I realize I owe many of you an update on our views about #lidar adoption and « #Tesla. Here are our preliminary conclusions in the thread below. 0/8 #ADAS

1) today, solving passive optical (camera only) is on the critical path towards L3 ADAS. Only optical can make full sense of what is happening - this is the way roads work. This is why Tesla is doing it, with a genius approach: using clients as training supervisors for now.

2) today, Lidar is not an option for L3. Way too expensive.

3) based on 1)+2), there is no doubt Lidar approaches to L3 today don’t make sense. QED. Tesla is right and Elon is not crazy. Lidar is used on AV platforms to bypass L3 and attempt L4 directly, faster but on a narrower range.

4) Whatever the eventual perception performance of passive optical, Lidar will always add enormous value, doubling the 9’s in safety performance.

5) if vision only is 99.99% safe, vision + Lidar will be 99.999999% safe. Rice more nines. This will always have value to public opinion, regulators, and insurance companies.

6) This means Lidars will be adopted systematically for L3 when they are affordable, whatever the performance of vision-only. The HW cost of vision only being $1-2,000, I consider sub $500 as the affordability barrier.

7) Conclusion: a small number of high-end premium cars will adopt Lidar in 2022-2023, and this will drive flash adoption by all in 2025-2026, including Tesla. Tesla will remain the L3 leader with the best vision system, which remains the most important sensing factor.

8) Next step for us: which Lidar technology gets us there? Comments and suggestions welcome, we love crowdsourced research!