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

Waymo One: The next step on our self-driving journey

medium.com

Post reply on HN
Print view View original

51–60 of 175 posts

Re: Waymo One: The next step on our self-driving journey

#51

Post by darkerside »

jbob2000 wrote: post #8
pliny wrote: post #6

Earlier quoted context omitted.

That's a prediction, but why do you think that's how it'll turn out?

The news would be different. This news item is focused around the driver and safety. They are slowly changing their messaging from being about self-driving to being about safe-driving. They can't come out right away and say "guys we failed at self driving, it doesn't work", that would tank the personal careers of everyone important in the company. They have to slowly move the needle away from self driving, and then t…

I agree strongly that the market for true self driving is a mirage. It's being sold as new free time in your vehicle, but it'll just end up more time for ads to saturate eyeballs.

Re: Waymo One: The next step on our self-driving journey

#52

Post by pavelrub »

bertil wrote: post #11

They announced a change of brand a month ago. I guess adding “One” at the end of the existing name counts, and I like the implication that they want to communicate clearly on phases of the project but… Was it worth announcing a brand change? I feel like I’m missing something.

I'm not sure I understand. What brand change? Wyamo is the name of the company. Waymo One is the name of the service they are launching.

Re: Waymo One: The next step on our self-driving journey

#53

Post by bryanlarsen »

TomMarius wrote: post #38
bryanlarsen wrote: post #17

Earlier quoted context omitted.

Mark these words. In 10 years time we'll still be hearing the same comments about Waymo & their competitors because there still will be certain combinations of passenger, route & conditions that Waymo will refuse to take without a safety driver along. But since Waymo will handle the vast majority of rides we'll just tune out those comments. Handling all rides & conditions is a very hard problem and it'll take decades…

I assume we're talking about city driving. I'm not sure about the USA, but being in Europe, the average speed my car says I go is 28.5 km/h. Do you think we could not solve traffic for 10 years if we slowed cars down to 30 km/h? With assumption being that a city-wide driverless system would easily 'swim' through the city without encountering red lights or slowdowns due to unpredictiveness of other drivers. We could e…

I'm a big fan of Sweden's Project Zero, one leg of which is setting and enforcing a speed limit of 30 km/h on any road or street without fully separated pedestrian & cycle infrastructure. I don't know if self-driving will help accomplish this everywhere, but I'd be ecstatic if it did.

> We could even tolerate a little bit more errors because accidents would not be harmful too much.

instead of "harmful too much" you can simply use "fatal". 30km/h is a pretty magic value: below that speed pedestrian collisions are very rarely fatal. Serious perhaps, but not fatal. Above that speed they are often fatal.

Re: Waymo One: The next step on our self-driving journey

#54

Post by Hydraulix989 »

aantix wrote: post #45
11thEarlOfMar wrote: post #5

I know this is a heated debate, but it's interesting to see the divergent approaches that are being taken to achieve self driving. In one corner, it's lidar-based, equipping vehicles produced by other companies, organically and incrementally growing the navigation technology. In the other, it's video-based with now 100s of thousands of vehicles feeding data to a massive neural net. Each has very well thought-out argu…

Curious; If you train a neural net to drive a car with X number of samples, and the car does something erratic, how do you determine which data item introduced the erratic behavior and how do you go about untraining that behavior?

Gradient descent is done in batches and over the aggregate collection of the training data so it's pretty lossy. Usually the error is more of the fault of the learned model itself, and then the culprit is attributed to neuron "mis"-activations post-hoc. Often, problems take the form of generalization error instead. And yes, all of these problems are still quite unsolved, which is why we don't have end-to-end neural network-based self-driving cars, we just have rule based systems with a perception framework that just uses convnets for vision.

Re: Waymo One: The next step on our self-driving journey

#55

Post by coffeedrinker »

seanmcdirmid wrote: post #44
coffeedrinker wrote: post #39

Earlier quoted context omitted.

Because if I'm Safeway and I can get elderly people to come to my business (for shopping, prescriptions, etc.) they won't be shopping somewhere else. The business gains/retains the clientele by providing the driving service.

Just fund a minibus from the retirement home. I think many people do that already? Ok, so it’s an interesting idea to use personal transit, but it could be done already without self driving vehicles, it might be economical. But it would be pretty niche.

It's just a thought, not a business plan :)

Re: Waymo One: The next step on our self-driving journey

#56

Post by GeorgeO84 »

jbob2000 wrote: post #14
TotempaaltJ wrote: post #10

Earlier quoted context omitted.

You realize those drivers won't actually be driving most of the time, right?

Doesn't matter, it's not self-driving if you need a driver. And when that driver takes control of the vehicle away from the AI, it's going to be at a startling and dangerous moment when the AI makes a really bad decision, like directing the car towards a barrier or a stopped bus.

If the driver has large uneventful periods of time where he/she has nothing to do. He/she will not react in time for an emergency.

Re: Waymo One: The next step on our self-driving journey

#57

Post by c2h5oh »

GeorgeO84 wrote: post #49
anticensor wrote: post #46

Earlier quoted context omitted.

That car would probably be called Windows Car (not to be confused with your car's windows), stop irregularly at weird places to allow for mandatory updates, and report your height and weight to Microsoft.

Critical Update. You have 2 minutes to close all the windows and leave the car.

More like "critical update is being installed" signaled by the car hitting the breaks hard until full stop in the middle of a highway.

Re: Waymo One: The next step on our self-driving journey

#58

Post by forgot-my-pw »

11thEarlOfMar wrote: post #5

I know this is a heated debate, but it's interesting to see the divergent approaches that are being taken to achieve self driving. In one corner, it's lidar-based, equipping vehicles produced by other companies, organically and incrementally growing the navigation technology. In the other, it's video-based with now 100s of thousands of vehicles feeding data to a massive neural net. Each has very well thought-out argu…

In 2004-2007, DARPA had a few self driving contests in the desert. The documentary was an entertaining watch. IIRC, most of the cars didn't even make it to the end.

https://en.wikipedia.org/wiki/DARPA_Grand_Challenge_(2004)

https://www.pbs.org/wgbh/nova/darpa/

Would be fun to get another test with current technology to see how far we've come. Though self driving cars nowadays are more focused on safety, not speed.

Re: Waymo One: The next step on our self-driving journey

#59

Post by rsync »

cromwellian wrote: post #47

Keeping the safety drivers in the beginning, scaling up, and getting a year or more worth of actual real world use, before slowly scaling back the drivers is what I'd call a very responsible choice. IMHO, to do less would be reckless showboating that could do unimaginable damage to the AV industry, as well as end up endangering lives. In the tech industry, we don't like to wait, we like moving fast, taking risks, shi…

"Keeping the safety drivers in the beginning, scaling up, and getting a year or more worth of actual real world use, before slowly scaling back the drivers is what I'd call a very responsible choice."

A whole year! Truly sage-like wisdom from the MoveFastBreakThings generation of technical development.

It is simultaneously comical, exasperating and sad to watch this tragedy unfold.

We will slowly build our own, modern Deuteronomy - instead of forbidding the consumption of shellfish or stipulating gifts for former slaves, ours will talk about agency and skin-in-the-game and aligned incentives ...

Meanwhile, there will be a multi year (decade ?) period of experimentation where we will all be terrified as we drive at highway speeds and another vehicle approaches that road at a right-angle. Currently, I know what that vehicle will do, because all other choices lead to death-of-driver.

But in this near (and short lived) future, that other vehicle could do anything.

Post reply on HN