Computers get more and more creative

Stephen Thaler's Creativity Machine

From the article:Also:And:The Web site of Imagination Engines, Inc. is quite straightforward in its portrayal of where we are headed:See Robotic Nation for details on where this is taking us.


Machine vision gets more and more capable

Smart Software Gives Surveillance Eyes a ‘Brain’

This article discusses software that makes video surveilliance cameras far more useful. From the article:Also:Each day machine vision gets a little more capable as new ideas emerge and high-speed computer hardware gets less and less expensive. In 10 to 20 years, computers will be able to "see" as well as human beings in many common tasks.



Robots replacing pilots

'I'm HAL; I'll Be Your Pilot'

From the article:The article has quotes from both sides of the aisle. However, Unmanned Aerial Vehicles (UAVs) are already flying routinely for the military -- It is only a matter of time before pilot-free commercial aircraft take to the skies. Pilots will succumb to the same forces that will remove truck drivers from trucks:The most telling quote in the article is this one from Boeing:In Robotic Nation I predict that pilots will be out of their jobs by 2015. In Manna I propose one possible scenario that speeds up the process:See Manna for details.


Using the cheapest labor

Tech execs look to expand--outside U.S.

This is not specifically a robotic article, but it does show how businesses behave when they have a cheaper source of labor available. From the article:In other words, as large U.S. technology companies grow in the future, they will not be adding any new employees in the U.S. All the jobs will be going overseas.

The same thing will happen as each new type of robotic technology starts to become available in the future. Companies will rapidly shed human workers in favor of much less expensive robotic labor.

The combination of offshore job loss and robotic job loss at approximately the same time means that unemployment rolls will saturate very quickly. We are seeing the leading edge of this trend in 2003/2004, which is why the job slump has lasted so long. See Robotic Nation for details.


How bad can it get?

How bad is the competition between automation and human workers getting? There was an article in the local paper recently entitled Garbage workers sound off that might give a hint as to where we are heading.

The city of Raleigh, NC is getting ready to buy a fleet of new, automated garbage trucks. In the process, 122 garbage collectors will lose their jobs. The city says it will give the displaced workers other jobs, but we all know what that means. That is what every employer tells workers when they install automation. Auto workers, for example, have heard that line for decades. As soon as a recession comes along, all of the workers replaced by automation -- the ones who were promised "other jobs" -- get laid off.

The article contains a wide variety of comments from the workers. Some of them make sense, some don't. That's really not the point. The point is, these guys are garbage collectors. They work, "for wages that force guys with families to take second jobs. Many of the guys cannot afford to live in Raleigh proper." So these garbage collectors are not getting paid tons of money. Yet, even though they are literally wading through garbage every working day and are not making much money, they took the time to call the reporter and talk with her about their jobs.

Why would they do that? Perhaps because they are worried. They are worried that if they lose their current jobs -- wading through garbage making not much money -- they may actually end up with even worse jobs once they are fired. It is hard to imagine jobs any worse, yet these guys have no trouble imagining it. Because of their fears, they are trying to save their current jobs. They don't want to be replaced by automated garbage trucks.

That tells you something about today's economy, today's job market, and the effects that automation is having on the employment landscape.

The article From programming to delivering pizza offers another perspective on the same problem. Understanding the severity of the current labor slump is also pertinent.



Robotic Drivers

Many articles have come out this month discussing the robotic vehicles that will be competing in DARPA's Grand Challenge race in March:

According to the Wired article:These vehicles competing in the Grand Challenge are true robots. They are totally autonomous. They do all their own thinking and make their own decisions. According to this article:The article continues:Most importantly, robot drivers mean that truck drivers will no longer be needed and they can all be laid off.

In the United States, there are over 1.5 million truck drivers operating tractor trailer rigs and heavy trucks (like dump trucks). There are nearly 3 million truck drivers total [ref]. Once the technology exists, all of these drivers will be out of a job very quickly for two reasons: 1) truck drivers are expensive, and 2) human error leads to a lot of accidents. In addition, a robotic truck can run 24 hours a day -- a robot never sleeps.

The same thing will happen to taxi drivers, and another 176,000 people will be out of work [ref].

