There are a number of potential bottlenecks that could slow down Tesla’s progress on autonomy (as I wrote about here). It might be a long while yet until we see Tesla’s data advantage trickle down to Autopilot and to the “Full Self-Driving Capability” software package.
But I predict that by the end of Q1 2022 (i.e. by March 31, 2022) Tesla’s partial autonomy features will operate in more environments (e.g. city streets), cover more driving tasks (e.g. turning at intersections), and handle those tasks more competently than any competing commercial Level 2 system that’s available in production cars from GM, Volkswagen, Toyota, or any other car manufacturer. The difference will be clear enough that reviewers like Consumer Reports and Motor Trend will agree Tesla’s system is the best.
I also predict that by the end of Q1 2022 the popular narrative in the media, among stock analysts, and among consultants that Waymo is the furthest ahead on full autonomy and that Tesla is one of the furthest behind will be called into doubt, if not disproven. By that time, Tesla’s demos (if they’re still just demos) will be at least on par with Waymo’s demos, if not superior. (Note: it’s possible that the prospect of full autonomy in general will be called into doubt, in which case the perception of Waymo leading and Tesla lagging might still disappear.)
What would change these predictions is if another company (such as GM, Mobileye, or Waymo) started emulating Tesla’s large-scale production fleet learning approach or if Tesla for some reason stopped pursuing its current approach (such as bankruptcy, although I consider that unlikely). As long as Tesla remains the only company pursuing this approach and as long as it keeps expanding its fleet by 50,000+ cars per quarter, these predictions hold.
I’d like to post these predictions on a site like Long Bets, but the minimum bet is $200 plus a $50 fee. Anyone know a good alternative?