George Hotz: Self-Driving Cars Are a Scam
cleantechnica.com
cleantechnica.com
If the giant metal box moving around at 100km/h is killing people, then slow the giant metal box down and make it smaller and lighter. We'd get where we're going faster without them if we hadn't dedicated 75% of the land to them in the first place.
Good job.
The wrong side of the tracks thing is political and social, not literally because of the tracks.
Just recently we had to call in the military to protect the train tracks, people just constantly crossing train tracks. Trains constantly driving over hordes of people is really slowing down the economy.
If there only was some kind of technology to would tell people if its safe to cross or not.
I live in the country that most people agree has the best trains in the world. Illegal crossing of trains is basically a non issue. It barley ever happens, its almost never discussed, its not in the media much at all.
The amount of issue the country has because of trains is almost 0, with freak accidents where some teenagers swim in the river and then climb onto a railway bridge.
Cars that are being replaced are about 1000x more of a problem for safety. Trains are universally safer and every traveler that moves from a car to a train is a waste improvement in safety.
The train stations actually connect everything together.
Its also pretty easy to cross a train, mostly its 1-2 tracks and most of the time there is no train going threw. You know what isn't easy to cross, crazy insane 8 lane strodes or 20 lane highways that are constantly full of fast driving cars.
Also if trains are not gigantic American cargo trains you can have the train cross the road, the road closes for like 2 min and then the train has passed by.
A train system needs to be designed for people, not just a bulldozer driving straight threw a neighborhood and then a train that goes from one side of the country to the other go threw there. The train stops need to happen often with fast electrified trains going stopping often to connected villages and sometimes even multiple stop in villages.
How so, since that's Bloomberg Businessweek quoting Mr. Hotz? https://archive.ph/ZnqFl
> “It’s a scam,” says George Hotz, whose company Comma.ai Inc. makes a driver-assistance system similar to Tesla Inc.’s Autopilot. “These companies have squandered tens of billions of dollars.”
The companies are a scam, not the technology. By Hotz words.
Yes, that's why I asked. According to the CleanTechnica article, Mr. Hotz is specifically referring to "autonomous cars", not companies, being a scam.
> George Hotz, the serial entrepreneur who founded self-driving startup Comma.ai, has this to say about autonomous cars: "It’s a scam".
Going back to the original source I linked to above, "Even After $100 Billion, Self-Driving Cars Are Going Nowhere", still effectively says "self-driving cars are a scam", and Mr. Hotz's quote is still being used to support that thesis.
So I think you're incorrect, but I also think the article's original author took Mr. Hotz's quote out of its original context, and then the interpretation was mangled further during syndication.
It starts as follows: > In a lengthy interview with Bloomberg News, republished by Autoblog, George Hotz, the serial entrepreneur who founded self-driving startup Comma.ai, has this to say about autonomous cars: “It’s a scam. These companies have squandered tens of billions of dollars.”
Which is likely accurate reporting of what he said.
Meanwhile, pouring money into something that doesn't work is par for the course in tech. Musk saying "it's almost ready" is a scam. So I disagree with half of George's statement there myself.
> has this to say about autonomous cars: “It’s a scam. These companies have squandered tens of billions of dollars.”
Hotz says that companies have used lots of money, so to me it looks like he is referring to the companies and not the technology.
But reading a bit more, it's all to vague and the writer of the article made it really difficult to understand what's really being said here. It can be interpreted multiple ways.
Although I'm in the skeptic camp for self-driving readiness, all writing should be fair and clear about it. I'm surprised that power-assisted human control of stuff is so far behind, all we have is lane assist and stuff like that. You'd think we'd have haptic pothole anticipation or seats that cringe a little when someone roars up from behind by now.
He's right you know.
This is just not really true. Every research on this suggest the opposite.
Of course if you blame everything on driver being bad misses the point. The road system AS DESIGNED encourages and intensities bad behavior and speeding.
Its the design of the road system that leads to fatalities. Ok, sure somebody that texts is more likely to be in an accident, but if that car is going 60mph then the chance of a fatality is far higher then if the care is going 20mph.
To ever solve anything, people need to stop blaming 'reckless behavior' when people have to drive on fast moving strodes (8-lane roads) with lots of car merging in form the side.
Ironically the horrible street design in the US also makes the whole self driving thing far harder. When I as a Swiss person look at videos of Tesla driving around the US I am constantly shocked about the horrible condition and absurd roads and highways they have to drive on.
Merging into 8 lane roads with a green belt in the middle, you have to merge in from a stop sign. That's crazy, in my whole country there isn't a single 8 lane road.
If Tesla (and friends) had spend all of its time optimizing for the Netherlands or Finland they would be much better off. Cars simply drive slower, roads built smarter, and the signals and lines are better.
> With V2X technology, a car whose sensors or cameras detect a pothole in the roadway will be able to notify their drivers, giving them time to take evasive maneuvers.
Yeah lets come up with the most complex solution ever, relay on terrible backwards car makers to implement that. Just so we can never slow down cars for any reason.
Making cars slower is what makes things safer, its really not that fucking hard and the best thing is, we can do it with technology we have today.
> How simple would it be to have a system that immobilizes a car unless a licensed driver is at the wheel?
Lets introduce Stalism so we don't have to slow down cars.
> There are so many ways driving can be made safer and it doesn’t require massive investments or hugely expensive options to make safe driving a reality.
Yeah massive complex software systems with complex hardware for every participant implemented by many standards groups is defiantly the way to achieve a cheap solution.
This is why basically nobody dies on the road in Finland. People there all have government issued chip in their brain that explodes if you go over the speed limit.
Its absolutely baffling at what incredibly low level of argument we are having. Like seriously, how about we look at the places that have done a good job at reducing the amount of people dying?
> This is just not really true. Every research on this suggest the opposite.
Can you explain?
In the US in 2020, 30% ( 11,654 ) of fatalities involved alcohol-impaired drivers. 29% of fatalities involved speeding (63% of speeding drivers involved in fatal crashes were not alcohol-impaired).
https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/... https://crashstats.nhtsa.dot.gov/Api/Public/ViewPublication/...
What research suggests that fatal accidents are not largely caused by speeding, drunk driving, distracted driving, and falling asleep? It seems like just drunk driving or speeding are involved in close to half of all fatalities.
If speeding is contributing to fatalities, changing road designs to discourage speeding is a reasonable response. We don't have to claim that fatal accidents are not related to reckless behavior to make changes to discourage reckless behavior.
So the solution to all of this is infrastructure change. However the solution that are often pushed will never be successful, ad campaigns against drunk driving, education about speeding and so on will never be the answer.
What statistics like that also hide is the people that did not die because they were not there. In the US, there is much less walking and biking because the roads are designed as they are. So of course people don't die often walking, simply because nobody is walking in the first place. And not many people walk because its dangerous.
So basically the US has reduced fatal crashes by simply not making it possible to crash into anybody else by the virtue of nobody else being insane enough to use this infrastructure.
This book goes into a lot of that: https://www.confessions.engineer/
The quote is so truncated as to be useless. He could be talking about the goal of self-driving cars or he could be talking about the amount of investment that has been taken and how it's been used.