Hacker News Reader

Comment by andreyf | Hacker News Reader

From a marketing perspective, it's a greater blunder still: asking people to think about what they'd want Glass to do it will quickly set expectations well beyond what's technologically feasible for an initial product. That and the wishful concept video almost guarantee the initial product will be a disappointment. And that's a damn shame, because given Google's talent, it'll nevertheless be a engineering feat of epic proportion.

That said, perhaps Google doesn't care about releasing perfectly marketed hit products a la Apple. If wearable computers catch on that make it easier to check-in to the places you go and +1 the stuff you like, they'll make tons more money from useful targeted ads than units of hardware shipped, just as they do now.

View on HN

"And that's a damn shame, because given Google's talent, it'll nevertheless be a engineering feat of epic proportion."

You know, I have respect for the people Google employs - they are certainly a very talented bunch. But what has Google ever engineered, at least physically, that has set your expectations so high?

I think the technical infrastructure they've built to support Youtube, Google Search, and the rest of their products is a modern technological marvel, and that was designed 100% in-house as far as I know.

I don't think they've developed a physical consumer product that is revolutionary, but I believe they have the talent to do so.

Everything I've heard about the self driving car gets me more excited than Search ever did. But then again I've never seen it.

But what has Google ever engineered, at least physically, that has set your expectations so high?

Driverless cars.

Quite so! I'm very excited to head on down to my Google dealership and pick one--

Oh.

There's a huge difference between a research project and a shipping product. I'm most excited for this development as well, but I'll be withholding my acclaim until they actually deliver it. So far, all of this stuff is an elaborate Montessori school science project.

If you've followed the DARPA Grand Challenge since 2004, you'll realize that it's _not_ just a school science project. The Google car is going to be ready for market far before the legal, insurance, and other non-technical issues are sorted out.

In other words, they could ship "now" if they weren't doing such a revolutionary product. If you make a trivially incremental improvement on an already-existing product (iPad3) then that's a lot easier

That still sounds silly. If you have the muscle of Google, wouldn't you get the legal ball rolling first, so that when your car is ready, you can sell immediately?

Does a driverless car even count as basic research that they can spin off into a consumer product?

Which they did. Nevada now has laws governing autonomously piloted cars, and California is working on it, last I checked.

If they can get such vehicles legal and available to the general public in California, it will be legal everywhere else within a year or two at most. California sets or has had the ability to set a lot of standards for the country, from vehicle emissions, textbooks, to now, apparently, driverless cars.

I think you're simplifying a little.

A driverless car is not just another car. It's not a fancy smartphone. It's a paradigm shift that will reshape society and erase tens of thousands of jobs in a relatively short timeframe. And it's one of the rare cases where you can actually predict and quantify these consequences with relatively good accuracy.

I'd assume quite a few parties (read: all of them) will want to have a say in such a rollout, beyond technical peculiarities. For the better or the worse.

ISTR people saying that about the Segway.

You're not seriously comparing the segway to driverless vehicles?

Nah, I'm just comparing that bit of hyperbole.

Like the "paradigm shift" that was to be the Segway, what problem does this solve exactly?

40,000 deaths annually due to driver error

Also: Cost of human labor in logistics. Driverless cars don't need sleep.

Driverless cars would more or less kill an entire job category overnight. Who wouldn't want a truck driver that doesn't make wrong turns, doesn't get distracted, forget to gas up, doesn't need to sleep, eat, or be paid (as much)?

Don't forget: Offloading almost all freight traffic that can't be done by rail to automated vehicles between 1-6am daily.

Airplanes fly constantly, because each second they're on the ground is a second they're not paying for their capital cost. Ever see how barren a 6 lane highway is at 3am in the morning on a weeknight?

Exactly, robots aren't constrained by a need for sleep; and is likely more aware of it's surroundings vs. a human facing the same conditions. A driverless car can "see" in infrared or radar just as easily as it can visible light (potentially better).

The challenges in getting a driverless car to market will take years to over come. lsiebert replied that a few states are already thinking about this. We'll see how that turns out.

In the mean time, isn't it prudent to have intermediate products? Products to help people get used to the idea of driverless vehicles? Products to give them experience on how to rollout their bigger vision?

