The hard truth is there is no reason to limit machines to only the tools humans are biologically born with. Cars always have crutches that humans don't possess. For example, wheels.
Turns out, when there's demand for LIDAR in this form factor, people invest in R&D to drive costs down and set up manufacturing facilities to achieve economies of scale. Wow, who could have predicted this‽
Western countries might not be smart enough to keep R&D because Wall Street sees it as a cost center.
[1] https://www.tomshardware.com/tech-industry/artificial-intell...
You know what else used to be expensive? Structured light sensors. They cost $$$$ in 2009. Then Microsoft started manufacturing the Kinect for a mass market, and in 2010 price went down to $150.
You know what's happened to LIDAR in the past decade? You guessed it, costs have come massively down because car manufacturers started buying more, and costs will continue to come down as they reach mass market adoption.
The prohibitive cost for LIDAR coming down was always just a matter of time. A "visionary" like Musk should have been able to see that. Instead he thought he could outsmart everyone by using a technology that was not suited for the job, but he made the wrong bet.
This should be expected when someone who is *not* an experienced engineer starts making engineering decisions.
FOSS is the obvious counterexample to your absurdly firm stance, but so are many artistic pursuits that use engineering techniques and principles, etc.
My intent was to exclude research efforts, which is fundamentally different from engineering, which is a practical concern and not a “get it to just work” concern.
In a true self-driving utopia, all of the cars are using multiple methods to observe the road and drive (vision, lidar, GPS, etc) AND they are all communicating with each other silently, constantly, about their intentions and status.
Why limit cars to what humans can do?
The reason this is clear is because, except for a brief period in late 2022, Teslas have included some combination of radar and ultrasonic sensors. [0]
[0] https://en.m.wikipedia.org/wiki/Tesla_Autopilot_hardware
A system that's only based on cameras is only as good as its ability to recognize all road hazards, with no fall back if that fails. With LIDAR, the vehicle might not know what's the solid object in front of the vehicle using cameras, but it knows that it's there and should avoid running into it.
This is a good example of why sensor fusion is good.
Crutch for what? AI does not have human intelligence yet and let’s stop pretending it does. There is no shame in that as the word crutch implies.
If a sensor provides additional data, why not use it? Sure, humans can drive withot lidars, but why limit the AI to using human-like sensors?
Why even call it a crutch? IMO It's an advantage over human sensors.
Quality of additional data matters. How often does a particular sensor give you false positives and false negatives? What do you do when sensor A contradicts sensor B?
“3.6 roentgen, not great, not terrible.”
Probably comes down to lidar (and Ai) failure modes.
Also don’t forget that as a human you can move your head any which way, and also draw on your past experiences driving in that area. “There is always an old man crossing the road at this intersection. There is a school nearby so there might be kids here at 3pm.” That stuff is not as accessible to a LIDAR.
That's because our stereoscopic vision has infinitely more dynamic range, focusing speed and processing power w.r.t. a computer vision system. Periphery vision is very good at detecting movement, and central view can process tremendous amount of visual data without even trying.
Even a state of the art professional action camera system can't rival our eyes in any of these categories. LIDARs and RADARs are useful and shall be present in any car.
This is the top reason I'm not considering a Tesla. Brain dead insistence on cameras with small sensors only.
You’re not considering them even though they have the best adas on the market lmao suit yourself
(and that's not even addressing that human vision is fundamentally a weird sensory mess full of strange evolutionary baggage that doesn't even make sense except for genetic legacy)
The assumption was that with similar sensors (or practically worse - digital cameras score worse than eyeballs in many concrete metrics), ‘AI’ could be dramatically better than humans.
At least with Tesla’s experience (and with some fudging based on things like actual fatal accident data) it isn’t clear that is actually what is possible. In fact, the systems seem to be prone to similar types of issues that human drivers are in many situations - and are incredibly, repeatedly, dumb in some situations many humans aren’t.
Waymo has gone full LiDAR/RADAR/Visual, and has had a much better track record. But their systems cost so much (or at least used to), that it isn’t clear the ‘replace every driver’ vision would ever make sense.
And that is before the downward pressure on the labor market started to happen post-COVID, which hurts the economics even more.
