Who will win the self-driving race? Here are eight possibilities | Hacker News Reader

Who will win the self-driving race? Here are eight possibilities

arstechnica.com

11 pointsPabloRobles33 comments

"A decade from now, we'll be able to look back and say which companies or approaches were on the right track. For now, we can only guess."

My first thought was that a decade from now there will be a very limited number of self driving cars on the road, still. Mostly because people (not all people, but most people) will be very scared of the technology, no matter how good it is. Every time there's an accident it'll make news. Any deaths will make the news. It will not matter if statistically self-driving vehicles are safer.

I do wonder, can it ever be as good as or better than a person? Won't they always hesitate more, be slower than, and take more time than a real person driving? If it costs a little less to take a self-driving taxi than a person-driving taxi, why would anyone bother? What's the motivation for people to go with a self-driven thing over a person driven thing? Will there really be a big price difference?

I think this is an overly cynical take. If the technology gets there, I think there will be massive uptake for in the urban commuter and taxi markets.

when pitting Actual safety + massive convenience VS. fear, I think the former will win out for a large part of the market.

I also think the majority of people would pay more for a self driven taxi than a person driven taxi. Some reasons might be personal space, privacy, safety, and reliability in terms of dispatch and cancelled rides.

Insurance companies may well drive (pun intended) the uptake. If self-driving cars are an order of magnitude safer, insurance can be an order of magnitude less expensive. (Or, more likely, human-driven cars will become an order of magnitude more expensive to insure.)

If cities had really good public transport, and I mean really good services, plus urban planning moved on from current zoning housing / commercial / work into more mixed zone approach the demand for self-driving kind of disappears.

My bet is that we will see self driving features in an emergency intervention, where you are still in control generally but if the system predicts a crash in a future time horizon, it will intervene.

Google will spin out Waymo and GM will buy it. Mark my words.

Tesla is the only serious contender so far. They're the only company with real life cars on real life roads all over the world. It's not just the technology, it's deployment of the technology at scale. It's a platform.

Comma.ai is good too and I think it'll be a great option for adding AI-like capabilities to future and past "dumb" cars.

I see right around 0% chance for Cruze and others. They'll never figure it out. Chinese companies might though.

I think we'll see Tesla, some emerging Chinese automaker contenders, lots of current brands with smart features, and technology like Comma.ai. The future isn't just "does a car have AI?", it's the combination of the hardware of the car, battery management, routing, and platform.

Still curious about what Apple intends to do here.

Well, I think Tesla is developing FSD recklessly and their accidents will set back the entire industry for years. It’s not about “figuring it out,” real self driving under all conditions a human can handle is decades away still.

https://arstechnica.com/cars/2020/10/tesla-is-now-beta-testi...

My pet theory is that Apple won’t produce a car; rather, they’ll produce a car cockpit. A kit of standard, high quality parts, comprising everything a driver touches. So Ford would make a car, Apple would make the screen, steering wheel, mirrors, etc. Might be a trim level or package on existing car models.

Every time I drive an unfamiliar car, I see so much opportunity for an Apple-like company to fix the UX of infotainment and quality feel of volume dials, etc.

This might be the way everyone else can catch up to Tesla’s UX. Yes, Tesla has issues, but their dang touchscreen doesn’t have a half second input delay.

The biggest problem with this theory is that it is so opposite of how apple does things. Apple always controls the entire stack.

Agree, but they've also done things like CarPlay. They obviously saw a problem to solve there; I think it wouldn't be hard at all to see them produce first-party CarPlay hardware.

Waymo is the spun out version of the Google self-driving car project.

Tesla is probably the only "famous" autonomy company that is not considered a serious contender, including by people sitting in them.

> Google will spin out Waymo and GM will buy it. Mark my words.

I expect this to happen in reverse. Google will buy up some more startups in the space then eventually a traditional car maker.

Margins are too low. They'll sell the software and equipment as add-ons before they do that.

I was mentally living in the world where google didnt sell motorola after 2 years.

I can see your point and agree.

Cheers!

Can you explain your reasoning and or insight for these conclusions? Personally I see Comma.ai as more likely to "never figure it out" than Cruise and others, but I lack special insight.

The missing possibility is that no one wins generalized self-driving.

Self-driving is not a solvable or a winnable problem, because the problem or question isn’t defined. It’s a probabilistic space, where the world presents scenarios, and a computer must make a “correct” or “good” decision as often as possible, for the widest range of inputs.

Because there isn’t a grand unified formula for self-driving, I’m not even sure how we can talk about it being “solved”. Maybe Tesla might be able to self-drive on US or EU or first world clearly marked streets with 99.99% automation, assuming training data exists for all road signs and weather conditions? Is that what we mean by self-driving being solved? That might be possible in the next few years.

Other than that, doesn’t seem like there’s going to be much more than driver assist. Driving is a general intelligence problem, not something machine learning can solve - by definition ML can only work on problems it has many known solutions to.

Why is this so complicated, to me self-driving being solved is when I get into a car to get to work and don't have to touch any controls other than the AC maybe. Pretty well-defined, no?

Self-driving can still be a huge value even if it isn't totally automated between every two drivable points. I'd pay a lot of money for a car that could safely drive itself on highways alone. Get on the highway, fall-asleep, wake up at a rest stop on the other side of the country. In some ways, even better than having a private jet.

It would depend on your route to work. There are also billions of routes to work and other places that should work as well.

Maybe if enough routes work that people find it worth to buy.

How's that a missing possibility? It's in number eight.

Describing self-driving car research as a "race" reminds me of a movie: https://www.youtube.com/watch?v=F2GEKV1dOgY

All the self-driving car work you see being done today is literally rooted in a race! https://en.wikipedia.org/wiki/DARPA_Grand_Challenge

I think no one will win full self driving until artificial general intelligence exists.

For those in the future, I bet by 2035 there will be no car that can take you coast to coast with you ever having intervene.

Seems like Cruise should be on this list?

is there a good way to "short" this industry of bullshit artists, via options or something

none of these are self-driving. They are just driver assist feature rebranded and marketed well.

Tesla can't even get their cars to drive in an empty tunnel

Waymo taxis self-drive. Many of them have no support driver any more.

How is Waymo's driverless taxi service not self driving? When they come to pick you up there is no-one in the car at all, and when you are collected you are a passenger in the back not a driver.

It's limited to a single suburb, but it's definitely full self driving (unlike a lot of other offerings).

Possibility 9 — nobody does, because there is no such thing as self-driving.

The whole name "self-driving" is invented to pull the wool on public's eyes, and is a misnomer.

It's fundamentally clear to anybody moderately technically literate that what "self-driving" companies can't be called "driving," yet there are way more studied, and credited people than us here working in these companies who know that, and stay silent.

Self-driving, as in there's no one driving the car, it drives itself. No human labor proportional to miles driven. How is that a misnomer?

I would consider something capable of passing an arbitrary variation of DARPA Grand Challenge in urban traffic conditions repeatedly as something close to genuine "driving," but not the current "driving on rails" approach.

1) you clearly don't know what the cars are capable of today.

2) many (most?) licensed drivers today could not pass difficult urban driving tests without violating a zillion laws and taking totally unsafe risks that people would never allow robots to gamble on.

Reply on news.ycombinator.com