K^2

Bait used to be believable. Anyways, FAA asking for a mishap investigation indicates that a mishap resulting in increased risk to public has occurred, in their opinion, and they want to know why. https://www.faa.gov/newsroom/statements/general-statements Notably, from the FAA statement: Emphasis for anyone with trouble parsing the statement whole. So the claim that, "The hazard zone was indicated in advance, and that's where the debris fell," does not hold water. The debris fell outside of the hazard area, creating danger to airline traffic.

Bezos is better at keeping low profile, but I agree that it's no reason to let him off the hook. People should scrutinize BO and its gov't contracts more. Both the past under current admin, and what they'll get under the incoming. Relevant read: Whataboutism I'm pointing out specific patterns in company statements and decisions in the context of technology used and stated and expected timelines for the development as well as past behavior of companies under the leadership of the same individuals - the political stance of said individuals was at no point brought up, and you trying to make it sound like it's about politics rather than company policy is exactly what you're accusing the rest of us of doing right now. Company policy is relevant to the discussion of technology and its safety, and so setting that aside because you want to make it sound like decisions made by company leadership are automatically politics, because it hurts your own narrative, is not helpful to anyone involved in the discussion.

Because there is no reason to expect that CEO of Sierra is going to get any special considerations from the admin. He is not politically involved and doesn't own media companies. I can't say the same about SpaceX or Blue Origin. It's not a reason to automatically dismiss either of these two companies, but it is a reason for significantly more scrutiny. Pretending not to understand that is very bad faith. Outside of that, the scrutiny should be on grounds of policy of the company itself, and that's all we've been talking about, so no need to pretend that it's anything else.

Generally, I'm on board for letting people put themselves at a risk they fully understand on a voluntary basis. When Mad Mike Hughes was building his steam rockets I had zero objections - he clearly knew what he was doing and how it can end. I start liking it less when large corporations are paying a crew to take on a risk, because now we're getting into significant power disbalances. It's worse if the crew is under military service or another obligation. Even so, this is a limited concern. If SpaceX says, "Hey, we have conducted enough tests, and can honestly conclude that there is a 10% chance it will blow up, but we want to try anyway," my ethical concerns stop at making sure the crew understand the gamble and are prepared to play Russian Roulette for whatever prize money is on the table. The thing I'm afraid of is SpaceX instead declaring that the Starship is safe when their tests do not fully support that conclusion. Consent to risk cannot be built on a lie. And yes, this hasn't happened yet. My fear is based entirely on how Tesla has done this exact thing with their "autopilot," lack of 3rd party testing on Cybertruck, and so on, and still standing on PR that their vehicles are safe. SpaceX has started as a very different company, but they have lost a lot of people who got SpaceX to where it is and built hat safety and transparency into the Falcon program. Current leadership seems to be going for a Boeing and Tesla approach instead. And yes, this is all hypothetical, and if none of it happens, great. I was worried over nothing. I'll take that gladly. But if in a couple of years we see SpaceX leadership stand there and declare that "Starship is safe," after it managed ~5 complete test flights, do think back to this discussion.

I agree with the broader sentiment, and the situation is very different between the tech and knowledge we have now and back when N-1 was developed. However, there are problems we can't address with simulation, because they are due to an interaction with other systems of nearly infinite complexity. Flight 7 is a good example. Simulating one engine and proving it to be perfectly stable wouldn't have prepared you for how it failed in concert with the rocket. And when you add more engines, you increase the number of potential points of failure that you did not foresee in a simulation. So instead of the problem with N-1, where you multiply failure rates of individual engines, you are multiplying factors that have nothing to do with an individual engine. The outcome is still an exponential growth. One you have better control over, but not one you can eliminate entirely. Naturally, building a small number of giant engines also has issues, especially in context of reusability. And the pros vs cons of Saturn V vs N-1 approach are on a completely different set of scales now because of history, technology, manufacturing processes, the economies involved and so on, and so on. But neither is it fair to say that the inherent problem of N-1, that, "More engines = more things that can go wrong," has disappeared. It's not something you get to just ignore and pretend it doesn't matter now, because the simulation tells you that the engine is 100% reliable, and all tests show that when running in conditions identical to the test, that really is so. Because the engine will never run in conditions identical to simulation in a real mission. Nothing we invented lets you take that risk factor to zero, which is the only case where you could claim that comparisons to N-1 are bogus. That said, without taking this risk, I'm not sure we'd be getting a super heavy at all. So they either get it to work and fly safely, or they don't, and it's not so much about the road not taken. Just so long as nobody pushes for a crewed launch before the system has been properly tested. The biggest risk of Starship due to above mentioned problems with the booster and the shifts in company culture is that when Starship is "done" and has undergone a few successful flights, we won't have nearly the certainty that the rest of them will fly safely until we launch enough for a statistically significant sample. And if the corporate types at SpaceX try to convince the regulators that the Falcon and Soyuz safety records are an indicator that Starship is ready, I only hope there are people in whatever admin's in charge at the time that will call them out on this BS. We need to see dozens of flight in a row without failure to declare the Starship safe enough to even start running crewed tests.

