Stadia wasn’t something that excites engineers in the first place. It only looks genius if you’ve not had prior experiences and assessment of issues with remote gaming.
I literally know someone who went to work on the Stadia team 2 years before it was announced because working on it was essentially his dream job. It doesn't have to be "genius" to be interesting to work on with the scale and backing of Google behind it.
And I know somebody who had been a 15 year Google employee who could choose to work any team decided to make it his new priority. This is somebody who could have worked on any platforms project they wanted. They left Google a couple years after it launched, I imagine probably at seeing their hard work go nowhere.
Not the same sort of scale. The technology part of it scaled fine, probably because Google invested the money and dev hours into making sure it did. The product itself didn't scale for a host of other reasons, but none of them had to do with the reasons it might be interesting to develop around it.
There's certain problem spaces that are just different if you have different amounts of money and dedication behind them. Working for AWS is likely much different than working for Linode, even aside from the culture of each company. Longer runways, better access to cutting-edge tech, a pre-existing global-scale infrastructure pattern... Even working on a failed project can be an interesting experience sometimes.
I wonder if even Google suffers from the premature scaling architecture astronaut problem? Perhaps if they’d spent more of their resources getting to 1000 and 10,000 games, before doing the engineering to support a billion users, they may have actually needed that scalability (and might have become at least a small cash cow alongside the surveillance capitalism asserting golden goose)?
