Imo AI just need to be able to code AI and write AI research. Once that happens the rest of the work to getting to human level AI should be trivial. And looking at the progress with GPT-2 it seems not to far off:
-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Tesla autopilot HW3
- Thread starter verygreen
- Start date
Imo AI just need to be able to code AI and write AI research. Once that happens the rest of the work to getting to human level AI should be trivial. And looking at the progress with GPT-2 it seems not to far off:
Good point.
Once AI is equally capable as the humans that made it, it means that the AI should be able to design another AI with the same capability. That is really close to the singularity where the AI/humans are able to design an AI slightly better than humans. Once that happens, this AI should be able to design an even better AI etc... and things become really scary.
This could be jump started by having non-general AI that are specialized at creating new patterns for slightly better/more general AI. Though not sure if that would work, atleast it's more difficult. (altough obviously it has happened before on earth through the evolution of intelligence)
Daniel in SD
(supervised)
This is just like when people say an EV wouldn't work for them (often true!). I think not owning a car would work for the vast majority of people. Of course this assumes that they can bring costs down lower than regular car ownership and actually get to level 5 no geofence (seems very unlikely anytime soon!).
ahkahn
Active Member
Well, the price is going to be come the new trade-off. As cars become increasingly expensive (I just heard this morning the average mid-size SUV has literally doubled in price over the last 15 years) people will start looking at alternative modes of transportation. I like having my car at my fingertips also, and don't feel that I would ever give that up, but as times change and you can literally dial-a-car to show up at your house within minutes and cover your entire monthly transportation expenses with a fraction of the cost, the car ownership experience then becomes a luxury one.
I think the biggest challenge that notion faces is not who will do it, but how do you keep the vandalism down and quality of experience up.
Daniel in SD
(supervised)
I would argue that the cost of cars hasn't really gone up, people are just getting much nicer and bigger cars. A Toyota Camry is only about 20%-30% more than it was 20 years ago and is probably better in every way so after you account for inflation it's actually way cheaper.
If level 5 AVs are actually invented then it will probably become illegal to drive yourself in most places. Is vandalism of rental cars a big problem? I expect it will be even less of a problem when there are cameras watching you all the time.
HW4 could be anything - the investor day mentions tell us there is a new computer in work, but we have no idea if there could be more sensors.
It is probably noteworthy that Tesla apparently didn't choose to add or change sensors with AP3 - at the moment they seem to think the collection they have is well suited to the task. They may or may not think the same thing in a year when they are making final decisions about the HW4 suite.
I guess we'll see. The current blind spots and lack of redundancy (even with radar) in side and rear areas seems crazy to me. They have redundant steering and brake controls, why not redundant vision? Maybe they can over come this, but they may also just simply reach a limit to what they can do. That limit is probably far enough off they don't have to worry about it for cars sold up to now, but it will come due soon!
Serious question - how many blind spots do human drivers have? Or do they have an unobstructed 360 degree view around the vehicle?
Where do you think they lack redundancy?
There are at least two cameras with an arc to see every area, except right up against the rear side of the car, forward of the rear camera’s arc - which is close enough to the car ultrasound can catch it.
Daniel in SD
(supervised)
More importantly, humans are outside the car before they get in and can survey the surroundings. Quite a few people have hit poles and garage openings while using Smart Summon that were clearly visible from outside the car.
I'm not convinced that sensor redundancy is needed for FSD. I think that if you can make a car drive itself then you can make it safely stop if there's a single sensor failure.
diplomat33
Average guy who loves autonomous vehicles
True but then we lose sight of the area that we were looking at before. We don't have 360 degree vision at the same time.
Counterargument is we look where we want to go and don't need to be looking everywhere all the time. Our sideways and rear vision processing is infinitely more capable.
diplomat33
Average guy who loves autonomous vehicles
There are plenty of accidents where human drivers did not see a car in their blind spot or did not see a car about to hit them from the side because they were not looking in that direction at the right time. Having 360 degree vision all the time is still better.
Happened to me today and my model 3 saved my ass. I was merging with my blinker on and didn’t see a car that cutover in my lane.
They don't have good views cross traffic, particularly on the rear end. They also can't see very close to the car (I believe the corners of the front bumper are blind...which are for a human as well).
Ultrasound is OK for large, solid objects. It won't do curbs, grass, sign posts that kinda thing.
I'm certainly not saying it's impossible. I am just curious how it will turn out and if they'll try to fill those gaps. Humans have the advantage that even if we can't see an area, we can fill it in based on context. That is the hardest thing for the AI to do...it would be much easier to just have a camera view!
You may be right. There also might be more to the system than we were ever told.
One of the interesting things I found in the public electronic parts catalog that I've never known quite what to make of is what appears to be a rear radar as part of the AP2.5 suite. Tesla has never said a word about it that I've seen - but look at this:
https://epc.tesla.com/#/systemGroups/66322
Assembly 7 appears to be a rear mounted radar unit... I haven't taken the bumper off to see if it is physically installed on my Raven, and the EPC doesn't spell out an exact location. Also don't know what to make of the front corner radar brackets mentioned here.
It might be interesting to take a battery powered radar detector around a modern Tesla and see how many radars it actually has and where they cover...
That would be sweet, but I feel like we'd know about it by now. Might be they left mounting points...just in case they need them for their promised FSD functionality or something. Or maybe for 'training cars' to help validate some functions faster or something...someone knows I'm sure!
This isn't like EVs, where the only real downside is range per dollar, and people who are legitimately affected by the downside (as opposed to just being scared of it) are rare. There are a lot of downsides to shared taxis, even if you don't factor in questions about self-driving safety, liability, etc., with cost being the biggest.
People driving more than 300 miles in a day are at least a couple of orders of magnitude less common than people traveling for more than half an hour each way. Even with a 30-minute commute, shared cars start to be problematic in terms of electricity spent, because the odds of the car having a passenger to take during its the 15-to-20-minute reverse commute are low. So you can assume that about 30% of the driving will be just moving the car from one place to another without anybody in the car. That adds to the cost significantly.
Also, self-driving taxis will require lots of extra maintenance, or else they will have all the same problems with cleanliness as taxis and city buses, only worse. People abuse things that they don't own, and if you think that your average taxi is in bad shape, can you imagine if it had no driver? Without cleaning them at least once a day (and maybe more), these things will smell like pee within a month, if not a week. Nobody wants to endure that for even a half-hour commute.
All of those extra costs add up. If taxi costs are any indication, the cost per mile of self-driving taxis, even if you take the cost of the driver out of the question, is likely to be about twice as expensive per mile as owning a personal vehicle, even after factoring in the cost of the vehicle (assuming your car lasts 200k miles). There's just no way that self-driving taxis can be nearly as cheap as driving a car that you own, in practice, unless you're one of those people who buys a new car every three years just to have a new car (and if you are, then you probably won't be caught dead riding in a car that smells like vomit because of the guy it took home from a bar the night before).
The folks who think self-driving taxis will replace personal vehicles clearly haven't done the math. The economics just don't work unless you live in one of those horrible places where you have to pay an extra $300 a month to rent a parking space.
Similar threads
- Article
