Just compare the effort to use Vulkan, where they don't provide bindings, point out to some random GitHub repository, and wish us luck.
Also like most stuff presented at IO, SceneForm hardly lasted more than one year.
Meanwhile iOS frameworks are usable from all major platform languages, and even if sometimes buggy, way easier to start than doing your own Vulkan based engine from scratch.
In other places they would long have been fired for underperforming, at Google they get to keep messing up because core business is performing too well.
Hopefully this changes soon.
They did the hard part (the inside-out tracking)! And 3D input was obviously a game changer and the future (for VR) to anyone who had tried it out, which included basically everyone who had been to a VR conference at the time (and the tech had been done decently at a consumer level as far back as 2010 with Sony's Move controllers).
Why go all the way to production on a platform that obviously has no future? Why give up right before the basic UX factors are met?
And now they're putting the same person in charge again under 'labs' branding?
It's really hard to be optimistic about this leading to products that go anywhere. Google obviously has tons of tech and talent, but sometimes it feels like they manage to be way less than the sum of their parts due to poor management.
Facebook went from ZERO firmware expertise to where they are today. They're still asking idiotic, irrelevant leetcode questions and filtering out a lot of good hardware/firmware people, but it's entirely possible that in a few years they'll actually have a world class device team up and running.
Maybe Google was just too short-sighted. They kept trying to jimmy VR into all the world's low-quality phones, and then gave up really quickly after that proved unworkable.