pull down to refresh

We all saw what happened in the 2010’s with Tesla, Musk might have another hit on his hands with Robotaxi.
FSD Strategy: Tesla’s Break from LiDAR-Based Systems
The key differentiator between Tesla and its Robotaxi competitors is its rejection of LiDAR in favor of a vision-based system driven by AI and neural networks. While most AV developers utilize LiDAR—a laser-based system capable of mapping high-resolution 3D environments—Tesla uses a combination of cameras, radar (in some models), and its proprietary Full Self-Driving (FSD) computer. Tesla’s bet is that neural networks trained on vision, not laser scanning, will ultimately lead to safer, more scalable autonomy.
Guess I missed the memo around Robotaxi was a “cost-effective AI-driven system” positioned for higher revenue growth compared to “LiDAR-based systems” like Waymo.
Table 2 presents a comparative breakdown of per-vehicle costs between Tesla's AI-based system and LiDAR-based systems. It also captures why Tesla’s approach could unlock margin and scale advantages over time.
💫 Kinda wild, but Tesla owners are literally one software update away from their AI robot car making them fiat.
Here is Robotaxi in the wild this weekend in Austin debut.
Someone will eventually run the numbers on the energy cost vs. rideshare earnings. Interesting times ahead.
This is what I truly love about Elon. Elon's guys say "you don't need LiDAR" and all the competing interests that are burning through 10x the funding at a 100x rate basically call them idiots. Then they fucking deliver like a boss. Who's the idiot now?
Happened with EVs, happened with rockets that couldn't land, now self driving cars. Next: OpenAI, plz.
reply
It makes sense, right? We humans don't use LiDAR to drive, why should it be necessary for robotaxis?
reply
236 sats \ 1 reply \ @Car OP 24 Jun
It does make me think what other sort of other applications could follow a similar path. I guess one that comes to mind is AR glasses, once the camera starts capturing similar data, it could be applied across a whole range of robots in the future.
reply
Yes! I've wanted a flashable google glasses kind of device since forever, for dev purposes. I don't care if it's safe if I wrote it, my FA to FO.
I'm working with a startup that wants to disrupt manufacturing the coming weeks (as my side-gig that hopefully for once pays when I cure, lol.) Maybe there is hope!!!
reply
Exactly!
reply
I knew a guy in Seattle who was all-in on LiDAR for self-driving. Seemed like the only way these things were going to go when I talked to him. The domain we are in often feels so certain. Isn't always.
reply