Robotaxi is also a pretty interesting bet. I'm skeptical that it'll pan out, or at least that it will ever reach the same level of safety as Waymo without lidar, but on the other hand AI has obviously significantly advanced and been significantly democratized compared to when Tesla first launched Autopilot over a decade ago. More significantly, I see the decentralized rollout strategy as a potentially viable path to beat Waymo to the punch of a full-scale deployment. Waymo might still get there first based on its headstart and massive warchest, but Tesla's strategy is essentially crowdfunding it while using the market to optimize geographic distribution. The downside is that allowing the Robotaxi to be used in arbitrary/novel locations might be significantly less safe and/or more computationally expensive than Waymo's approach, so it's a more uncertain bet than Waymo.

It just seems that self driving is staying a "hard problem" (harder than expected!)

Both would be insanely economically valuable.

I do fear humanoid robots will face intense competition from Chinese mass manufacturing, undercutting them with cost while mostly matching features.

Elon dangles all sorts of bullshit to get you thinking about nerd nonsense. “OMG, a Tesla robot will do my laundry on Mars with my robot rocket ship!”

It’s all bullshit. They’ve been dangling the imminent $25k car for a decade. The Tesla roadster that people put huge deposits on is nowhere to be found.

People are figuring it out. My previous comment would have created a 100+ comment thread in HN 5 years ago.