This guy knows his shit. I'm looking into and researching self-driving cars ever since this post. The technology and implications are fascinating, to say the least.
PSA: veen is a bright and inquisitive young man with a penchant for maps, GIS and technology. Driverless cars, by way of comparison, are poorly understood, murkily-reported and probably the most important/disruptive technological advancement we're likely to see in our lifetimes. ________________________________ As such, it behooves you all to follow technology.veen and nag this cat to keep us up on stories like this, because veen gets it. ________________________________ I've told him before that this stuff should be his thesis, and that there's a lifetime of journalism/research/invention/punditry/whatever he wants out of this. If he had a kickstarter, I'd donate, and the only other kickstarter I've given money to is Robot Turtles. Also, you oughtta check out that linked post again. That is all. (Or not: the author's ebook is 99 cents for 2 more days, then jumps to $9.99. I bought it. Have not read it.)
Thanks, kb! I'll do my best to fulfill your expectation. :) Part of me still fears it ends up like a Segway. I've read his 7-part series on self-driving cars he wrote last year (first link of the article). While he uses way too many words to make his argument, he's got a lot of interesting points, and he connects this technology with the arguments in his book. Bought it, I'll try and read it this week. Maybe the gist of the book is woven into the 7-parter, but for only a dollar I'd be happy to find that out.murkily-reported and probably the most important/disruptive technological advancement we're likely to see in our lifetimes.
So Dean Kamen and the Segway are kind of a sore point with me. The old-timers among us may or may not remember the crazy speculation around "ginger", which was going to change the world if only we knew how. Then we ended up with a Razor scooter that needed a Celeron in order to roll. You look back over the course of history and there has never once been a demand for a sideways scooter. It's just not a thing. You look back into the murky pool of the future and you can't see a demand for a sideways scooter. It turns "walking" into something requiring an operating system. And honestly? Speaking as someone who worked in the industry where Dean Kamen's "revolutionary" dosing pumps supposedly overturned paradigms and shit, he basically slapped a stepper motor on a syringe. I fucking built one of those in the lab because I needed one and it was obvious - C-clamp plus syringe plus motor equals dosing pump. There's a lot more hype there than innovation. Driverless cars are another matter entirely. If there were no purpose for them there would be no car service. There would be no taxis. There would be no chauffeurs. There would be no stagecoach drivers. There would be no rickshaw pullers. There would be no litter-bearers. "let someone else drive" is a human need that predates the wheel and a driverless car, if nothing else, automates a job done at some point by everyone with a driver's license. I've said it before and I'll say it again: I'm a driving enthusiast. I enjoy driving. I look forward to a time when everyone else's car is driving them because it's gonna be a lot easier for me. And hey - there are plenty of times when I wish I could multitask and do things other than drive. Have that option available to North Americans regularly and cheaply? It's gonna be a game-changer on every level. I differ with the conventional wisdom on when it'll happen, how it'll happen and what the biggest effects will be but once we hit critical mass it's gonna be a different culture.Part of me still fears it ends up like a Segway.
Seem to remember very earnest speculation as to whether he'd found a way to teleport matter. Jesus, somebody shoulda punched that guy.The old-timers among us may or may not remember the crazy speculation around "ginger", which was going to change the world if only we knew how. Then we ended up with a Razor scooter that needed a Celeron in order to roll.
I wish I could invest specifically in this subset of Google. Shoot, I don't know why Google isn't throwing even more money at this project to accelerate it's implementation, it has the potential to generate mad revenue for them. I know there is considerable risk, but from a risk vs. reward standpoint... well, the reward is quite a jewel. Thanks for this article, veen, great post.
Who says they haven't already? They bought out most of the DARPA Grand Challenge winners, numerous robotics and AI startups, got some really good professors from Stanford and MIT on the team... And that's just the part of the project we happen to know about. The self-driving car is what birthed the Google X wing, their experimental secret-ish department that also brought us Google Glass, and possibly the very interesting Project Tango - which is basically the same tech as the Google car but for inside your home. Let that sink in for a minute.Shoot, I don't know why Google isn't throwing even more money at this project to accelerate it's implementation
There's a great video from Pieter Abbeel, professor at Berkeley, where they show a robot doing all types of household tasks. Accompanying it, he notes that everything shown in the video is controlled by a human. The tech is all there (costs, non-withstanding...), they just need a brain to direct the motors. linky Now, that "just" turns out to be pretty complex, even if you hard-wire the highest level of logic "Do dishes" => "Collect dishes" => "find dishes" ; "grab dishes", you still have many black-boxes in there that you all need to perform with a high degree of accuracy. You don't want to put the TV remote in the dishwasher, and determining how to pick up a dish--even what orientation a dish is in--is not a straightforward problem. Then you start wondering if it is robust against the cat walking in front of a sensor... Each of these is generally tackle-able, but I can't help but wonder if version 1.0 will only support circular plates with only minimal scraps of food. Still even if your robots don't have any sense of objects or time, and are only grabbing and moving items with pre-assigned tracking tags, that's enough to make entire jobs obsolete.
