Classic. I love it. My kind of engineer.
When Tesla ships an autopilot on mass market cars that fails in edge cases, commenters are up in arms that it wasn’t tested to perfection in every scenario. Big companies are punished if they don’t deliver perfection.
When an underdog company hacks together an autopilot proof of concept and takes a reporter for a ride with it, they’re heroes for pulling off a technical feat like that. Underdog stories will always draw applause.
The challenge with a company like Comma is that they can’t maintain underdog status forever. The product is very impressive in the context of an underdog hacker success story, but outside of a few early magazine shootout wins it just can’t hang with the efforts of the big companies throwing huge budgets at their own solutions. This puts them in a difficult spot because the underdog-hacker story can’t scale forever.
If anything, your comment is proof.
It is absolutely the wrong attitude for self-driving cars or anything that is safety critical.
Why do people keep printing things like this, which are objectively wrong? I have had a completely autonomous waymo come to my location, pick me up, and take me to another location.
That didn't exist five years ago, it does exist now. How is this not "closer" than it was 5 years ago when it literally exists now, and didn't exist then?
Not that the technology hasn't improved, but with these things there might be many factors involved that might answer the question "why we have this today and not 5 years ago".
Self-driving is still incredibly limited and progress is often overstated because people make headway on some tiny issue. A thing I always liked for people who think progress is rapid, this is Germany in the 1980s where Ernst Dickmann had autonmous cars drive thousands of miles: https://youtu.be/_HbVWm7wdmE
No, it would be like if somebody said "some day we will travel to the moon", and then after the Apollo missions there were articles being published that said "we are no closer to traveling to the moon than we were 5 years ago".
https://www.youtube.com/watch?v=AHdKm0kW4l0
This is a video of a person riding in a fully self driving car.
----
Eliezer Yudkowsky is fond of shitting on the AI developers of the 1960s for thinking they could write `APPLE` in the source code of a symbolic language and that that made them weeks away from a human intelligence which could reason about apples, and how simplistic that looks now.
Like YouTube auto-transcribed subtitles are useful but they are obviously transcribing sounds without understanding, they lack understanding of where the context indicates that a spoken thing should be a name, or they will transcribe the same word two different ways in two different sentences with no understanding that it was the same object as before being referred to again, or where a sound is unclear I can fill in what was intended but the auto transcriber can't, or I can see from lip movement that the transcription was wrong, the audio processor can't integrate multiple inputs in that way, and they will transcribe sentences which are grammatically correct but human background knowledge of the world tells you it makes no sense.
Similar with self driving cars, it's pretty clear from the outside that you can't have a car which can reason about the state of a city, its roads, the things in the roads, the environmental conditions, without having a large amount of interconnected human level background understanding of the world and the things in it. e.g. not just seeing a shape and identifying it as a cyclist, but knowing that you passed a cyclist a few seconds ago and now you are slowing down for traffic lights the cyclist will be coming back alongside you momentarily. Not just identifying a parked car, but seeing a car stop moving and turn its lights off as it parks implies the doors are about to open. Not just seeing lane markers in the road, but seeing no lane markers and being able to complete the pattern of where the lane markers should be because you understand how humans design roads. Not just seeing rain and slowing down, but the hinkiness feeling of "these conditions are dangerous" from the way other cars are driving, the road conditions, and slowing down in advance of anything objectively happening because you predict what could happen. Not just seeing a sign saying 'Diversion' but being able to look around expecting to see the next diversion route sign either down this turning or up ahead by another turning, and using that extra information to decide what to do. Not just identifying an erratically moving vehicle when you see it, but hearing a siren and seeing a flash of blue in the mirror and thinking ahead that an ambulance is coming and then looking for places to pull over to let it past and expecting the cars around you might move like that as well. Not just seeing the car in front slowing down, but seeing the driver inside it move and understanding that they ware waving you past because they are double-parking to drop someone off or pick someone up instead of slowing down because of traffic. And countless other situations.
Humans have good reaction time when it comes to touching something hot and pulling our hands away before we understand and are aware of what happened. Sensor equipped cars have good reaction time when it comes to ultrasound sensing a thing up ahead and applying the brakes without understanding what's happening. Humans have bad reaction times when driving because we can't feel the thing in the road, it has to go through our slower higher level thinking to understand what's happening before we can choose to respond.
