[stanlarroque]

As the founder of a startup creating an AR headset, I think it’s true but the AR market is not as mature as VR. In every tech product, first comes the hardware and then the software adds the magic at every update. In the AR world, we are still at the hardware stage. No one has the answer today for a compelling device and experience for B2C. There is no Moore law in optics and the compute power needed for SLAM, computer vision and stereo rendering is crazy if you want an untethered device (like hololens). You need more than the traditional CPU/GPU config, you need DSPs, ASICs and perhaps FPGA. Combined with the optics your NRE is high. We are far from a market that has the same ROI as VR now, even on the entreprise the adoption is slow. This game is hard but it will be worth it, and it’s the most exciting venture I’ve been a part of.

[erikpukinskis]

> There is no Moore law in optics and the compute power needed for SLAM

Are you sure? My instinct is that the number of probes required to reach the same level of accuracy would be halving regularly.

What makes you think it’s not? Are there studies?

The tiniest of creatures can do it very well with very little energy.

[bitL]

All open source state-of-art SLAM libraries fail on real-world data, often the same scene that was processed once successfully fails in another pass. The problem is with the computer vision algorithms themselves, the "classical" ones are super susceptible to per-scene constants tuning and randomization effect; the only hope IMO is Deep Learning at the moment, but that requires massive computational capability for real-time inference.

[stanlarroque]

Yes, the best SLAM we have today based on visual inertial data (Camera+IMU) uses 500mW of power. It uses DSPs and ASICs. Also, the custom silicon team from MSFT did a great job for their SLAM and display engine, it's built on ASICs in what's called the HPU for Hololens: https://youtu.be/u0eBd2m_wEs?t=1641