Yes, some of the photography is amazing. Part of the success is from some optics math on depth of field -- better with small cameras, and the solid state senso

Yes, some of the photography is amazing. Part of the success is from some optics math on depth of field -- better with small cameras, and the solid state sensors permit that. E.g., the old idea was just a pinhole camera that didn't have to focus at all. Then the electronics and software processing the signal from the sensor can make the new, really small cameras look really smart. And you may have still more advanced work in mind. For AR and VR, I see some markets but am reluctant to expect wide usage. The last really big idea was, what, Facebook? I know: There have been a lot of $1+ billion exits since the start of Facebook, but I haven't kept up on the small fry!!! I started my career near DC in applied math and computing and there saw a lot of fantastic national security projects from research. E.g., I was in the group that did the Navy's version of GPS, before the USAF did GPS. There was some really nice math, physics, and research, powerful stuff, in that group. E.g., how to design a satellite that will orbit the earth not very high up and have essentially no "drag" at all? Cute. Important. In an important sense, GPS, etc., are information technology (IT), but from all I can see the role of applied math, physics, and research so heavily used by US national security is missing from Sand Hill Road funded IT projects. E.g., if I were looking for some promising background at Stanford, I'd go to D. Luenberger, B. Efron, or P. Diaconis and not the computer science department. At Berkeley I would have gone to one of my favorite authors, L. Breiman, but of course eventually the computer science people did that to some significant extent. For Breiman's boosting and bagging, I suspect some important connections with resampling and some math in a paper I published. I suspect that there is a nice, more powerful, valuable way to explain such ideas, maybe with connections with sufficient statistics. E.g., order statistics are always sufficient. But such things are not relevant to my startup and, thus, on the back burner. Bluntly the US DoD, NSF, NIH, etc. eagerly support and exploit research with a batting average much higher than VCs, while as best as I can see Sand Hill Road, for IT projects, hearing research, soils their clothes and runs to the rest room. It appears that Sand Hill Road believes that the "technology" in IT is mostly just software with no connections with anything new, correct, and significant from research. Yes, they are pursuing AI and ML, but my view is that those are not very promising directions. The VCs need stuff that, among other things, is NEW, and we know where the "new, significant, and correct" stuff comes from -- research. So far Sand Hill Road for IT (but not for bio-medical technology) has with great determination refused to have anything to do with research, even if already done, rock solid, in production quality code. So the VCs just want to see traction, significant and growing quickly "up and to the right". But the research is a crucial part of estimating the future of the company, past current traction. IMHO, Sand Hill Road needs to get serious about "new, correct, significant", powerful and valuable in IT.