234
Tesla Full Self Driving Is Now 'End-To-End AI'
(odysee.com)
This is a most excellent place for technology news and articles.
I feel like the NTSB need to draft a min spec for self driving cars and a testing course that involves some of the worst circtimstances to get approved. I feel like all self driving cars should have to have lidar, and other sensors. Computer vision really isn't working out.
You build a benchmark and tesla will train on that benchmark, says nothing about real world use but gets them signed off.
But yes western society is currently in a hellscape of refusing to do even basic regulation of any new technology so it'll probably be a good 20 years of murder robots on the streets before anything gets written down.
By “western society” do you mean the US? Because the EU doesn’t seem to have any qualms about regulating new technologies. That seems to be a uniquely American thing.
Which somehow means that Europeans suddenly have headlights that makes sense while we’re over here dying from aftermarket HIDs that should be treated like the VA Highway Patrol treats radar detectors ( rip ‘em out and smash them with a sledgehammer on the side of the road)
To be fair we already have giant metal murder boxes zooming around on the streets. If AI kills even a single person everyone flips out even though over 40,000 people die every year in the US from car accidents. And that is just the deaths, not including injuries. Yet I don't really see anyone calling for more regulations on driving tests for humans.
People want AI to somehow be perfect when in reality as long as AI is even 1% better than humans that's saving over 400 lives per year. AI doesn't get sleepy, distracted, drunk, etc. so it probably already is at least 1% better in most situations. Humans are horrible drivers.
Speak for yourself, chief
But yes western society is currently in a hellscape of refusing to do even basic regulation
US regulations are only written in blood or money. the united states was built on the backs of slaves, and then wage-slaves. literal graveyards filled with workers.
im not disagreeing with you, i just found this comically disparate to history... ie, its always been a regulation hellscape.
I don't think mandating lidar specifically by name is right, seeing as computer vision is definitely a software problem. Instead they should mandate some method to detect objects in any light condition + a performance standard, which in practice during certification could mean lidar. Regulations should be as minimal and specific as possible.
Good point. Mandate the ends rather than the means. If they get better functionality with some new tech in a few years, we don't want outdated regulations holding the industry back.
Should a self-driving car face more rigorous tests than actual human drivers? Honest question.
Yep.
Yes, because when there’s an accident with a person driving, you usually know exactly who is legally to blame in an accident. With self-driving, if the car accidentally hits and kills someone, who do you charge for it? There’s no one person you can point to for responsibility for if something goes wrong, like you can for a person responsible for an accident.
Human drivers should be facing more rigorous testing regardless. It’s horrifically easy to get a license… and then they never test you again for the rest of your life. That’s just insane when you think about it. My test was in 2002. Feels like I should have to retake it at some point.
And take them away for bad driving. But we don’t because our entire transportation infrastructure, outside of a few cities namely NY, is built around everyone driving a car.
Yes. A human brain can handle edge cases it’s never encountered before. Can a self driving car?
Ever stop at a red light only to have a police officer wave you through?
Ever encounter a car driving the wrong way down a one way street?
Ever come across a flooded out stretch of road? (if the road has no lines and the water is still it can be very deceptive looking)
These are a tiny number of things I’ve encountered over the past few years. I’m sure plenty of other drivers can provide other good examples. I’d want to know how a self driving car would handle itself in situations like these.
How will the bot car handle itself out in the country? Dirt roads? Deer? Roadblock checkpoints full of bored, mean spirited cops.
Those are pretty basic conditions that I hope are already in the training data.
What about a wildfire evacuation? Police might have people driving on the wrong side of the highway to make use of all the lanes. Smoke might be obscuring everything. A human driver would know not to pay attention to any of the road signs in that situation without ever having been trained on it, but would a self-driving car?
Or, how about any situation where a police officer has to have a driver roll down the window to give them instructions for dealing with some unusual situation, like a chemical spill or a landslide.
Or, what about highway signs that have been shot by a shotgun so that it's hard to read? Or, what about novelty highway signs that a business might put up as a joke?
Self-driving cars definitely need to be tested against a much bigger range of situations than a human driver. Much as we might be baffled by their lack of common sense, the common sense of an average 16-year-old is still off the charts compared to an AI. Having said that, I know how bad many drivers are, and I wouldn't be surprised if the competent self-driving car organizations (Cruize, Waymo, etc.) are already better than an average driver under 99.9% of common scenarios.
Yes because each person must learn on their own and have limited experience relative to the general public as a whole.
Self driving cars can 'learn' from all self driving cars and don't get tired, forget, or anything like that. While they shouldn't be held to perfection, they should absolutely be held to a higher standard than a human.
Only Tesla self driving cars need to have more rigorous tests. Other brands are fine as it is because they have lidar.
LiDAR isn’t some sort of magic eye. The self driving system is only as good as the software that takes the inputs from cameras, LiDAR, etc., processes them, and ensures safe operation of the car.
Finally someone who actually uses critical thinking instead of being an anti-Elon bandwagoner.
I feel like all them do, have you seen wayze nearly getting black people killed cause it didn't stop for s cop. And it can't recognize construction zones.
Five LiDAR sensors hasn't stopped Cruise from running into a bus, multiple cars, and a fire truck. Maybe self-driving is a myth?
Maybe we should just build buses and trains and pay people good salaries to operate them??
Throw I some pot holes and child pedestrian crossing the street, etc and they'd even come out with a powerful marketing ad.
It's not really a sensor issue, as much as having software that can interpret the sensor data and act on it. Cameras and lidar effectively provide same thing, distance to objects in 2d/3d. But u need software to process that data and identify where the road is, where little jonny is, and what to do...arguably, the distance measuring problem has been solved for a while with lidar or with cameras, it's object identification and reaction to that info that's not solved. You can't really solve it with traditional if/else programming, while AI gives you only a probability of what something is or what action to do...so the problem is hard.
But ntsb/dmv whatever needs to come up with a way to test and classify autonomous driving software...probably doing real world test and identifying edge cases where it fails.
bUT thAt WOuLD StiFle proDucT iNnoVatIoN!!!
AND have triplicate back up system that runs in parallel.
Pretty much what the UNECE did.. there are standards for these things. Tesla doesn't meet them, which is why FSD 'beta' is still 'seeking regulatory approval' in the rest of the world.
Everyone would build to pass the test track. This does get at the problem though: the permutations of scenarios an L5 system has to correctly process is a huge number. Trying to build a system that can do that appears to be beyond anyone’s av system right now. This is why the most advanced deployments are all geofenced. That way at least the traffic signs and signals, lane markings, etc all understood and tested. Even then ‘shit happens’. Untested scenarios still occur. Also the maps are always out of date.
The problem really requires AGI, and nobody has one of those, or if they do it’s a secret.