Category Archives: Singularity Watch

Atomic Computing

It has been possible for some years now to use atoms to represent 0s and 1s as part of a computing scheme. Getting people to believe this is actually possible and practical seems to be hard, although this new demonstration by IBM should make it clearer just how much potential there is at this level.

And if you’re sitting there thinking “yes but is it practical?” contemplate that the first major commercially sold electronic computer used 5,200 vacuum tubes, had the equivalent of about 1K of memory, weighed 13 metric tons, consumed and performed about 1,900 operations per second, and cost millions of dollars. What is just a demonstration today can become an everyday reality in the blink of an eye in historical terms.

The exponential growth curve of technology is not going to be hitting a plateau any time soon methinks.

Unmanned fighters in near future

The US Navy has launched the first full fighter-sized robot plane from an aircraft carrier. Those of us who think growth in technology is linear rather than exponential often make the mistake of thinking this is no big deal because it hasn’t landed on a carrier yet. My prediction is they’ll have planes like this landing on carriers in fairly short time… and that by the end of this decade (which is only 6.5 years away) we’ll not only have them taking off and landing from carriers, but we’ll also be beginning to discuss publicly when and how to begin phasing out human pilots for most purposes. (Well, some people already are, I just mean that’s when the general public will become aware that it’s seriously under consideration.)

The thought of fighters, not to mention bombers, doing combat at accelerations and rolls and speeds that would kill a human in the cockpit is a little frightening–a machine can handle 20, 30, 50 Gs easily I would guess. But it also looks pretty inevitable now, and I don’t just mean “within my lifetime,” I mean within the next decade or so.

By the way, at this point I’m also calling mid-2020s (I’d say 2026 at the latest) before we begin openly contemplating making self-driving cars pretty much mandatory and openly questioning whether we should even allow people without special needs and training to drive themselves. To be clear, that’s when the discussions will begin in earnest.

Technological growth is rarely linear. That’s worth remembering.

Driverless Car Update

So now the team at Oxford University is getting into the act, and they are setting their sites on finding a way to cheaply retrofit existing cars to boot.

Audi, Volkswagen, and others are also still at work on it of course. Google is projecting 3-5 years before general availability. It still looks like my prediction that some rich person will purchase one just to have it as a toy by the end of this year may be overoptimistic–although on the flip side, that may have already happened and no one knows it yet.

This is especially exciting news for the visually impaired, although currently car companies are being cautious in what they say they’ll offer; they’re starting small, by incorporating more and more features they’re calling “driver assistance.”

I remain pretty staunch in my position that by 2020 these cars are going to be generally available.

No More Saturday Mail?

It appears the U.S. Post Office will stop delivering regular mail on weekends later this year.

I’m not surprised. I expect to see them further and further contract. Although I hate to leave the business entirely to private enterprise, I’m not sure the postal service is relevant anymore. I can think of only one or two things I use it for, and if you took away my use of Netflix, that would be something I do maybe a half dozen times a year. Mostly, what’s in my mailbox is spam, just like my email.

On the other hand, these days I regularly talk to friends on three different continents and in at least four different countries, daily, for free. I like living in the future, don’t you? :-)

Driverless Car Liability Non-Issue

First I was told driverless cars weren’t technically feasible in our lifetimes. Then I was told they would be too expensive to be practical. Then I was told it would be too regulated by government for anyone to make a profit. Now that every one of those hurdles was easily overcome or turned out to be groundless, the normally sensible Megan McArdle says driverless cars will be stopped by lawsuit liability, and then goes further and suggests the “fix” for that will be to require the driverless car to have someone behind the wheel, which will render the technology moot.

I call bullshit. I and others have heard these lawsuit liability arguments before and the answer is always the same: You don’t need the driver to be in the car. The solution is a regulatory environment (gasp! yes! we actually need those!) which makes the responsible party the owner of the vehicle, just like the owner of a horse or a dog or whatever is responsible for the actions of that animal whether the owner is present when the animal behaves inappropriately or not, or just like you’re responsible now if you forget your baby in a hot car and the baby suffocates. In this scenario the owner of the vehicle might then turn around and sue the automaker of course, but that’s happened in other industries and usually sorts itself out. As long as the direct liability follows the owner of the vehicle, or whoever last pressed the “go” button on it to send it off somewhere on the mission where the accident happened, the issue will sort itself. I predict this will be as much of an issue as all the other objections raised to it to date, i.e. a challenge not a massive insurmountable roadblock.