That's the problem with the google boys, they're a one hit wonder. They lucked out with search, a research project, and now they think they know what they're doing. They need to take their bozo hat off and realize they need to partner with the rest of the world to get stuff done. Apple's success is not just due to Job's taste but also extensive ties with chinese manufacturers and real focused hard work. A "secret" X lair in the second most expensive city in america ain't gonna cut it.

That's they problem with google, they still haven't figured out they're a one hit wonder, real business is much harder edged than the "playtime" for kids that goes on there.

> That's the problem with the google boys, they're a one hit wonder.

You don't consider GMail, their dominance in ads, Android's ubiquity, the YouTube acquisition, etc as successes? Because all of those seem to have improved vastly since Google took control of them.

Given that Google is not about search anymore but about Google all those acquisitions make sense. Now Google totally controls the biggest assets online (not counting facebook). They could also know as much now as they would like to know about you to give them, data-wise, more than just a controlling position online. So the seperate entities don't even have to be profitable if they generate enough data exhaust for the mothership (what they clearly do).

Google is not just about data collection, though data collection does help improve products (think distributed spam filtering). Most Google products exist because someone at Google thought people would like them. Making money is good, but not a primary concern. It seems crazy from the outside, but really, developers at Google are not sitting around trying to figure out how to build a huge and scary profile about you. We just want to play with fun things.

(Yes, that generates a lot of data, and we do use a lot of that data for things that might not thrill you, like targeting ads. But ultimately, a lot of companies know a lot about you, and few try to be as transparent about it as Google does. At least you can stop giving Google data. Try to opt out of the credit reporting system.)

Google has a competitive edge in search because of it's data. All their other so-called successes are based directly off that (either in user share or because they have the big bucks to bankroll it).

Take a look at other tech companies, they've created fortunes based on almost nothing

* apple, consumer electronics that anyone could do

* amazon, selling stuff online that anyone could do

* oracle, selling databases when their competition includes ibm and even free software

Youtube should have been itunes, amazon instant, netflix, hulu all rolled in one. But they're not, because they don't actually know how to do business - one hit wonder.

Another example, google, until recently, had no lobbyists - why on earth would you do that? (when you have enough piles of cash to buy ountries). Investors should be filing lawsuits for this reason alone.

Another example of business cluelessness, they're not flexing their patent muscles.

Android ubiquity - linux is ubiquitous too, far more so - hint, it's FREE. Not very hard to give away free stuff.

Another example of business cluelessness, they're not flexing their patent muscles.

It's worth noting that Google's prospectus [1] as filed with the SEC says that Google will always pursue long-term opportunities even if they hurt Google in the short term. Patents are an example here; Google could sue all of its "competitors" out of existence today, but would that really benefit Google in the long term? Would you work at Google if they were the company that put every tech startup out of business with "method for computing the sum of two integer" patents? Probably not. So aggressive patent lawsuits are not what Google's shareholders want. And doing what the shareholders want is not exactly "business cluelessness" :)

[1] http://www.buec.udel.edu/pollacks/Acct351/handouts/SEC%20For...

Remember Communicating with Care?

You might want to go review that. Or just stop posting so much about Google altogether.

So you're saying it's careless to respond to comments on the Internet?

I'm saying that speculating about what Google can and cannot do to competitors in public is definitely not Communicating with Care, and that your comment history is littered with borderline cases like this one. Just stop.

Further, you don't actually need to respond to every stupid, incorrect statement on the Internet that's written about your new employer. Really.

I see where you're coming from. Why do you have to use a misspelled throwaway account to say this, though?

(Ultimately, I see the risk in saying "I think XXX" if Google decides to someday do the opposite of XXX. But there is some limit; I'm still going to say that writing unit tests is good even if Google's official message to investors becomes "testing is a waste of time". That's because I'm me, and I have my own opinions. If Google wants to make an official statement, they have a blog for that. If someone wants to misconstrue a personal opinion of mine as an official statement, well... let's just say that their Pulitzer hopes may be dashed...)

If they're borderline report him and let whomever take care of it. Best case scenario here is that you're not only a hypocrite but one who feels the need to say things he considers so bombastic it requires hiding behind a throwaway account.

By the way, his email address is in his profile. If you were really that concerned about "communicating with care" you could be a professional about it and mention it privately to him. Of course, that wouldn't satisfy your need to respond to every stupid, incorrect statement he makes about his new employer, really.

>Another example of business cluelessness, they're not flexing their patent muscles.