I think that's the crux. It's a rocket test. Everyone involved in it should at least have been prepared for it to go spectacularly wrong. And investigation is primarily for the benefit of the team building the thing, so they might as well be the ones to do it. If safety concerns are raised, these should be investigated separately, but even then, access to actual debris and establishing cause of the accident might be entirely unnecessary. Again, outside of SpaceX and these curious, nobody really cares why it blew up, and it wasn't a surprise to anyone that it could. So safety precautions would have to be built on that assumption, and if they were insufficient, the investigation would be into why proper safety procedures weren't followed or if there is a reason why these were not enough. These kinds of investigation happen in aviation all the time, and that's how it gets to be a safe industry. There is a question of costs due to disruption to normal traffic and environmental impact and how the company conducting the test should be recompensing all these impacted. But again, in this particular case, it's probably on airline companies to sue SpaceX if they believe their time tables were sufficiently delayed. Given that the impact on traffic was that of a moderate thunderstorm, I doubt anyone will want to waste the effort. If such disruptions happen often and without proper warning, I can see it being escalated, though. I'd also complain more and louder if this was a hydrazine rocket. MethaLOx gets a 'meh' from me. It's not that dumping rocket debris of any kind into the ocean is great for the environment, but as far as I'm aware, it's all pretty benign. A shipping vessel crossing the ocean will dump more hazardous materials (through leaks, corrosion, and accidents) during its voyage than what ended up in the ocean from Flight 7, if I had to hazard a guess. (I'm happy to stand corrected.) I have general concerns about the program as mentioned earlier. And everything I was able to find on procedurals and hazard zones before the explosion makes me feel uncomfortable, as the traffic exclusion zones got activated only after the breakup. I haven't flown in nearly a decade, though, and even if there were deficiencies, it's on FAA to figure them out, and amend how they work with SpaceX going forward. Either way, it's not really a complaint about Flight 7.

Engine. Unity is just not a good fit in 2025. Lacking budget for something dedicated, Unreal is a much better foundation. Rigid physics for crafts. Look, there's flex to real structures, but it's hard to simulate in a way that works. You're better off simulating the stress across a rigid structure. You'll still have failures of overstressed parts, but no flappy rockets. Assembly building and part manager are a mess. This is less concrete - it just needs work. From setting up unit tests on saving/loading the craft so that the system isn't broken half the time, to the UI/UX side of things.

My concern is that, based purely on preliminary, the part that failed is not related to systems that have been changed for this flight. The failure happened due to insufficient safety margin on systems that have already passed the tests. And as a consequence, some of the new changes still need to be first-tested on the next flight on top of things that need to be improved for the parts that failed. Technical debt has gone up, not down, as an outcome of this flight. There is a part of the design process where you move fast and break things. But you have to ramp it down gradually. You're not going to arrive at a working model by changing everything every time. You have to arrive at smaller and smaller changes as time goes on. Sometimes, that means flying with something imperfect knowing it will work well enough and get the mission to results. This is something that all good engineers understand, and something we've seen with Falcon's design process. They've flown older engine configs when they were itching to test new ones, because there were other flight systems that had to be checked first. I'm not seeing that with Starship. Now, maybe, it's still very far away from completion. Maybe nobody expects it to even remotely reliably until flight 100-something, and I'm expecting too much consistency from flight 7, and SpaceX is happy to burn $100M per flight to keep doing these tests where they overhaul multiple major systems for each flight. But that's not congruent with the promises that SpaceX is making on the schedule. And it sounds like the team's under pressure to test more things in every flight precisely because management is trying to cut cost by flying fewer tests. Which is, predictably, backfiring. And even that isn't a huge deal if all that happens is that SpaceX burns more money. But we've seen how self-driving on Tesla was handled. The progress was slow, and instead of that resulting in more tests, the company pushed a raw product around the safety checks that should have prevented it. Hopefully, I'm wrong here, but it sounds like SpaceX is making all the same mistakes that Boeing has been, from how they cut corners, to how they handle tests and safety, except that SpaceX now has an ability to bully its project through by using the gov't connections leverage, regardless of the safety in ways that even Boeing couldn't. If that is the case, this largest rocket might also turn into the largest disaster with loss of human life. Maybe I'm being too pessimistic here. I am confident that the Starship project is not in as good of a shape as the PR tries to show it, but some delays and overruns aren't the worst thing in the world - so long as appropriate steps are taken when it starts to become unsafe. Given how close the debris track came with threatening a few airliners this time, hopefully, we'll see FAA take steps to increase the safety in the future by requiring that the airspace is properly cleared. That will put additional costs and constraints on SpaceX. If the SpaceX plays along - good. They're taking responsibility and eating the costs. If not, and they bully FAA into allowing the future launches to continue putting passenger flights at risk instead, then we should all start being way, way more concerned about how SpaceX is handling the Starship project. FAA's investigation shouldn't be into why the rocket failed. Rockets do that sometimes. It's into why airliners managed to make it so close to the debris track. The warning to traffic was only given for the area immediately around the launch site, presumably, because the rocket was above relevant flight levels. That's fine if you have a well-established rocket with known failure characteristics. It's irresponsible if you're flying an experimental rocket that may fail in novel ways, like what just happened. Clearly, the simplified procedures were some sort of an agreement with FAA that SpaceX takes responsibility for keeping the launches safe to air traffic. That didn't happen. FAA must clamp down on that. It is part of the agency's direct responsibility. And yes, if that doesn't happen, we should be worried about influence of corporate interests over regulatory agency, because that's putting all of us at risk. Especially in light of the Boeing's recent failures on that front.