The current niche of Taxis kinda makes sense - centrally maintained and capitalized Taxis with outsourced labor has been a viable model for a long time, it lets them control/restrict the operating environment (important to avoid those bad edge cases!), and lets them continue to gather more and more data to identify and address the statistical outliers.
They are still targeting areas with good climates and relatively sane driving environments because even with all their models and sensors, heavy snow/rain, icy roads, etc. are still a real problem.
When the argument was Phoenix is too pleasant I could buy that. Most places aren't Phoenix. But SF and LA are both much more like a reasonable place other humans live. It rains, but not always, it's misty, but not always. Snow I do accept as a thing, lots of places humans live have some snow, these cities don't really have snow.
However for ice when I watch one of those "ha, most drivers can make this turn in the ice" videos I'm not thinking "I bet Waymo wouldn't be able to do this" I'm thinking "That's a terrible idea, nobody should be attempting it". There's a big difference between "Can it drive on a road with some laying snow?" and "Can it drive on ice?".
Both SF and LA climates are super cushy compared to say, Northern Michigan. Or most of the eastern seaboard. Or even Kansas, Wyoming, etc. in the winter.
In those climates, if you don’t drive in what you’re calling ‘nobody should be attempting it’ weather, you - starve to death in your house over the winter. Because many months are just like that.
Self driving has a very similar issue with the vast majority of, say, Asia. Because similarly “this is crazy, no one should be driving in these conditions” is the norm. So if it can’t keep up, it’s useless.
Eastern and far Northern Europe has a lot of kinda similar stuff going on.
Self driving cars are easy if you ignore the real world haha.
I live in the middle of a city, so, no, in terrible weather just like great weather I walk to the store, no need to "starve to death" even if conditions are too treacherous for people to sensibly drive cars. Because I'm an old man, and I used to live somewhere far from a city, I have had situations where you can't use a car to go fetch groceries because even if you don't care about safety the car can't go up an icy hill, it loses traction, gravity takes over, you slide back down (and maybe wreck the car).
Because as an old man who has actually lived in all these places - and also has ridden in Waymos before and has had friends on the Waymo team in the past, your comments seem pretty ridiculous.
A lot of the large population centres in the US are in these what you're calling "super cushy" zones where there's not much snow let alone ice. More launches in cities in Florida, Texas, California will address millions more people but won't mean more ice AFAIK. So I guess for you the most interesting announcement is probably New York, since New York certainly does have real snow. 2026 isn't that long, although I can imagine that maybe a President who thinks he's entitled to choose the Mayor of New York could mess that up.
As to the "But people in some places are crazy drivers" I saw that objection from San Francisco before it was announced. "Oh they'll never try here, nobody here drives properly. Can you imagine a Waymo trying to move anywhere in the Mission?". So I don't have much time for that.
and propellers on a plane are not strictly necessary because birds can fly without them? The history of machines show that while nature can sometimes inspire the _what_ of the machine, it is a very bad source of inspiration for the _how_.
When was the last time you had full attention on the road and a reflection of light made you super confused and suddenly drive crazy? When was the last time you experienced objects behaving erratically around you, jumping in and out of place, and perhaps morphing?
We were somewhere around Barstow on the edge of the desert when the drugs began to take hold. I remember saying something like, “I feel a bit lightheaded; maybe you should drive . . .”And suddenly there was a terrible roar all around us and the sky was full of what looked like huge bats, all swooping and screeching and diving around the car, which was going about 100 miles an hour with the top down to Las Vegas. And a voice was screaming: “Holy Jesus! What are these goddamn animals?” [0]
When was the last time you saw a paper bag blown across the street and mistook it for a cat or a fox? (Did you even notice your mistake, or do you still think it was an animal?)
Do you naturally drive faster on wide streets, slower on narrow streets, because the distance to the side of the road changes your subconcious feeling of how fast you're going? Do you even know, or are you limited to your memories rather than a dashcam whose footage can be reviewed later?
etc.
Now don't get me wrong, AI today is, I think, worse than humans at safe driving; but I'm not sure how much of that is that AI is more hallucinate-y than us vs. how much of it is that human vision system failures are a thing we compensate for (or even actively make use of) in the design of our roads, and the AI just makes different mistakes.
