Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Welcoming our first riders in San Francisco

blog.waymo.com

Post reply on HN
Print view View original

391–400 of 691 posts

Re: Welcoming our first riders in San Francisco

#391

Post by rhacker »

burnished wrote: post #303
actusual wrote: post #111

Earlier quoted context omitted.

Definitionally, the achievement of self driving vehicles does not require AGI. Doing one task very well requires a subset of AGI called Weak AI.

Part of driving well requires a diverse of array of abilities, right? You know what is litter and what is debris because you can make a guess about material properties based on some observations, like looking at something that moves without being touched and is translucent, you probably conclude that its a plastic bag of some sort and not a hazard. Similarly you probably use a wealth of experience to judge that a sma…

I think the engineer's answer to the child entering the roadway would be: The car SHOULD never drive at such a speed that if the child WERE to enter the visible zone that it could swerve+slow enough to not hit it, forget the toy. After that we can move the goal posts and say it's a FAST child on a bike - but then the reasonable solution to that is a human driver may have also hit the biking child. Then, of course, we get into the ethics of fault for the accident.

Re: Welcoming our first riders in San Francisco

#392

Post by whimsicalism »

standardUser wrote: post #140
noneeeed wrote: post #79

As someone who lives in Bath, an old european city with roads (and other drivers) that can give experienced human drivers a nervous breakdown I'm rather more interested in how they do in SF than their current deployment in Phoenix. From what I've heard repeatedly, SF sounds much more irregular and messy than Phoenix, so it should be something of a stepping stone to making them usable more widely if they can crack it.…

The streets of pre-pandemic San Francisco were an unmitigated clusterfuck, arguably the worst in the nation at its 2019 peak. I imagine it's much easier these days for an autonomous vehicle.

> arguably the worst in the nation at its 2019 peak

Not even close - I've lived in the DC metro, Boston, and LA and all three are certainly worse than SF, and I think that is also backed up by evidence.

Re: Welcoming our first riders in San Francisco

#393

Post by mdoms »

ra7 wrote: post #82
bluGill wrote: post #38

Earlier quoted context omitted.

It may not scale to rural areas though. There are some roads were you don't need to look at the road in front of your: it is there and nothing else is. Instead you need to watch the ditches in a wide area around because that is where wildlife will jump out of in front of you.

Is there a market for taxis in the rural areas? They have little incentive to expand there if there's no money to be made.

Absolutely! I live in outskirts and I would LOVE to be able to get a taxi to the pub and back! Unfortunately they don't service me here.

Re: Welcoming our first riders in San Francisco

#394

Post by oakfr »

panick21 wrote: post #170
pbreit wrote: post #89

Earlier quoted context omitted.

Most of the commercial value is achieved at Level 4. I'm curious what all the folks who claimed this was decades away are thinking?

Most commercial value is at Level 2 for quite a while.

Level 2 is becoming a commodity. Its commercial value will fall to zero very quickly.

Re: Welcoming our first riders in San Francisco

#395

Post by whimsicalism »

Ftuuky wrote: post #122
noneeeed wrote: post #79

As someone who lives in Bath, an old european city with roads (and other drivers) that can give experienced human drivers a nervous breakdown I'm rather more interested in how they do in SF than their current deployment in Phoenix. From what I've heard repeatedly, SF sounds much more irregular and messy than Phoenix, so it should be something of a stepping stone to making them usable more widely if they can crack it.…

Same with Coimbra, Portugal. Would love to see a Waymo car trying to navigate those old intricate roads.

I wonder if Waymo can use street mirrors.

Re: Welcoming our first riders in San Francisco

#396

Post by ctvo »

ElFitz wrote: post #338
sschueller wrote: post #123

If Google can't do it with its massive resources and sensors don't ever expect a Tesla to perform better than this. It's time for Elon to pay the piper and refund all users who were hoodwinked into purchasing FSD. Especially people who's cars are nearing end of life.

While I do get your point and have my own doubts, I fail to see how this is a fair argument: > If Google can't do it with its massive resources The companies behind the SLS had and have an enormous amount of resources. Yet, they have failed to even launch it. Arianespace has a huge amount of resources and experience too. Yet, when asked, they said landing a rocket couldn’t be done. Then, back-pedalling, that it was p…

> Yet, SpaceX has developed three generations of rockets (Falcon 1, 9 and heavy), builds them, launches them, iterates on them, lands them, reuses them, and is now working on a fourth rocket (SuperHeavy).