All it takes is that inconceivable thing that apparently libertarianoid/rightists can no longer imagine existing anywhere: a government policy that makes sense. One of the reasons I’ve become so skeptical of so many libertarian and/or right-ish economic thinkers is what I have begun to think of as the libertarian, “government can never, ever do anything right, ever, except maybe blow things up, and will always cause disaster no matter what it does” mentality. It’s become less a thoughtful position than a twitch in some corners, and I say that with all due respect to McArdle, who’s usually pretty damned sensible.

Yes, liberal/leftists have their own incredible array of petty hypocrisies, stupidities, and blind spots too. It’s called “being human” folks, and no matter where you sit on the political spectrum you’re going to have those.

Anyway, on the self-driving cars: First it was “oh they’ll never work.” Now that they indisputably do work, it was “the government won’t allow them on the road,” but now multiple states already do. Then it was “they’ll cost way too much money to make” but it turns out it won’t. Then it was “no one will trust them,” but tons of people are saying they’d love one, and multiple car companies, having researched the market, are promising to produce them–not just one company taking a risk, but the entire industry in multiple countries has researched it and believes consumers want them. The only remaining argument that seems to have any legs at all is “the liability costs will be sky high,” but the simple fix is, you make the operator responsible for the vehicle just like you do now, but define “operator” as “whoever turned it on and sent it off on the mission where it got in trouble.”

The only hitch will be if something goes drastically wrong and there’s a huge series of calamities, all at once, due to a design flaw. Which all the big car companies have gone through and managed to survive. Otherwise there’ll be a few settlements, most likely, and progress will continue, because these things will be safer than human drivers by far.

Driverless cars are coming. Period. As I said a few years ago, the only really interesting question left is when they start ratcheting up the requirements for human drivers to make it tougher to get a license, and then when they will start moving to outlaw human drivers entirely because it’s too unsafe and irresponsible to allow humans to drive on public roads anymore. I give that maybe 20 years before people begin thinking seriously about it.

Robots Genuinely Taking Jobs From People?

For years and years I said it doesn’t matter when jobs get replaced by automation, it didn’t matter, because it freed people up to do other more creative and interesting things. And I do believe that’s been a general trend for generations. That said, I increasingly find myself wondering how much longer it will be before there are essentially almost no jobs people are needed for.

What will “wealth” and “capital” mean in a society when most of the population has no clear purpose? I’m just not sure. But I’m not the only one who’s wondering, so do professors at places like MIT.

I’m no longer so blithe about assuming this is a self-solving problem. I think one problem futurists have always faced is predicting future events too soon (i.e. before they actually happen, even though they eventually do) or predicting things will happen a lot later than they do. But I suspect we’re in a decade where we’re growing to a point where an ever-increasing and noticeable percentage of the population actually doensn’t have anything they’re needed for, strictly speaking. And I wonder what happens then?

The High Price of Materialism

The funny thing is, there was a time when there would have been nothing at all controversial about this video. I remember, I was there. Yet now, somehow, today it seems like it is controversial:

There is not a word of that I disagree with. I also think it’s describing the issues we will wrestle harder and harder with in the emerging 21st century economy, as everything our way of life is dependent on is changing so rapidly no one can keep up. The better we have the lower two tiers of Maslow’s Hierarchy of Needs covered, the faster we seem to forget or are frustrated in our desire to climb to the higher layers.

But I do know one thing: the least happy people I know are often the most driven to “succeed” and define “success” as acquisition of power and status and money (which, really, are all part of the same thing, and are in my view much like a narcotic: pleasant, positive, potentially useful, and seductively poisonous). Living a life where you consider “competition” the ultimate in nobility is extremely likely to end in misery, like the Silverback Gorilla Male who takes over his harem and then just waits until he’s old or injured so he can be killed and have his place taken. It’s not a happy life.