Self-driving is probably “AI-hard” as you’d need extensive “world knowledge” and be able to reason about your environment and tolerate faulty sensors (the human eyes are super crappy with all kinds of things that obscure it, such as veins and floaters).
Also, if the Waymo UI accurately represents what it thinks is going on “out there” it is surprisingly crappy. If your conscious experience was like that when you were driving you’d think you had been drugged.
The human brain's vision system makes pretty much the exact opposite mistake, which is a fun trick that is often exploited by stage magicians: https://www.youtube.com/watch?v=v3iPrBrGSJM&pp
And is also emphasised by driving safety awareness videos: https://www.youtube.com/watch?v=LRFMuGBP15U
I wonder what we'd seem like to each other, if we could look at each other's perception as directly as we can look at an AI's perception?
Most of us don't realise how much we mispercieve because it doesn't feel different in the moment to percieve incorrectly; it can't feel different in the moment, because if it did, we'd notice we were mispercieving.
Robots are supposed to make up for our limitations by doing things we can't do, not do the things we can already do, but differently. The latter only serves to replace humans, not augment them.
So maybe LIDAR isn't necessary but also if Tesla were actually investing in cameras with a memory bus that could approximate the speed of human vision I doubt it would be cheaper than LIDAR to get the same result.
One of the ways it's better is that humans can sense individual photons. Not 100% reliably, but pretty well, which is why humans can see faint stars on a dark night without any special tools even though the star is thousands of light years away. On the other hand, our resolution for most of our field of vision is pretty bad - this is compensated for by changing what we're looking it when we care about details we can just look directly at it and the resolution is better right in the centre of the picture.
This was only plausible to people who had no experience in robotics, autonomy, and vision systems.
Everyone knew LIDAR was the enabling technology thanks to the 2007 DARPA Urban challenge.
But the ignoramus Elon Musk decided he knew better and spent the last decade+ trashing the robotics industry. He set us back as far as safety protocols in research and development, caused the first death due to robotic cars, deployed them on public roads without the consent of the public by hoisting around his massive wealth, lied consistently for a DECADE about the capabilities of these machines, defrauded customers and shareholders while becoming richer and richer, all to finally admit defeat while he still maintains the growth story of for Tesla's future remains in robotics. The nerve of this fucking guy.
I agree that Tesla may have made the right hardware decision when they started with this. It was probably a bad idea to lock themselves into that path by over-promising.
The correct move for Tesla would have been to split the difference and add LIDAR to some subset of their fleet, ideally targeted in the most difficult to debug environments.
Somewhat like Google/Waymo are doing with their Jaguars.
Don't LIDAR 100% of Teslas, but add it to >0%.
Reportedly, they no longer use this widely - but they still have some LIDAR-equipped "scout vehicles" they send into certain environments to collect extra data.
Tesla would subsidize them and offer them at the same price as non-LIDAR models, to select customers in target areas.
And yes, you answered the second part of your own question.
The cold hard truth is that LIDARs are a crutch, they're not strictly necessary. We know this because humans can drive without a LIDAR, however they are a super useful crutch. They give you super high positional accuracy (something that's not always easy to estimate in a vision-only system). Radars are also a super useful crutch because they give really good radial velocity. (Little anecdote, when we finally got the Radars working properly at work it made a massive difference to the ability for our car to follow other cars, ACC, in a comfortable way).
Yes machine learning vision systems hallucinate, but so do humans. The trick for Tesla would be to get it good enough to where it hallucinates less than humans do (they're nowhere near yet - human's don't hallucinate very often).
It's also worth adding that last I checked the state of the art for object detection is early fusion where you chuck the LIDAR and Radar point clouds into a neural net with the camera input so it's not like you'd necessarily have the classical methods guardrails with the Lidar anyway.
Anyway, I don't think Tesla were wrong to not use LIDAR - they had good reasons to not go down that route. They were excessively expensive and the old style spinning LIDARs were not robust. You could not have sold them on a production car in 2018. Vision systems were improving a lot back then so the idea you could have a FSD on vision alone was plausible.