That was very pretty. And I don't want to make a big deal about any particular failure - this purely a superstitious side of me speaking - wow, we're off to some bad omens this year.

"Contractor working at <place>" is the standard phrasing. We have a lot of "Contractor v. FTE" (Full Time Employee) discussions around here in the Valley. If your resume lists FAANG (MAANA????) it makes a huge difference if it's a "worked at" vs "worked for". Personally, I think it's ridiculous, because these people are doing literally the same job most of the time, but I have a lot of empirical evidence for it making a huge impact on your ability to find jobs in the future. Being an FTE in one of these automatically passes you on a pre-screen and sometimes the first round of interviews.

I'll let you figure out what the problem with referring to that article is based on this screenshot from it.

Thanks, I almost drowned. It'd be embarrassing. "How'd Kat die?" - "Oh, she was drinking water while reading Joe's post."

Ok, so first of all, you're 2 for 2 on YouTube channels I would strongly recommend avoiding. Astrum has a lot of the similar problems, where it takes sensationalist approach to an observation and runs with a hypothesis that's, to put it lightly, not seen as likely, and without any deeper understanding of cosmology involved. Secondly, you need to understand that any acceleration that we observe has to be explainable by matter within observable universe. Because we'd be observing it. Another way to look at it is noting that the path from gravity's source to us is at most as long as the sum of paths from source to observed mass flow and from there to us. (Worst case scenario, that is the shortest path.) Meaning, if the gravity's source is outside of our horizon, it can't produce acceleration detectable by us. However, it is very important to keep in mind that the earliest universe is opaque. What we call the observable universe is everything down to the age when it cleared, around 380ky after the big bang. Anything older than that is effectively obscured by the cosmic background. Because the universe is expanding at an accelerated rate, some galaxies that started out in that shell have moved beyond it since, making them invisible to us now. Meaning, there are masses we don't account for, and any random fluctuation anisotropy in them would be exacerbated by the higher density of the early universe. So if you naively compare the observable universe to movement of distant galaxies, there is an anomaly, that's trivially explained by the clusters and superstructure that goes just beyond, and are still falling within the normal distribution of densities consistent with the part of the universe we can observe. In simple, simple terms - if these galaxies were accelerating, it'd be weird. Movement is easily explained. In fact, if you simply started with the Wikipedia entry for Dark Flow, you would have discovered that a) evidence for it is inconclusive and somewhat controversial, and b) even if we were to assume it exists, the amount of flow consistent with measurements could be explained by density fluctuations in the early, opaque universe well within fluctuations allowed by the cosmic background radiation that we do observe. One paper puts the detection significance at 0.7 sigma. By the way, note the date. It's from 2009. As well as the original claim that the Astrum video talks about. Yes, there hasn't been anything new recently discovered. This is something that astronomers looked into more than a decade and a half ago, and there's been little change since, yet it's presented across several sensationalist pseudo-scientific channels as sensational new discovery. In other words, please, stop watching junk. If you want a recommendation on a good astronomy channel that goes over interesting discoveries and is done by somebody who is actually an active researcher, you can't go wrong with Dr. Becky. I would much rather you be legitimately confused by Hubble Tension than fall for that pseudo-scientific garbage.

His videos are full of misinformation. Anton's background in astronomy is rather rudimentary, and in cosmology non-existent. He often brings up speculation about models that aren't just unconfirmed, but retracted, and does so without understanding any of the math that went into them in the first place. This particular video is a good example.

I don't know if you've seen any footage from the ongoing, but the cities don't get occupied. They get leveled. Yes, once you firmly establish territorial control, you need to bring in supplies, establish new infrastructure, possibly rebuild and repopulate in time. But at the point of actual contact? It's a meat grinder, and there is zero reason to send in humans when unmanned means of destruction are clearly doing a better job at everything from scouting, to laying mines, to clearing trenches and buildings. I'm not saying equipment like APCs is becoming obsolete in this kind of warfare. You still need evacuation transport for when the battle lines shift quickly or general transport along threatened routes. But the current use as an infantry delivery vehicle? That's just a waste of resources now - both human and materiel.