Elon Musk has said that Tesla will move FSD (Full Self Driving) out of beta with Version 12.
Version 12 won’t be beta
— Elon Musk (@elonmusk) June 27, 2023
They are actually much more than point releases, but the team is reserving 11.420 for the big one
— Elon Musk (@elonmusk) June 21, 2023
FSD Beta 12 “maybe later this year”
— Whole Mars Catalog (@WholeMarsBlog) June 26, 2023
Tesla AI training compute power will be drastically increasing and the miles of driving with FSD is doubling every 4 months.
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.
Q: Will the new fsd be able to drive autonomously from LA to NY as Elon advertised now many many years ago?
A: NO
Q: Ok, but that’s a lot of miles so, for sure the new fsd will be able to take control for long stretches of road reacting with blinding speed and allowing the user to relax. Then the fsd will be able to slow down in a timely manner and prompt the driver to intervine just to clarify a course of action, right?
A: No.
Q: But is it safer than a human driver, right?
A: No, as there is always a human driver that has to manage the problematic situations.
Q: Will it be dirty cheap as comma ai and other companies have products with similar capabilities right?
A: Sure.
Tesla’s self-driving suite is the best in the world…but it’s not what people want, which is good enough tech so they can sleep or watch a movie. They’re improving with time though, but I have a hard time believing they will reach level 3 in just 6 months. But then again, Mercedes has…for California only…and only highways, but still, impressive.
I do think new Tesla’s will be able to drive themselves with level 3 autonomy in 99% of scenarios, but my guess is late next year. Which again, is still amazing how close we are to being able to jump in your car, say where you want to go, and it whisks you there.
I think what people want is not to run into stationary emergency vehicles, or highway dividers or . . .
Thing is, when things are so crazy it can’t handle it, it hands it to you, but you are then out of practice, and in a mess.
Unless it can communicate with police officers and road workers, and distinguish them from the homeless, pranksters and thieves, it is incomplete as a driver, and not “full”.
If there is a road closure, and there is an officer or road worker, and they are directing drivers on how to get around the situation, this AI driver can’t hear and understand them, or acknowledge that it understands. It would need microphones outside the car and a speaker. Oh, another one, a military soldier. I took an off-ramp because I realized I was going away from my house instead of toward my house on the freeway and wanted to get going the other way. But it was to a military base, and there was no way to turn around. They photocopied my license, had me sign stuff, fingerprint me and gave me ridiculous directions for how to get out of there…eventually. I was supposed to turn at various objects that I could not see in the dark, or understand what they were even talking about. Is AI going to hear those directions, understand those directions, and get off the base?
Can it distinguish between a 12-year-old with a picture of a stop sign and a road worker? That kid could see an AI car coming down the street at 50 MPH and when 50 feet away whip out the Stop sign. I strongly suspect your Tesla will slam on the breaks, and you will get a harsh stop, maybe even get rear ended.
And what about sinkholes and sheer cliffs? Will it recognize these things in the dark? Does it know the difference between a tumbleweed and a runaway grocery cart? I might or might not want my car to stop for a tumbleweed, depending on the speed and traffic conditions. And even it I decided to hit it, I would pull over soon and insure none of it is lodged under my car. But a grocery cart? Gotta stop for that under most conditions, or I am going to be liable when I launch that into another car, and if there is stuff in that cart, it could easily go through my windshield. And what about distinguishing a rolling tumbleweed from a rolling boulder? All this is doable. But, I highly doubt Teslas are equipped with the hardware needed. They need sonar, LIDAR, microphones and speakers. Just cameras, does not cut it.
As they have since amended the legalese to express how UN full self driving the Full Self Driving feature is it’s entirely redundant whether it’s called beta or not at this point because it is just as useless for the advertised use case.
ie this is a completely hollow gesture designed to pad their stock price.
Again.
Indeed
“the miles of driving with FSD is doubling every 4 months.”
Well if the numbers are unfrauded and the progress benefits from the data then they have a chance to turn the trend around and actually deliver something like what was promised within the next years, which I doubted (more) before.
I would like to hear your thoughts on new hardware.
Also would like to see criticism, not just ball licking
Also that compute exponential curve is wrongly represented (exponential function is not followed looks to increase more than exponentially) and even then can not be inferred like that, its doubling every 4-6 months