yesterday’s orbital test flight, but on the other hand, this was also a less guarded moment where he seemed to be just saying what he honestly thought, before he was under media and investor pressure to manage expectations as aggressively. In response to a question on the goals for the first orbital launch, he said:
“I mean our goal with the first one is, for the first orbital launch, our goal is to make it to orbit without blowing up. That’s our goal. And frankly if we even get the… If the booster even does it’s job and something goes wrong with the ship I would count that as good progress. Like basically, actually, to be totally frank, if it takes off without blowing up the stand, stage zero, which is much harder to replace than the booster, that would be a victory. Please do not blow up on the stand. That’s the number one concern.” (16:20)
Musk enumerated three additional goals there and the launch only achieved one of them, sort of. The booster did seem to do its job, depending on how you define its job. It seems to have “only” gotten about 24.2 miles, or about 1/4 the way even to low Earth Orbit. You could probably infer as much from the fact that it was visible on camera the whole time if faintly. But that’s still a decent chunk of the way. It’s not clear why the ship went tumbling or to what extent the booster was malfunctioning to cause that. SpaceX’s stream of the event live didn’t really report any problems until the attempted separation for the boostback. The rocket was described as nominal to that point. It’s possible that boostback was supposed to start that early in flight, and that’s why the ship flipped. If it wasn’t—which seems equally plausible to me, but then I’m not a rocket scientist— perhaps the maneuver was intentionally initiated to try to get some data on the process even though they didn’t get to the desired altitude. If either of those is the case, the separation mechanism seems to have failed. It may also be that the ship was just tumbling. Starship looked visibly angled seconds after launch. The engines out on the Booster seem like obvious potential culprits.
But perhaps the most unfortunate of all these is that the launch site appears to have been seriously damaged. The launch tower looks intact, but it seems like in retrospect perhaps Stage Zero should have been more comprehensively considered the launch site in general rather than just the tower. Starship dug an enormous crater underneath itself. Retrospectively it seems obvious it would do this. The Falcon Heavy, which has six fewer engines and a lot less power overall, is seen in videos of its test flight being launched from Kennedy Space Center over trenches meant to divert flame and soaked in water. It’s not clear at the time of this writing why Starship launched off a flat, dry surface. Maybe SpaceX wanted to see if it was viable? They certainly have their answer! Some joke that they let the rocket demonstrate how big its minimum flame trench would be. This is an admittedly effective but kind of messy way of doing that. Whatever the reason, I would imagine— given the long term goal is to make the rocket rapidly re-useable, that maybe the biggest thing getting reworked after the launch is, strangely, the launch site, which at this point looks decidedly one-use.
Some people are speculating the flying concrete chunks contributed to the engine outs and even to the failure of stage separation. While I’ll hold off on saying absolutely that that’s not the case until the final report is issued, to me it doesn’t seem very plausible. I just don’t think concrete chunks would be able to overcome the headwinds generated by 17 million pounds of thrust in order to strike the engines, no matter how they ricocheted. As for the failure of stage separation, well, forgive me for stating the obvious, but that’s way up at the top of the rocket. How exactly would that work?
Lest I come off as too much of a negative Nancy, allow me to point out that a lot went right with the flight. It didn’t accomplish the things that Musk has described wanting it to in the past, and it only got through a limited number of mission milestones. But… they were big milestones. The biggest and most powerful rocket in history cleared the pad without exploding, in its first launch test as a full stack. It pushed through Max-Q relatively intact. It was also fairly stable (if canted) even with a few engines out, including being able to—presumably as a result of its onboard electrical igniter—relight an engine that went out in flight (it’s worth noting that Falcons have also had engine-outs on missions that were ultimately successful, and of course engine-out capability was specifically discussed as something the rockets should have by Musk above). And the biggest positive of all, SpaceX now has a ton of new data on what the rocket looks like when launched that it didn’t have before. And for a company that plans to fly the rocket until they know every way it’s prone to breaking, well, they now know several more ways it can break. And there’s a good chance those ways will be truly novel. Musk noted in part 1 of the interview:
“If you look at like the various reasons why we blew up Starship and you looked at the risk list, none of the reasons they blew up were on the risk list. Maybe you could argue one of them, maybe, was on somebody’s risk list, but it wasn’t brought up beforehand, put it that way.” (32:58)
I think we have good reason from the track record of the company to hope this test provided the necessary information to build bigger and better things. People forget that there was a time in the recent past that landing and refurbishing boosters was a pipe dream, and SpaceX blew up a lot of rockets learning how to do it, to the derision of the same idiots now perversely semi-celebrating the explosion of Starship and telling people this proves we should just stay on Earth. Now SpaceX is using re-useable rockets to beat all comers at launch costs. He who laughs last…
In the final estimation, my own take on the Starship launch is that, honestly, it wasn’t a success, not on the day of the launch—but it could become a success. If SpaceX learns the things they need to from the 4/20 launch to either get closer to orbit in the next launch, or at least to discover a new and exciting failure mode, then the launch will indeed have been a success. As it stands today it is a potent seed for success. I hope we get to see it bloom. I hope that Musk isn’t being too optimistic—or overcome by bravado—in wanting another launch in a few months.
But after all of the above, don’t be surprised if it blows up too. That’s what test flights are for.
I will leave you with a final quote from Musk, from part 3 of the interview, because I think understanding what drives him may be the biggest reason to bet on SpaceX over other private aerospace companies.:
“I think if we operate with extreme urgency, then we have a chance of making life multi-planetary. Still just a chance, not for sure. If we don’t act with extreme urgency, that chance is probably zero.” (13:11)








![The Other Princess by [Mary Catelli]](https://m.media-amazon.com/images/I/41J5CnXXAGL.jpg)









