I would love to hear the "underpants planning" of self driving cars iterated, critiqued and discussed.<p>Telsa's underpants plan appears to be (still?): [1] Base self driving software on affordable components that already exist in production models. [2] Get lots of high quality data from an existing fleet using this hardware [3] Climb the "levels of automation" ladder step-by-step, straight to consumer [4] Robotaxis, with Tesla owners lending their cars to Tesla's Uber (Tuber)?.<p>If/when version X of Tesla's self driving software "solves self driving" in the sense that it can use roads/infrastructure as-is and achieve superhuman safety... Tesla hit a massive jackpot. They'll already have millions of cars on the road. Producing millions of cars takes time, and a lot of capital. This amounts to few years' head start and is a big advantage.<p>Waymo's underpants plan is more of a traditional prototype/proof-of-concept/R&D lab thing. Google already poured >$20bn into Waymo without a start date for a business model. But... they have every advantage. Expensive sensors & hardware. No limits imposed by industrial engineering, manufacturing, marketing or cost concerns. Limited, hand picked routes areas of operation. Waymo is set up to smash performance milestones as quickly as possible.<p>Waymo, on paper, are better set up to achieve milestones... a working L5 vehicle. But, the road from this milestone to profitable revenue & scale is still pretty long. What if the hardware costs too much? Even if it doesn't cost too much, how long to start and scale production to Tesla levels?<p>George Hotz has his "Android to Tesla's Apple" underpants plan. No vehicle.<p>It seems to me Cruise (kvogt, the ceo has commented) is doing something similar to Waymo, strategically. There's a lot of interesting stuff to discuss. Personally, I'm interested in takes on how self driving "rolls out." Will the infrastructure (roads, signs, etc.) need to be adapted? How can this all play out?<p>BTW... kvogt. It would be great to hear you on Lex Fridman.