I set an alarm for 6:45am yesterday just to watch a rocket, which is not a thing I normally do, but Starship's fourth test flight was worth losing sleep over. SpaceX launched it from Starbase around 7:50am Central, and for the first time the whole stack more or less did what it was supposed to do. No fireball. No "rapid unscheduled disassembly," which is still the funniest euphemism in the entire aerospace industry and I will die on that hill.
Quick recap for anyone who wasn't hunched over a laptop with a mug of coffee going cold: the Super Heavy booster came back down and did a controlled landing burn over the Gulf of Mexico, basically simulating a catch even though there was no tower there to catch it this time. Then the ship itself, the upper stage, made it through reentry despite losing a chunk of a flap and, from the footage, cooking one of its camera housings so badly the feed turned into this weird orange plasma smear for a few minutes. It kept flying anyway and splashed down softly in the Indian Ocean something like 64 minutes after launch. Three previous flights and this is the first one where both halves of the vehicle survived their whole job. That's the actual headline, not "SpaceX launches rocket," which is a sentence so generic it could apply to basically any Tuesday since 2010.
What got me wasn't the launch itself, it was that camera feed during reentry. You could see a literal piece of the flap glowing and flapping around (sorry) while plasma built up around the edges, and the ship just kept going. There's something almost unsettling about watching hardware take visible damage in real time and not die. I've watched a lot of these livestreams over the years and usually you're bracing for the "well, that's it folks" moment where the feed cuts and the commentators go quiet. This time it just didn't happen. My kid, who has approximately zero interest in rockets, wandered in halfway through, saw the orange screen, and asked if the video was broken. Fair question honestly.
I'll admit I've been pretty skeptical of the whole "rapid iterative testing, blow things up on purpose, learn fast" approach, mostly because from the outside it looks like a company setting its own money on fire for content. But watching four flights in a row, each one visibly further along than the last, I get the argument now even if I still think the messaging around it is a little too pleased with itself. There's a difference between "we're testing to failure on purpose" and "we're testing to failure and also please clap," and SpaceX's PR has always leaned toward the second one for my taste.
None of this means Starship is anywhere near flying people, or landing anything on the moon on the timeline everyone keeps repeating. A soft splashdown of an unmanned test article is a real milestone but it is not "orbital taxi service," and I say that as someone who wants this stuff to work. The gap between "didn't explode" and "reliable enough to trust with a crew" is enormous, and it's the part nobody puts in the highlight reel.
Also, unrelated complaint since I'm here: whoever decided that important livestreams for a US audience should default to Central time needs to be sat down and told that not everyone lives in Texas. I had to do math at 6am to figure out when T-minus zero actually was, which is not a great use of a brain that hadn't had coffee yet.
Anyway. If you missed it live, the replay is on SpaceX's stream and it's worth the twenty minutes even if you skip straight to the reentry sequence around the 45 minute mark, that's the part that actually matters. The launch itself is cool but you've seen a rocket go up before. You probably haven't seen one come apart a little and keep flying.
I'll be setting the alarm again for flight five, whenever that ends up being. Probably still too early. Worth it, probably.