[PaulHoule]

I'm going to argue VR has camp 1 problems. My persona for this is the owner of a few Thai restaurants who is brilliant at social media and SEO marketing. I could sell him a VR project easily if he believed in the ROI. Part of that is the user base but part of it is authoring and VR authoring is expensive.

If VR is going to be like the web we need some way he can get his business in the metaverse for $5000 not $500,000. Horizon Worlds falls down flat not because Meta is stupid but because the problem is difficult -- I'd like to make WebXR content based on my photography (and stereography) but once you have big textures you start to feel the 8GB limit of the device. The art gallery I want to make would require low resolution images or would require some of the programming techniques used in open world games.

In my mind VR seems to be the future of gaming, when I see many action games like Monster Hunter World or Rise of the Tomb Raider I think I'd like to experience them in VR but practically I still keep playing a lot of flat games like Dome Keeper and Dynasty Warriors 9 because there are a lot of them and they don't take the dedication that it takes to play through a game like Asgard's Wrath 2.

====

At the time I believed that better training data (business process, UX, and lots of things go into that) rather than better models was the key to products (I saw projects, including mine, that went nowhere because people did not muster the will to collect this data) so I felt we were getting a lot out of being engaged with customers.

I advocated a lot for drawing a clean line so you could reuse the same training data from different models in which case we could have had a team working on advanced models while the customer facing team gathered the data we needed to eval and refined those models over time. It would have been good if we could have gotten more VC money to hire up.