Contextualization of machine vision has been the bugbear of artificial intelligence for 20+ years. The advances we've seen over the past 10 or so have all been a function of the steady advances in computing power and speed prompted by the Internet and smart phones. We're past the need for a quantum leap; at this point it's the incremental crawl to asymptotic perfection. Which is kind of a drag. There will be no "holy shit" moment. It'll be "huh. I just bought an Asimo that does windows at Best Buy - future, yo" moment.
Yeah, it's sort of funny how people gave up on "programming consciousness" and just figured out that if they can abstract away actions, they can script the highest level and get a robot / program that is still perfectly useful. Makes me think of machine learning as "the working-man's AI".
This guy might know his shit, but he misses a GIANT point when he says: Look at the picture of the "Google" car. It's not a Google car. It's a Prius which Google has retrofitted with its technology. That's a non-trivial difference. If anything, this technology should strengthen car manufacturers when all is said and done. It's easy to get Foxxcon to manufacture your new smart phone to unseat BB. It's no easy task to build a car. The car process starts in an art studio and continues its journey for sometimes up to a decade to make it to production. It's a complicated, global logistics nightmare that can't be pulled off by a technology company, no matter how big and rich (for example, just the engine of your car probably has components from at least three continents and parts form a dozen different [specialized] companies). Unless Google is planning on buying one of the major players, they will never be more than a supplier, and it will be in everyone's interest to cooperate. Sure, as a supplier Google is going to command a lot more than some company that raw casts engine blocks, or molds plastic for your bumper, but they will still need manufacturers as much as they're needed in turn. And anyway all this rests upon people not wanting to be autonomous, which is still far from clear. Google is betting that people don't want autonomy, but the automakers' market research suggests otherwise. Market research can, obviously, be incorrect, because consumers don't always know what they want until they have it. But for now, all indications are that people like driving, and want the option to drive when they feel like it. There's a reason that GM has four brands, and each brand has a bunch of offerings, and that GM is just one company among a dozen major players internationally: people like choice. It's a sad world if Google takes over, and we're all forced to drive Model T's again. Maybe I'm a luddite (and in full disclosure I'm also a person who is deeply invested in the auto industry both financially and personally), but I just don't think that monopoly is a viable direction for the industry. And I think that especially in the US, it will be a very difficult sell to get people to completely abandon autonomy.Unfortunately, disruptive innovation can upend even the most well-conceived incremental strategies. Just look at Blackberry, Blockbuster, Borders, DEC, IBM, Kodak, Microsoft, Motorola, Nokia, Sears, Yahoo and most of the TV, newspaper and publishing industries—to name just a few bankrupt or struggling former industry leaders. They all bet that non-incremental approaches would not pan out, and struggled with the consequences when those bets went sour.
Very valid point. I think you might find this video interesting: (I'd love to know what your wife thinks of the design and / or their design process. How capable is Google at designing cars?) The thing is, it's too early in the production process for Google to make a final decision on how to implement their system. That's what we're really talking about, a system of computers and some sensors. Making those do something useful is what Google's good at. All Google wants is to bring their system into as many cars as possible. Here's what could happen: - Google will build a car themselves, or acquire a car company. Remember, they acquired Motorola for $11 billion just to get into the smartphone game. - Google will partner with a large car company. Collaboration means that Google doesn't have to manage the car part except from some sensor placements. - Google will license their system to other car companies. Ever heard of George B. Selden? Held the patent for the automobile, became rich off royalties. Google could end up in the same luxurious position. They're far ahead of the competition. Most people (including the author of that piece), after seeing the prototype car released last month, assumed that Google would automatically go for option #1. They're making a car themselves now, surely the future Google car is from Google? I'm of the opinion that the third option is the most likely - Google having their proprietary system, selling it to GM, VW, Toyota etc as an option on regular cars, like cruise control is now. Further in the future it might be the other way around, with cars having GoogleDrive as default and steering as an option.
They're rich enough to poach designers from established companies, or to hire new grads. There are only three major schools in the US that train car designers, and most car designers come from one of them, or from one of the handfull of others in UK, France, Korea or Japan. So my guess is they could figure it out. But as to which option Google will choose, I think it most certainly won't be option one. Take total employment, as an example. Google employs about 47,000 people, if Wikipedia is to be trusted. GM alone employs 205,000 people directly, with hundreds of thousands of others working for suppliers needed to support their manufacturing operations. GM's domestic market share is 18%, with their world market share at about 11%, according to the most recent figures I could find. Do the simple math to extrapolate the total global automotive workforce. It's staggering. Unless Google is planning to increase it's workforce by two orders of magnitude (in a wholly different industry than the one in which they have expertise), they aren't really in a position to dominate anyone, are they? Any asshole with some knowledge, money, and time can make a prototype. Only a mega manufacturing company can make safe, affordable, attractive, mass production vehicles. So yeah, option two or three are the only real options, but really option three is the only logical one, unless Google really wants to get into the auto business, which I'm sure they don't (that $11 billion is chump change compared to what a major auto company is worth; even if their market cap is lower, one needs to value assets to assess a sale price, and all the factories and capital equipment of a major player adds up to a boatload of cash). (And yes, I used to hang out at a bar on Selden street in Detroit when I lived a block away--everything here is named after auto company men.)(I'd love to know what your wife thinks of the design and / or their design process. How capable is Google at designing cars?)