A business with a basic sense of decency... imagine that.

Android was an acquisition....

Yes, 7 years ago, before they were well known. It seems a bit disingenuous to not give Google credit for Android as a success. They didn't initially create it, but they did shape it into a success.

Mountain View is the second most expensive city in the country? I wonder where LA and Chicago rank...

Google Street Map cars, then.

You want to pick nits over Google being able to ship?

That's not a consumer product, dude.

Driverless cars. Google Glasses.

The CEO's job is to allocate resources to initiatives that will make the company money. NOT to fool around with worthless projects like driver less cars and glasses that are just complete waste of money and resources.

Shouldn't they be spending time and money fixing Android and the Android market place?

The car business is fantastically large. If Google can license its technology to power the next stage of personal transportation, the revenue received will dwarf its 30% vig on selling Angry Birds.

Either way, it's not that they are taking money and resources from Android and dedicating them to cars and augmented reality. I don't think more money and more people is what Android needs.

If they were man enough to actually license. I recently say sebastian thrun speaking and he turns investor off. Hand a professor a cheque and he'll hand you a pipe dream. Google needs to be hiring more wall street sharks.

There are plenty of Wall Street types at Google, the last thing they need is to have them in control of innovation. The scientists won't be in charge of sales or licensing, I think it's great that the MBAs aren't in charge of product development.

Maybe Google won't license it, maybe Google will give it away like they do Android. Since I don't see them manufacturing cars, those are about the only options other than canceling the effort.

Can't tell if you're joking.

The only joke here is Google.

You are correct that "The CEO's job is to allocate resources to initiatives that will make the company money" and with driverless cars and google glasses the CEO is doing exactly that.

Also people working on these projects are scientists at googleX they are completely different from the Android team ... Google has enough resources (people and money) to pursue these long term research project without hurting their short term business too much.

> You are correct that "The CEO's job is to allocate resources to initiatives that will make the company money"

Actually that's not strictly true. The CEO's job is to satisfy the shareholders/board of directors. Normally that coincides with making the company money, but not always - the idea of a legal duty to maximise share value is a widely held myth. In the case of Google, most of the shares are still held by the founders. If they just want the company to do cool and impractical things they have the right and the capacity to make it do so.

The effect of a driverless car isn't necessarily that it drives itself. It's a car that's aware of its surroundings and provides feedback to the driver. Mercedes already implemented collision avoidance, driverless car can potentially prevent running over curbs, pedestrians or provide better control. In any case, lack of innovation is a pretty easy way eventually lose ground.

A lot of people would argue that a CEO's job is to innovate and that profit is just a means. Starting a company (and making yourself CEO) doesn't mean you have to change your priorities from improving the world to increasing your annual return. Lots of world-changing technology has financial gains that aren't realized for years. As long as you have enough money in the short term, which Google does, you can afford such R&D. I can't imagine the inventions at Bell Labs were profitable in the short term, but we wouldn't be typing posts on HN if they didn't exist. Do you think expensive technological innovation should be relegated to the government? Who do you trust to get us to Mars first, NASA or companies like SpaceX?

Thanks to a number of wars and billions being pumped into the UAV market, the likes of Raytheon, Thales Optronics, FLIR Systems, ITT, BAE, IAI, and probably a hundred other companies have made self driving cars possible. Not Google.

Even when I was working on autonomous vehicles 10+ years ago, the entire setup was basically a "solved problem" minus environmental sensing accurate and reliable enough to be put in the critical path of human life. The state machine for how to behave on paved roads has existed in various forms for a long time.

They have however put money and legal weight behind the problem. If Google deserves credit for anything, it would be getting them allowed on public streets.

UAV control has basically nothing to do with self-driving cars. It's a completely different problem. The only piece that's shared is the GPS/IMU, and it's a small piece. And if you think "the state machine for how to behave on paved roads" is the hard part, you're delusional.

"Environmental sensing accurate and reliable enough" isn't the problem either. Although fancy lasers and radars can help, we have had cameras, microphones, and accelerometers superior to human eyes, ears, and vestibular system for years now. The missing piece is the software, which is exactly the piece Google is working on.

Sorry for the confusing use of the term. UAV means unmanned autonomous vehicles. UAV in reference to aerial platforms is mostly the result of marketing/media.

