There's been lots of interesting reactions by SLers to Leslie Jamison's seminal Atlantic feature on Second Life, including observations by reader "Pulsar", reflecting on this passage from Leslie's article:
But when I spoke with Peter Gray, Linden Lab’s global communications director, and Bjorn Laurin, its vice president of product, they insisted that the problem doesn’t lie in the concept, but in the challenge of perfecting its execution. The user plateau simply testifies to interface difficulties, they told me, and to the fact that the technology hasn’t yet advanced enough to deliver fully on what the media hype suggested Second Life might become: an utterly immersive virtual world. They are hoping virtual reality can change that.
Puslar demurs there, writing in comments:
I remember when Blizzard released World of Warcraft, they said they weren't making the best graphics, but something that would work for a larger audience. That was a good idea. Of course they had the already successful Warcraft brand pushing the new game, but making the game more accessible I guess it gave a good hand. It doesn't mean it should run on a Commodore 64, but they look at the then current average machine. I remember Blue Mars, built around CryEngine 2, sporting some top notch graphics for the time. And the hair! OMG! Moving naturally, falling on your shoulders as you move, like real hair would do. That rendering engine was a technology marvel, way more advanced than SL, but... do you know that old joke and meme? "But can it run Crysis?". Of course it wasn't the only factor, but it didn't help to make Blue Mars massive success.
From my personal perspective having briefly consulted for Blue Mars, I can confirm the heavy Crysis engine (which lacked a Mac version), was a huge challenge. In any case, Pulsar compares the rise of Facebook against Second Life's stumbling -- and why it seems like Linden Lab is repeating that history with Sansar:
Now let's go back to 2003-2008. But let's take SL as it was hyped, as the next Internet, or as a the 3D social thing that SL is (among other things), rather than a game. Second Life offered another level of user-to-user interaction, compared to Facebook, any forum or Internet chat, even compared to massively online games. On the other hand, Facebook was much more accessible. All you needed (and still need) was a web browser. The simpler computer of the time, connected even with just a 56k modem, could access to Facebook comfortably without extra technical complexities (not to mention special devices such as VR headsets).
Now Linden Lab is doing the same mistake again, with Sansar, both by offering a similar model and by requiring demanding hardware and network speeds that restricts considerably the amount of people for whom Sansar would be accessible and enjoyable. So far Google Trends shows a tiny peak at Sansar's launch, which went down quickly, just to remain flat since then. And they think the key to make it massively used is to focus on more advanced tech, VR headset, a somewhat good gaming machine, and a too much demanding network speed for 2017 averages (again, up to an half hour to access to a scene, at the recommended speed of 10 Mbps. Was it good at that speed, it would be fine for today average speeds).
In fairness to Linden Lab, the company has promised a mobile version of Sansar, which should increase the platform's accessibility. Then again, that brings up another challenge: On iOS and Android, the biggest virtual experiences aren't extremely immersive in terms of graphics (or even audio), but 2D/2.5D games like the Candy Crush and Clash of Clans franchises -- while the biggest virtual world-type app for mobile is not the one with the most high-end graphics, but Minecraft.