Self-driving cars, then, are either the pretense that you can put a human level AI on top of the car's unconscious reactions, without compromising the reaction time, to get a superhuman level driver. And that's not something you can do because human level AI doesn't exist. Or they are the unfounded claim that you can drive through humanspace without human understanding, which is about as convincing as saying you can send a machine to the butcher, baker and candlestick maker to do your shopping without it having any AI. As soon as anything goes off-plan the robot is stuck. And you get into "well, we'll hard code a workaround for this situation and simply enumerate everything which could go wrong in a decision tree". Shop door closed with a sign saying "please use other door"? Hard code that, OK now are we good? Shop door closed with a sign saying "please ring bell for attention"? OK, hard-code that, now are we good? Shop door propped open with a mop and bucket and a sign saying "caution, wet floor"? OK, hard-code that, now are we good? Butcher says "sorry we have no liver but we're expecting a delivery in 5 minutes are you OK to wait?"? OK, hard-code that, now do we have AI? And then you get to Amazon which controls the warehouse layout, temperature, environment, shelving, can put tracks in the floor, put all items into regular sized boxes tagged with machine readable labels, which is more analogous to trains and trams on rails, and still Amazon use humans to pick and pack things.
Huh, so are you saying maybe we're a little close to autonomous cars than we were 5 years ago?
Do you expect to wake up one day and have self driving cars work in every city? That's just not what today's technology can accomplish. You either end up with broadly applicable L2/L3 (Tesla, Comma) driving, or you get narrow scoped L4 driving.
The scope of L4 widening is a real change.
As long as the remote operators and assistance teams are an order of magnitude smaller than putting a driver in every car, then it's close enough to autonomy to count as "closer" and to be useful.
"As long as the remote operators and assistance teams are an order of magnitude smaller than putting a driver in every car" - the entire gig is way to expensive and requires "time travel" level of scientific achievements, which is 100% fiction and 0% reality.
No, but it does qualify for "closer to reality today than they were 5 years ago"
> Pattern recognition software alone (A.I.) would never be able to match human driving performances.
That's okay. A trained human can do much better than necessary, and geofenced pattern recognition software doesn't have to be as good, especially because it should have better reaction times and braking force than a human.
> "As long as the remote operators and assistance teams are an order of magnitude smaller than putting a driver in every car" - the entire gig is way to expensive and requires "time travel" level of scientific achievements, which is 100% fiction and 0% reality.
Why?
If you can run a fleet of 300 cars with 30 people, that's already enough to make tons of money once you get well-established. You don't need any scientific improvements for that, let alone the ones you're exaggerating.
What are the steps back?
They're slow but they're improving. And they don't need to reach their original lofty goal.
> "Why?" - It's unsustainable
Sorry, the "Why" was directed at the level of scientific achievement you claim they need.
"the level of scientific achievement" - every single step, every single minute and every single individual (the financial input), is prohibitively expensive for this R&D project, and it is not justified by any means by the results (the financial output), Companies and investors don't care about progress. They care about profits, and, in case progress would stay in their path to make profits, they'll fight against it. You should check waymo salaries, hardware prices, operations costs, and fleet management costs. From operational POV, every mile covered by those vehicles translates into a price payed by the company, money that are not recovered whatsoever at this point. Vehicle lifecycle, insurance, maintenance, cleaning and the electricity used, adds up very quickly and could go as high as half a billion dollars per year - "Argo has about 1,300 employees and is likely burning through at least $500 million a year, industry participants say." (https://www.theinformation.com/articles/argo-ai-planning-pub...). Now remember how in business, any investor usually expects to make 10 times his or her investment, in this case (the Argo.ai example) meaning that the profits (after all expenses and taxes are substracted) to be around $5 billion per year. This is the reason why Ford decided to shut down Argo, which was burning half a Billion a year with no end in sight. To directly address your statement - the scientific level needed would require way too much money to justify the road to accomplish it. Basically, all those parts interested either do not have those money, or are part of a business model that requires substantial returns on a relatively short term, and cannot afford to finance projects with constantly moving delivery dates for fictional ideas.
What about the regulators that allow drunk and distracted drivers everywhere
In what places is drunk driving legal? The laws exist and are rigorously--if imperfectly--enforced everywhere I've ever lived.
Are folks building transportation businesses employing drunk drivers?
Vs autonomous driving where none of those impediments come into play
If planes were designed by safety committees we would have never figured out how to make them light enough to fly.
If you want to help US pedestrians you'd be much better going after SUVs and pickups https://www.webmd.com/first-aid/news/20220318/turning-pickup...
With your attitude, we would never have allowed cars in the first place.