A lot of people still seem to not understand the difference both in system cost and complexity of other autonomous platforms like Waymo compared to what Tesla is doing. Tesla taking the hard path here will start paying big dividends very soon. I've always really liked this quote to sum up Tesla's approach. "I think it's pretty obvious that the road system is geared towards passive optical. We have to solve passive optical image recognition, extremely well in order to be able to drive in any given environment and the changing environment. We must solve passive optical image recognition. We must solve it extremely well. At the point at which you have solved it extremely well, what is the point in having active optical, meaning lidar, which does not – which cannot read signs; it's just giving you – in my view, it is a crutch that will drive companies to a local maximum that they will find very difficult to get out of. If you take the hard path of a sophisticated neural net that's capable of advanced image recognition, then I think you achieve the goal maximum." @elonmusk, Q4 2017 @Tesla Earnings Call
People need to experience FSD. I’m not a long term beta tester. Started about 6 months ago with 11.4.9. Was pretty cool but almost never used it off highway. Off highway needs to intervene every mile. Have had 12.3 for a couple of weeks and use it on every drive. Night and day difference. It is absolutely incredible.
@MatchasmMatt If waymo achieves true autonomy first, the high cost of LiDAR is negligible relative to the value of the system.
@MatchasmMatt Customers don’t care. And Google has plenty of cf to subsidize Waymo.
@MatchasmMatt “Mark my words…LiDAR is a fool’s errand” - King Musk
Wow! I love this quote! So powerful and says it all! Thank you for sharing it. I’m convinced as a driver of FSD 12.3.3 that this is the answer. It’s surreal…I wish my grandfather, father in law and grandmother got to see this….they would have appreciated it so much! My grandmother would have said “this must be the end of the world” her favorite thing to say any time she was amazed by technology. I have no doubts in this technology, I pray daily for it to do good things in this world 🌎
@MatchasmMatt Pure, packed, and true information. No wonder he is little understood.
I have a slightly different take on this. Not that I disagree with the approach, but I would slightly modify it by adding step goals. I suggest taking a vision-only approach with the current sensor suite, then solving for one geopolitical location as a test bed (e.g., San Francisco). Turn FSD into an L4/5 system with Tesla assuming responsibility. This would prove several things: that in-car compute and cameras are sufficient, and that the approach and models used can achieve full autonomy. This approach will drive the FSD adoption rate exponentially. Potential OEMs that could license FSD would now have the confidence to proceed. In my opinion, OEMs would not license or start deploying a sensor suite now, not knowing its capabilities.
@MatchasmMatt Looks like Tesla has got past the hard path and is now accelerating towards autonomy!
@MatchasmMatt Initial cost is somewhat irrelevant. Power usage of the system is key FSD computer+cameras is roughly 100-200 watts. Waymo and others are probably 5-10x that at least. 15-60 watts per lidar ( and you need like 5-8) 300-900 watts Computer + GPU.