So it's been three days and I'm still thinking about that stream.
If you somehow missed it: last Sunday, Felix Baumgartner rode a helium balloon up to 128,100 feet and jumped out of it, wearing basically a fancy pressure suit, and fell back to Earth breaking the sound barrier on the way down. Everyone's already written the "man falls from space" post. I don't really want to write that one. What I keep coming back to is what happened on the other end of the camera, the part where something like 8 million people were watching the same YouTube stream at the same time, and it didn't fall over.
I was at my desk when it happened (yes, on a Sunday, don't ask, deadline stuff), and by the time the balloon was near altitude half the office chat was just a wall of people pasting the same YouTube link. Within twenty minutes there were four of us hunched around one laptop because somebody's stream kept buffering and mine didn't, which felt like a tiny personal victory for absolutely no reason.
That number, 8 million concurrent viewers, is apparently a record now for live stream viewership. I don't know exactly how skeptical to be about how these things get certified, but even if the real number is off by a million in either direction it's still an enormous audience for a live video that wasn't the Olympics or a Super Bowl. And it mostly held together. I saw some complaints on Twitter about pixelation and one friend who got dumped into a lower-bitrate stream right as Baumgartner stepped off the capsule, which is about the worst possible moment for your video to turn into a smear of gray blocks. But compared to basically every big live event I can remember trying to watch online, this one worked.
Red Bull spent something like five years and reportedly north of $30 million on the whole Stratos project, which is its own kind of insane when you think about it purely as a marketing budget. But the part that's actually interesting to me isn't the marketing, it's that a live video stream just quietly became infrastructure that can handle a global audience the size of a mid-sized country's population, in real time, with a guy in freefall as the payload. Two years ago this would've been a disaster. I remember watching launch streams that choked with a fraction of that traffic.
A small complaint, since I'm apparently in a complaining mood today: the commentary during the ascent was rough. Not the science part, the science was fine, it's just there's only so many ways to say "we are still going up" over ninety minutes and the broadcast found most of them twice. I ended up muting it and just watching the altitude counter climb, which honestly might be the correct way to consume that kind of event anyway — a big number ticking upward is its own kind of tension, you don't need someone narrating it at you.
There was also a genuinely scary few seconds where he went into a flat spin during the fall, spinning fast enough that people watching (including, from what I've read, people at mission control) were worried he might black out before the drogue chute deployed. He didn't, he stabilized it himself, but that's not really something you get from the highlight clips going around now — in the moment, on the live feed, nobody knew if that was going to end fine.
Anyway. I don't have a grand point about what this means for the future of live video, other than: the plumbing is there now. Whatever the next huge live moment is (an election night, a product launch, whatever), the assumption going forward is that millions of people can watch it happen at the same time without the whole thing dissolving into a spinning wheel icon. That's not nothing. Two Octobers ago I don't think anyone would've bet on that.
Also, completely unrelated, but the second presidential debate was the same week and my Twitter timeline could not decide which live event it wanted to be losing its mind about. Split-screen chaos. My brain is still recovering.