What does SpaceX have to do with Tesla? They're owned by a billionaire with his attention split, but otherwise are structured differently, are in different industries, and most likely have different cultures.

The cult of worshipping a single person has got to stop.

Re: Welcoming our first riders in San Francisco

#397

Post by arduinomancer »

pbreit wrote: post #89
espadrine wrote: post #47

Waymo is currently the only SAE Level 4[0] self-driving car in use, right? The wheel must not be used. Even though it is limited to two cities. Their technology seems an order of magnitude above the rest, although factoring in its cost could make it look worse. (Can it turn a profit, when including the development costs?) Tesla FSD is at best SAE level 3 (can need human fallback), and at worst level 2 (needs constant…

Most of the commercial value is achieved at Level 4. I'm curious what all the folks who claimed this was decades away are thinking?

How is there commercial value if you still need a driver in the car?

Re: Welcoming our first riders in San Francisco

#398

Post by njarboe »

paganel wrote: post #227
VelkaMorava wrote: post #197

Earlier quoted context omitted.

I wonder where "a country road with no lanes which barely fits 1.5 car in winter in the Czech republic" is on your scale... Something like this, just imagine the snowdrifts around it https://www.google.com/maps/@49.080269,16.4569252,3a,75y,307...

Or an ambulance going on the opposite direction (because that’s the only available choice) on a boulevard in a busy capital city like Bucharest. Saw that a couple of hours ago, the ambulance met a taxi which was going the right way but of course that the taxi had to stop and find a way for the ambulance to pass (by partly going on the sidewalk). I said to myself that unless we get to AGI there’s no way for an “autono…

People at Tesla and other autonomous driving companies, of course are aware and worry about such situations. If you have a few hours and want to see many of the technologies and methods that Tesla is using to solve them, check out Tesla's recent "AI day" presentation. Tesla is quite cool about openly discussing the problems they have solved, problems they still have, and how they are trying to solve them.

An incomplete list includes:

1) Integrating all the camera views into one 3-D vector space before training the neural network(s).

2) A large in-house group (~1000 people) doing manually labeling of objects in that vector space, not on each camera.

3) Training neural networks for labeling objects.

4) Finding edge cases where the autocar failed (example is when it loses track of a vehicle in front of it when the autocar's view is obscured by a flurry of snow knocked off the roof of the car in front of it), and then querying the large fleet of cars on the road to get back thousands of similar situations to help training.

5) Overlaying multiple views of the world from many cars to get a better vector space mapping of intersections, parking lots, etc

6) New custom build hardware for high speed training of neural nets.

7) Simulations to train rarely encountered situations, like you describe, or very difficult to label situations (like a plaza with 100 people in it or a road in an Indian city).

8) Matching 3-D simulations to what the cars cameras would see using many software techniques.

Re: Welcoming our first riders in San Francisco

#399

Post by bluGill »

jeffbee wrote: post #60
bluGill wrote: post #38

Earlier quoted context omitted.

It may not scale to rural areas though. There are some roads were you don't need to look at the road in front of your: it is there and nothing else is. Instead you need to watch the ditches in a wide area around because that is where wildlife will jump out of in front of you.

The superiority of a blended computer vision system for this task, over a human performance, is almost impossible to overstate. The computer is not going to overlook even one deer.

It isn't possible to not overlook deer because they are often doing things such that you cannot spot them. Unless you mean they won't fail to see a deer 2 meters in front of the car - but it is too late to do anything about it then.

Re: Welcoming our first riders in San Francisco

#400

Post by whimsicalism »

boplicity wrote: post #180
MattGaiser wrote: post #49

Earlier quoted context omitted.

I assume that they just don't solve the really wacky edge cases which might require AGI. Many people assume that a self driving car must be 100% successful. To me it is sufficient to beat the top 20%/bottom 80% of drivers. So the car will crash if a seagull flies in front of the camera and there is a deer just beyond it. The human would probably crash too.

> To me it is sufficient to beat the top 20% of drivers. Computerized safety systems can, should, and are raising the bar, in terms of safety for the top 20% of drivers. "Self-Driving Cars" aren't competing in a static environment, similar tools are raising the safety bar that they have to compete against. This means it will get increasingly difficult to beat non-self-driving vehicles.

Don't let silly little empirics around traffic fatalities in recent decades get in the way of a good narrative like this one!
Post reply on HN