Lively. One of the many Google products they killed that nobody mourns or remembers.
Wave, Lively and Second Life have some useful lessons in them that it may be worth recalling. Second Life is/was of course the most ambitious attempt to build a metaverse, by far. It never broke out of its niche in the way Minecraft did partly due to insurmountable frame rate and resource usage problems that Linden Lab assumed at the start they'd inevitably overcome, but which they never really did. It's also the reason SL couldn't have been used in VR. Lively was to some extent an attempt to solve this problem with Second Life. However it simply created a different, worse set of problems. Lively also suffered from being seen as the pet product of one of the (at the time) extremely rare female product managers at Google, so she was basically given a team and budget to do what she wanted because hey! Isn't that great! Go women! But the project didn't actually have any executive support and the cutesy design was seen as wildly out of step with Google's brand at the time. So when it failed to set the world on fire immediately it was quickly shelved and the PM moved on.
SL's biggest problem is that it's very difficult to render user-generated content performantly, in ways that look good. Minecraft solves this problem by sacrificing the 'looks good' aspect, partly because it had no pretence of being a metaverse, but SL explicitly wanted to do the Snow Crash thing and thus allowed you to place more or less arbitrary scripted 3D meshes inside the world. Unfortunately the structure of the world, and how users wanted to use it, were the opposite of how you do performant 3D graphics:
1. SL is set outdoors. Thus draw distances are huge and many, many objects can be captured by the camera simultaneously. This places huge load on the CPU and GPU. At the time, the standard was for 3D games to be set indoors, largely to limit draw distance.
2. Many constructions in SL are buildings that contain translucent windows. This is much more intensive to render (requires overdraw).
3. Many constructions in SL are very odd shapes which make it difficult to rapidly determine if they intersect things. Part of why "land" in SL was so expensive was the need to run physics simulations and collision detection against objects that were not designed to make it cheap.
4. Because every object was fully dynamic and the user could change the world at any time, all optimizations based on batch processing of static data e.g. lightmap baking, were unavailable to SL, trashing their performance still further.
5. Because land was all adjacent in one uniform world, at the edges renderer performance was effectively a tragedy of the commons. Even in the rare cases that an SL content creator learned about the limits of the SL 3D engine and worked within them, their hard work could be undone by someone in land next to them constructing a giant tower filled with translucent windows.
Lively attempted to solve these problems by limiting art to a team of professional 3D artists. However this gave the world an entirely predictable, sterile corporate feel that appealed to nobody. Same problem as why Lego Worlds failed to compete with Minecraft. Additionally the Lively team had to spend a lot of engineering effort dealing with Google's infrastructure decisions, which at the time were optimized for apps that could use eventual consistency. See my comment from a few days ago on this topic [1].
The relevance to Wave is mostly that Wave was sort of the 2D content version of Second Life. It went all-in on very hard computer science problems, and did so in a web browser, on the assumption that they'd just figure out how to make it performant later. But they never did. Moreover the flexibility of the tool meant it was often confusing to figure out and quickly developed a perception that it was half baked, buggy and required a mastery of the tool to use. For something explicitly about collaboration rather than content creation, that was fatal. Finally it also suffered the same problem as Lively in that the project was basically a gift to the co-founders of Maps for their success and the Australia office as a whole. The team played corporate politics very badly, adopting a culture of internal secrecy within Google that was not only alien and controversial but which also meant they pissed off other teams (whom they were competing with), and failed to build support amongst the senior executive level. When they failed to take off quickly enough they lost executive support and the project was quickly killed.
