Hacker News Reader

Comment by espadrine | Hacker News Reader

Contrast Google's old way of managing bets with Tesla.

When Google started working on self-driving cars, they went about it academically. They did not plan a sequence of stepping stones that they could sell. They meant to have a product in an indeterminate future which should be immediately perfect and better than a human, essentially not needing a wheel.

Tesla went about it with an engineering perspective. First, cars that can send accumulated data over the Internet. Then, equipped with cameras. Then, with limited assistance, warning the user to take the wheel back in difficult situations. And sporadic updates adding support for more complex cases.

The end goal is identical; but they make money along the way.

View on HN

> They meant to have a product in an indeterminate future which should be immediately perfect and better than a human, essentially not needing a wheel.

This isn't what happened. The first iteration of Google self-driving cars, the modified Lexus', had wheels alright. What Google found and told the industry was "almost autonomous seems to be more dangerous than not autonomous at all", explained by the driver getting bored and complacent. This isn't surprising to people doing Human Factors, and while Tesla certainly made money off ignoring it, the result has already been tragic.

Ugh. The sooner Tesla pass by the "we're safer than human drivers" mark the better because hearing stories about how a single guy who died, while absolutely tragic, is getting tedious when put in context of the the vast number of regular deaths on the road. My heart only bleeds so much for a guy I never met, who took a chance that he knew the risks of, when considering the grander goal at play here.

The problem is, a human driver wouldn't have had any trouble avoiding that accident -- the truck was visible for some distance.

It's not just about the overall fatality rate per mile. Self-driving cars also need to not have any accidents that humans would have easily avoided. Until they demonstrate that, they're not going to be widely accepted.

Teslas don't come equipped with a LIDAR, that cruise, uber, udacity and google have been adding as a very important sensor.

Lidar sees a 3D cloud around the car with almost a cm accuracy. Better sensors could have most definitely prevented the accident.

However, to add to tesla's point. I have a car with adaptive cruise control. Not having to constantly check distance with car head has probably saved me once or twice from a crash.

Good auto-pilot's are definitely going to reduce human deaths on roads. Not only just deaths, but the amount of traffic all of us have to endure because someone just didn't keep the right distance from the car infront and bumped into them.

I'm all for open smart driving technology.

Yup, but I hope you see that an always-on driver assistance system is really different from a completely self-driven car.

I see so many people (like the one two replies above) push the argument for the latter using the advantages that even the first would give.

> Self-driving cars also need to not have any accidents that humans would have easily avoided.

Why? If they're overall much safer, then that seems like a net win to me—even if their failure modes are not identical to human ones.

I could be wrong, but I just don't think people are going to accept that. It's related to the illusion of control that driving provides: everyone thinks they're a safer-than-average driver. So even if self-driving cars are safe enough for everyone else, they're not safe enough for me.

There will, of course, be exceptions, as we've already seen (poor guy).

And my point wasn't "boohoo look at Tesla". It was that, contrary to what OP claimed, Google made the explicit decision to shoot for fully autonomous after having tried what Tesla did (and with LIDAR-equipped cars!).

Never understood this argument "s getting tedious when put in context of the the vast number of regular deaths on the road."

What are you comparing? a 70,000$ car with better safety standards to a 10 year old car! A car that's aimed to be completely autonomous (with driver assistance features) compared with a car completely human driven?

Can you not think beyond extremes? Here's a simple example. An always-on "Autopilot" (or any other driver assistance system) that only intervenes if the driver does a tragic mistake. We still have the important general intelligence of a human and the faster and reliable response of a machine.

"explained by the driver getting bored and complacent. "

I think it's more than that. I think driving requires a general level intelligence, and not a rule based AI.

For example, driving in the suburbs, you see a ball bounce in front of you. Expecting a chasing kid or a dog, you come to a stop. This isn't something a completely self driving car will do.

On the other hand, from the demos I've seen of "what the car is seeing", it looks like the car would probably know the child is running in his direction even before the ball passes in front of it.

We use general intelligence to deduce what we don't see. But the car sees much much more than we do.

