Ark Invests described how compute and data are the keys to self driving (autonomy).
Tesla has both large amounts of compute and driving data. Waymo and Cruise have 100 to 1000 times less driving data than Tesla. Tesla probabably has 10x more compute and will scale to 200X more compute. Waymo and Cruise are likely data constrained on self driving.
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
Comma Ai is accumulating data as community effort and literally every car manufacturer could do the same. For what us worth a lot of insurance companies in Russia force drivers to have cameras recording, so paradoxically they too have million of hours of data. Many truck companies do the same and also many taxi companies. So yes Tesla has data, but so do others that might not be so vocal about it but might have even started collecting them before tesla.
If you don’t have some kind of full self-driving sensor suite, then your data probably isn’t that helpful for self-driving. As for Russia, they barely have compute available. They’ve been reduced to cannibalizing home appliances just to get low-end chips for military equipment.
If the development of large language models provides some insight into other AI fields, data and computing power are not as important as a good architecture. ChatGPT was revolutionary thanks to the implementation of “attention” architecture THEN big dataset and computing power made the difference allowing the development of ChatGPT 2,3 4 and so on. But when that happened people managed to use chatgpt to finetune simpler and cheaper models spending few hundred bucks. This is not surprising as humanity learns collectively new stuff exploring reality at great cost, but then everyone is able to transmit what they know quite efficiently. It is not really clear who will come on top of the AI race for self driving. My two cents are that when a promising architecture is found there will be alliances between groups of major car manufacturers to develop one or very few common standards (to share develipement costs, regulatory boundaries and risks). If a very powerful architecture is discovered (making training cheap, in the order of few hundred thousand dollars to few millions) we will see also academic models and something similar to open source llms or community efforts as per the ai arts community with stable diffusion and similar. All this to say that it might happen to fsd what happened to dall-e Tesla might come out on top with billions of dollars of investments just to stay on top for a couple of years or less.
This was based on Llama being made open source and there being several large well funded efforts that all had access to big data for training their LLMs.
This is NOT the case for machine vision and self driving. There is ONLY Tesla and its data is completely proprietary. Tesla structured the whole company many years ago around generating that data.
There won’t be anything analogous to what we’ve seen with proliferation of open source LLMs with autonomous vehicles both because the approaches and data aren’t open source and because the system testing is so highly regulated and inherently risky.
When Tesla gets to Robo Taxi level autonomy it will be completely alone and there will be no opportunities for any competitor to catch up. Because of this, Tesla will license it’s tech fairly freely – but it will unambiguously be it’s tech.
Dude, who cares, what Ark Invest is doing ?
They ware about as wrong as Cramer has been, for about 3 years now. So ???
Yeah,Teslas are not autonomous.
Yet.
Tesla “full self driving” is always 2 years away. They won’t be first to level 3 certification; they already lost that race and I don’t see any reason to think they’ll be first to level 4 or 5 either without something new and revolutionary.
Finally, someone with some sense. Tesla hasn’t even put in any applications for permits to use level 3 driving. When that happens then I’ll believe. Until then it’s 2 or 3 years away since what, 2016?