And I hope it’s not the type of life we’re evolving toward as a species.

(Thanks Paul.)

Self-Driving Car Update

One of my New Year’s predictions for 2012 was that the first self-driving car would go on sale this year. Now, what I meant by that was that someone somewhere would sell one, in the same way that you can buy an Asimo robot right now if you’ve got 7 figures handy. We’ll see if that comes true, but it’s starting to look like the growth in self-driving cars will be faster than I expected even on the consumer end; BMW plans on having semi-autonomous cars available for sale to the general public next year, and General Motors is promising same by the middle of the decade. GM is also planning to have the fully autonomous car available for sale by 2020–Which pretty much matches with what my longer-term prediction has been for a while now, that such things will be available before the decade is out, because if GM plans on having it by 2020 you can expect others will be to market sooner.

At this point the self-driving car is a solved problem from a technical standpoint. No special modifications to the roadways required. In fact, it’s so well-solved they’re now turning to much more mundane questions like how robot cars will communicate to pedestrians.

Technologically, it’s a done deal. It’s a matter of working out the legalities (which are nowhere near as hard as some people seem to think), as well as cost and reliability factors. Just as with all other proven and working technologies, it will get cheaper and more reliable all the time. First, rich people. Then, upper middle class people. Then, middle class. Then, lower-middle-class. Then, everybody. To me, the only remaining question is how long before we start ratcheting up requirements for humans to even be allowed to drive instead of robots? 2030-ish perhaps?

Self-Driving Cars Legalized In Nevada, Other States To Follow

The future has arrived in Nevada, where they have passed a law to license robotic cars. Florida, Hawaii, and Oklahoma already have plans to follow suit. Story here. With a neat video:

Up until now people testing self-driving cars have been doing so in the “ask forgiveness instead of permission” mode, which isn’t necessarily all that bad legally:

A basic concept of US law is that most of the time, pretty much anything is legal unless there’s a law against it. With no laws against robot vehicles, the only real question with one of these things is who will be liable if when one of them is involved in an accident. If the law said nothing specific about it, then a judge would get to decide on his own who would be liable, and that would then be the law until some legislature or higher court ruled different. My guess is that a court would wind up deciding that whoever had ultimate control of the vehicle–i.e. whoever was at the keyboard of the computer that told the car to go–would be considered the “driver” for liability and criminal purposes, and that would be that. The main problem being we can only guess. But otherwise, once a judge decided who was liable, the law would just go from there, almost certainly treating some specific party as having the same liability as if they were actually driving it; Google wasn’t taking all that big a risk.

People also tend to assume the financial liability would be astronomical if someone got hurt, but the reality is–sad or not, depending on your viewpoint–that we know pretty close to the dollar how much it costs in monetary terms if you kill or injure someone. Even if Google couldn’t obtain special insurance, a few tens of millions of dollars in a special account would take care of most liability concerns on any such experimental vehicle, there being only so much damage one car can do. But likely it could be insured anyway, because companies like Lloyds of London will insure pretty much any weird thing you want so long as you can pay for it. If you can insure your own private launch into space (and you certainly can) you can insure a self-driving compact car.

All this legal status indicates is that life will (most likely) be easier on judges, and the terms of liability will be clearer to everybody. It’s really less a remarkable legal development than a sign that this technology is coming sooner than people think. Although Google says selling these things is some years away, they aren’t the only people working on such a project, and it would not surprise me a bit to see at least one self-driving car sold by, and to, somebody who wants to make a headline. Which multimillionare wants to step forward to say he bought the very first legally-licensed driverless car in Nevada and, hence, the world? Hey, if I were super rich I’d be vying for that spot right now, even if I had no use at all for the thing. If I were a multimillionaire, I’d approach Google now and ask if I could buy one as soon as it was legal, and tell them they could keep it and do anything they wanted to it to keep testing so long as it was registered in my name. ;-)