Yes, what about the case where the child is hidden from the view by a street parked car.

My point is not a specific example, but the importance of general intelligence.

It will be interesting to see what end-game is for Google. Is it to provide the hardware specs to car manufacturers for cameras/components and use a Google operating/control system?

In the most recent quarter, Tesla posted a $22 million profit, which was their first profit since first quarter 2013. There have been many times in Tesla's history where things were tight and risky. Things are looking good now, so it is exciting and Tesla seems to be going in a great direction.

It all comes back to Google having tons of money and thus being able to do things Googley.

It also helps when you're already making cars, which Tesla was.

True. That said, Apple's phone was successful against established manufacturers, and Android won without having to build a phone at all.

iPod to iPhone was certainly not 0 to 1.

I'm an electrical engineer and know people who were on the first iPhone team. I think it was about as close to 0 to 1 as you get in a big company like that.

I worked for a major RF test and measurement company when the first iPhone came out and everyone was blown away that a company could design and tool up production for a cell phone without tipping anyone off in the industry.

Compare Apple, a consumer electronics company, building a phone to Google, an ad tech company, building a self driving car.

One seems to leverage the core competencies of an organization better than the other... One seems closer to starting from 0...

I mean, google is not quite going from 0 to 1 there. it's a machine learning problem and google has some of the brightest machine learning people out there already working on search.

That's just because you don't understand Google.

It was a ML company going from building ML for the internet to building ML for a car.

Google hired the team that won the Darpa Grand Challenge.

The iPhone was released in June 2007. At that point, the iPod models were Classic (click-wheel), Mini (click-wheel), Nano (click-wheel), and Shuffle (screenless).

The first iPod Touch wasn't until September 2007, 3 months after the first iPhone.

When the iPhone 1 was released, the iPod range didn't have much on any other MP3 player. All the things that make an iPhone an iPhone - the touch-screen, multi-touch, iOS, didn't come from the iPod.

Perhaps iPod to iPhone wasn't quite 0 to 1, but it wouldn't have been much more than 0.001 to 1.

COPY: Compare Apple, a consumer electronics company, building a phone to Google, an ad tech company, building a self driving car.

One seems to leverage the core competencies of an organization better than the other... One seems closer to starting from 0...

But those are different classes of bets, no? I mean, Tesla makes cars, and a self-driving car is still a car. Google's primary business has nothing to do with cars. Their "X" bets are in completely orthogonal fields from their primary business. They are (in theory, anyway) hedges against a complete collapse of that primary business, while Tesla's is more evolving their primary business to position it for the future.

Which is not to say that Tesla is Doing It Wrong -- I personally am a much bigger fan of focused, iterative improvement than of wasteful throw-spaghetti-at-the-wall "moon shots." It just feels unfair to compare the two.

Also, Glass was Brin going off script.

I wonder if he bought into the press likening him to Tony Stark...

Every company has it's misfires - before Glasses was Wave, Microsoft Bob, Win ME, Apple Newton and that hockey-puck mouse.

The fact that he was screwing Glass's marketing manager probably helped.

> The end goal is identical; but they make money along the way.

Not really, there's a fundamental difference in the approach and desired result. Tesla is going for Level 3 autonomy (some human help needed on trip), Google attitude is more of Level 4 or bust.

There are pros and cons to both approaches.

googles car also costs twice as much to make due to the expensive sensors

Tesla's business strategy is a moving target. I don't think anyone put the words autonomous driving and Tesla in the same sentence until 2014. And now they turned into a solar power company all of a sudden.

Having a car drive itself was more tough/challenging than big or worth doing. Why didn't they try for something like teleportation? That's tough/challenging, big, and worth doing, though I bet some idiot would say, "no, because it would wreck the economy" or "no, because it would put people out of work."

> but they make money along the way

Only killing a few people in the process, but hey who am I to judge?

Please don't toss off inflammatory, unsubstantive comments on HN. If you're going to comment here, make it substantive, and if the topic is divisive, please make it less inflammatory, not more.