Even the early Daimler-Benz projects of the 90's had fully functional software stacks that had implemented all the required decision making logic of driving.

While it is true that we have cameras and microphones that work better than our own eyes and ears, I don't think anyone would dispute the fact that the sensory processing isn't to human levels yet. The advancements that have been made are in different types of sensing devices that eliminate the need for heavy post-processing of data (image pattern recognition, etc) in software.

I do give credit to Google for making it happen, both as an integrator, and in lobbying efforts. But there is a lot of standing on the shoulders of giants, in comparison to pure software fields where they have made huge industry disrupting shifts (like BigTable, GFS, etc).

I don't think anyone would dispute the fact that the sensory processing isn't to human levels yet

And that's the exact problem Google is working on, and where all previous efforts have failed miserably, no matter how good their "rules of the road" state machines were in theory. Saying Daimler-Benz solved autonomous driving years ago is like saying SHRDLU solved natural language processing.

The advancements that have been made are in different types of sensing devices that eliminate the need for heavy post-processing of data

There are exactly two new kinds of sensing devices used on the Google cars AFAIK: small automotive radars and the Velodyne spinning LIDAR. The latter owes its existence not to giant defense contractors or car manufacturers but to a small loudspeaker company, of all things. Furthermore, it absolutely does not eliminate the need for heavy post-processing of data.

I think we could start putting self driving cars on the road today that would lower both the number of fatality's and accidents. We don't need more processing power, better sensors, or lower costs. What we need is slightly better software and the willingness to put it into production.

After-all the 'worst case' in a car is basically solved 99.9% of the time by staying in the correct lane, obeying stop lights / signs, speed limits, and simply hitting the breaks if your going to hit something. Sure, you could improve on that, but get that to work reliably and your already doing better than human drivers, who get distracted, drunk, tired, impatient, angry, and just plain overwhelmed.

Well, I guess "autonomous vehicles" were a solved problem the same way "portable music devices" or "smart phones" or "Tablet Computers" were a solved problem. But it takes _someone_ to push the concept over and move the category forward.

In all three categories (particularly portable music devices) - I would argue that Apple didn't so much as make the technology possible, as it did make them _popular_.

Likewise, with a autonomous vehicle - it isn't so much building a car that can drive itself - it's making a car that can drive itself _that I can buy_ that I get excited about.

Of course, Google has yet to deliver on the part about me buying them - but can you identify any other company that has done so much to get them driving on roads _with other people_?

Could your "solved problem" vehicles reliably tell the difference between a kid running into the street and a empty garbage bag blown in the wind?

If you designed the system, would you handle them any differently?

A buddy of mine, when he was a kid, used to tie toilet paper across the road at windshield height then hide in the bushes. A car would come along at ~30 mph, see it, then come to a screeching halt. Some would get out and tear it down, some would stomp the gas. Point is, every human stopped for a single strand of toilet paper blocking the road.

Perhaps on a 30 mph residential street in light traffic.

Screeching halts have a different effect at higher speeds in denser traffic. People would be less likely to slam on the brakes for a trash bag on the freeway.

But either way it is a dangerous, controversial, split second, human risk calculation. I can't say how I would handle it in advance, much less how a robot car should be programmed to handle ambiguous inputs at highway speed.

I would posit that the secret sauce is the integration of Google maps and street views. The driverless car isnt just reacting to its sensor inputs, it has a pretty good idea of what is coming up including potential problems.

"The difference between theory and practice is that, in theory, there is no difference but in practice, there is" -- numerous possible sources

I wonder if Google will end up going the same route as Philips.

I'm from the home-town of Philips, the company that became big selling light-bulbs but then expanded into a myriad of great technical innovations.

We were always joking however about how they were able to make the greatest products, but were never able to sell them.

There are so many examples of absolutely great products/innovations which were eventually canned or, if they weren't IP protected, were successfully released to market several years later by a competitor.

I really hope Google doesn't fall into this trap, because the even the greatest still need to be sold...

>If wearable computers catch on that make it easier to check-in to the places you go and +1 the stuff you like, they'll make tons more money from useful targeted ads than units of hardware shipped, just as they do now.

I expect that they have patented "location based XYZ implemented by wearable computing devices" up the wazoo before this announcement. Even if Google Glass doesn't become the market leader, they're set to reap its rewards in more ways than one.

Reply on news.ycombinator.com