Tesla Autopilot FSD Beta drives from San Francisco to Sausalito with 0 disengagements.
FSD Beta 8.2 does well in suburban and rural areas but it still needs improvement for major urban areas.
Tesla FSD Beta 8.2 had a couple of disengagements and close calls while driving in Oakland.
SOURCES- AI Driver, Whole Mars Catalog, Dirty Tesla, Tesla Driver, AI Addict
Written By Brian Wang, Nextbigfuture.com (Brian owns shares of Tesla)
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.
I've driven countless times from SF to Sausalito and other places in Calif. using "full self-driving" without "disconnect" over the last 50 years in a variety of cars. When Teslas can claim the same, perhaps I'll have more faith in their ability.
I don't think that's a good financial move right now. The battery packs are the limiting factor in vehicle production. So a company could either produce one EV with two spare battery packs or it could produce three new EVs. The second option is much more profitable.
What's shown in that first video is as demanding, or moreso, than 90% of my driving here in suburban and rural TX, and good enough for our flattish, "dispersed urban" driving.
Trucks, pickups, big company stuff.
Musk is building sports cars.
Yeah, overnight sleeper vehicle if the journey was 6-12 hours, you save on all that security and queuing and possibly a hotel bill. Then there's the 4am get up to catch your 7am flight. You may or may not be a little crusty for your power breakfast meeting though, maybe a business opportunity for some kind of traveller bathrooms. Maybe the vehicles themselves can have something self contained.
From the videos I've seen of Tesla FSD, it is a long way from being able to drive in NYC.
no doubt the technology is better suited for interstate travel at this time and will be adapted there first.. inner city driving is another animal although I will say a huge part of the appeal for me to be in an autonomous car is actually intercity travel on interstates of hour plus driving… driving down the block isnt as big a deal to me
That would certainly work, with standard packs ready to go. Musk is going the other way, using the batteries in ways that are also structural to the car itself, thus could not be swapped. Still, H is so nice in contrast.
FSD needs to spend more time learning to drive in NYC. If it can drive in NYC, it can drive anywhere.
That is part of why I think the quick charging is a stupid idea & EV batteries should be made 1 standard size & easy to swap out. Each car would have several batteries that can be easily swapped out.
No kidding. Given the expense and hassle of flying, and the amount of time consumed going through security and baggage check, and the need for a vehicle anyway at the destination, self-driving cars that could handle expressway driving, but required a driver when you were stopping for gas or passing through a city, would be very competitive with air travel for anything under a thousand miles.
I concur: The problem with unreliable autonomous driving, is that you will NOT have enough warning of the need to take control. Regular driving, you just need to monitor driving conditions, and respond. With the unreliable autonomous driving, you need to monitor driving conditions AND what the car is doing in response to them, and be ready to correct if the car tries to do something dangerous.
Highway driving in good weather? Sure, I think we're already there for that. Anything else is questionable.
Ideally, you want the car to be actively driven by the human, but with the car also making decisions, and flagging whenever its decisions are notably different from the human's, for later analysis. And maybe alerting the driver when the car thinks the driver is doing something dangerous. When trips are routinely completed without any meaningful divergences, THEN autonomous driving is ready.
Those vids would be more credible if we were sure they weren't recorded by Tesla fans or/and Tesla investors because of bias. So only independent reports from people/experts who we know aren't betting on Tesla or against Tesla could be regarded as objective.
Humans are far from perfect either, but they fail along a mostly predictable path. We malfunction too often, due to drugs, distraction and lack of sleep, and then we are found guilty of a crime and can be punished, but who/what is liable when a car's AI fails? This is still unclear, with Tesla selling insurance, but the driver still responsible even when the car was driving itself. That may be a convenient position for Tesla and insurance companies to take, but it's untenable for true AVs.
This is a good illustration of why "miles driven" alone, is not a good indication of autonomous driving success. I only watched about half the last video and the near collisions, especially with a curb, and mistaking a parking lane for a right hand turning lane, are enough to convince me the software has a long way to go.
Instead of the "uncanny valley" there is the "dangerous valley" where one can almost, but not quite, trust the car. That's very dangerous, because you have to be constantly alert, and would actually have less time to react than if you were manually driving, by the time it's obvious the car isn't coping. I can't imagine a more tense way to drive than constantly poised with hands hovering over the steering wheel, and foot poised to stomp on the brake (or accelerator to get out of another car's way?). It would also be highly fatiguing and unnatural to hold one's hands and feet in mid-air.
We are a very long way from being able to watch a movie or read a book while the car drives wherever one has to go.
I live in NYC where the difficult urban situations in the videos above are multiplied 10X. We also have newly installed bike lanes, bus lanes (all different colors), outdoor dining due to Covid, trucks parked in bike lanes, edging into the street (illegally), newly paved streets that don't have lane markings yet, or old streets with worn out markings. Then, there's weather, night, uneven lighting etc.
Very easy during no traffic time at a route pre programed or at least pre selected…
I prefer keep getting real users videos.
As cars and trucks drive themselves, they have no time off to recharge batteries. H cars will keep going.
City driving remains difficult, but maybe they will solve it. Either way, I think fully automated *highway* driving is a very reachable, valuable, near term goal.
Based on the videos, that seems realistic. We should be talking about that.
That would already be highly valuable for many. One wouldn't be able to unlock robotaxi's just yet, but it would be a very valuable thing for personal users, longhaul and Tesloop-business models.