Edit: you also became uncivil downthread. We ban accounts that do that. Please (re-)read the site guidelines and follow them:

https://news.ycombinator.com/newsguidelines.html

https://news.ycombinator.com/newswelcome.html

Are you implying that Tesla killed the (1 person, not people) who decided to let his car slam into a tractor-trailer on cruise control?

It's not a problem to you that, say, Apple has "financed the suicides of dozens of Foxconn workers," or that GM killed 140 people by suppressing information on a defective ignition switch, or that Lockheed Martin and Halliburton buoy their stock by selling weapons to kill hundreds of thousands of people?

Evidently, you must find totally acceptable the gun-running, terrorist-sponsoring activities of HSBC in the last few years. I bet Google's permabanning of people from their own Google Apps accounts (because they were reselling a couple Pixel phones) is ethically fine with you, as well.

Never mind the tens of thousands of preventable automobile deaths yearly because car manufacturers refuse to implement even the most rudimentary of Autopilot functions.

Yeah, totally reasonable to pick on the electric car / solar battery company with $3b in revenue since one dope decided to drive into a truck and behead himself. There's nobody else deserving of criticism whatsoever.

Let's not even get started with how cocky their CEO has been with his fancy sustainable business plans!

p.s. I won't even attempt an ad hominem at the guy selling homeopathics and 'nutritional supplements' in this thread.

This is uncivil and stoops to personal attack. It's just the ante-upping, signal-noise-lowering sort of comment we ask people to abstain from here. Please don't post like this to HN.

https://news.ycombinator.com/newsguidelines.html

https://news.ycombinator.com/newswelcome.html

I'm going to ignore the whataboutism you spew.

Also, Elon musk has done amazing work with electric cars, sustainable energy and space travel. No doubt about that.

But this is about Autopilot, a collection of driver assistance program. Not sure about OP, but my spite with Tesla is that they are pushing the above-mentioned driver assistance programs as self-driving features. This over-selling is going to kill people.

And for your information, lane keeping assist, lane driving assist, automated cruise control (which is pretty much AP1.0) have been present in several cars. Even self parking has been demoed by several cars. Add in automatic lane changing (which doesn't sound difficult to me), it wraps up Enhanced AP1.0.

Does that sound like a completely autonomous self driving car? Because that's what the public perception is like.

Please don't call names ('you spew') in arguments. If someone is wrong, explain so civilly.

https://news.ycombinator.com/newsguidelines.html

https://news.ycombinator.com/newswelcome.html

How is "you spew" namecalling?

I was the one personally attacked in this thread, not the other way around. Just because my comment doesn't fit the Musk Worship narrative doesn't make me wrong.

Was someone not killed by a malfunctioning / underprepared Tesla? What am I missing here????

"You spew" is a pejorative. Examples:

https://twitter.com/search?q="you%20spew"

https://www.google.com/webhp#q=site%3Atwitter.com%20%22you%2...

What you're missing is that even though someone was uncivil to you on HN, you needed to remain civil in return. That's the social contract here, and we need it to prevent discourse from degenerating even further.

Given that there is no report button or ability to downvote someone that replies to you (at least not on mobile for me), it is only reasonable to defend myself in that situation.

The ensuing attacker went to my profile, researched me and my business, and personally attacked me because they disagreed with my sentiment.

What is my recourse for action in such a case? Take it on the chin like a wimp? I don't think so.

If you don't provide appropriate tools to stop a tit attack, then you are fostering an unfair and "unsafe" environment by banning tit-for-tat response.

You are turning this community into a one-way echo chamber.

Your recourse is to email hn@ycombinator.com, as the site guidelines ask. Note that I chided that person for what they did, which of course was wrong. Two wrongs don't make a right, though, and you're continuing to break the rules while they at least stopped.

For civil discussion we all need to hold ourselves to a higher standard than the one we perceive other people to be keeping, because the only alternative is a downward spiral, and an accelerating one. If you want to keep commenting here, we need you to internalize that.

Reply on news.ycombinator.com