If it takes the economy approximately 5 years to absorb 4,800 factory layoffs in North Carolina, how long will it take the economy to absorb two or three million unemployed truck drivers and taxi drivers? Unfortunately, at exactly the same time, millions of other workers will be getting replaced by robots as well. See Robotic Nation for details.

This list contains most of the teams that are competing:It's fascinating to look at some of the technology they are using.

As with everything else robotic, the key feature is incremental improvement. Even if none of these teams reach the goal this year, they will have a year to improve and then they will be back at it next year. And then the next year. And so on. Just as chess computers eventually beat the best human chess players, robots will eventually beat the best human truck drivers. One day in the not-too-distant future we will have robotic drivers that are much safer than human drivers, much more reliable than human drivers, much less expensive than human drivers, and 24x7.



Robots and Receptionists

University Unveils Robot Receptionist

From the article:This first version of Valerie is rough around the edges. However, it is a start, and engineering refinement will make her better and better. Think about how many businesses have receptionists sitting at the front desk. Imagine, over the next 5 years or so, Valerie improving dramatically as voice recognition capabilities improve, software improves, etc. Now imagine most of those human receptionists being replaced by inexpensive robots.

That sounds far-fetched today because we've never seen it before. However, a receptionist has a fairly limited repertoire of necessary skills. The receptionist needs to:Over the course of several years, a programmer can use standard technology to handle all of these different situations (and variations) to create a very effective robotic receptionist. With facial recognition software and a little tweaking, the robotic receptionist will likely be better than most human receptionists. Using something like Microsoft Speech Server, the robotic receptionist can also answer the phone.

Let's imagine that the first robotic receptionists that get deployed can handle 95% of the situations they encounter. That means that, for 19 out of 20 arriving customers, the robotic receptionist will be able to handle the situation autonomously and route the person correctly. On one out of 20 arriving customers, the robotic receptionist will say, "Hang on one moment while I call someone to assist you." Then the rate will improve to 97%. Then 99%. Then 99.8%. And so on. That ability to incrementally improve -- to get better and better over time -- is why robots will be able to take over so many jobs.

Now imagine a mobile receptionist -- one who can walk or roll around. Imagine that this receptionist knows about the location, price and in-stock status of every product in a large retail store. Imagine 20 of these robots roaming around the store to help customers. If you walk into the store looking for Guacamole dip, the robot immediately walks with you to the Guacamole dip location in the store. As we get used to interacting with these robots in our everyday lives, we will take them for granted (just like we take ATM machines for granted today), and they will seem completely normal when we interact with one sitting in the receptionist's chair at any local business.



Catalog of 5 robots

Five Robots That Will Change Your Life

A good article with great pictures:

This particular robot is designed to help firefighters.

The mention of firefighters brings up this thought. You may recall, back during 9/11, that one of the big problems was getting firefighters up to the affected floors. To climb 80 stories of steps carrying all their equipment, human firefighters took over an hour to reach the crisis.

Once autonomous robots are available, it is easy to imagine robotic firefighters that are packed into closets on every floor. When a fire is detected, these robots instantly burst from their closets to fight the blaze and rescue the people in the building. This will save a lot of lives (343 firefighters died at the World Trade Center in 2001), but it will also be the end of human firefighters as a profession. 250,000 paid firefighters in the U.S. will be out of work.



Quote of the week

From this article:It certainly is not surprising -- it is exactly what you would expect in a Robotic Nation. Someone needs to point Mr. Greenspan toward this blog.

From the same article:That's a direct result of automation and robots. And the pace is going to accelerate.



Stuntmen replaced by a PC

Attack of the Stuntbots

From the article:



Robots and Librarians

School hires robot librarian

Form the article:I expect that libraries will begin the process of extinction at some point over the next 20 to 30 years. In the meantime, robots will be able to run the libraries that remain.

This same technology is allowing the automation of warehouses, and will also move into the stockrooms of retail stores to eliminate workers there.



Robots and Marriage Counselors

Math and Marriage: A Match Made in Heaven? and Researcher uses math formulas for marriage

Elsewhere in this blog you can find posts that describe the vulnerability of fast food workers, pharmacists, teachers, truck drivers, nurses, lab assistants, umpires, actors, surgeons, factory workers, farm hands, convenience store clerks, air traffic controllers, etc. to robotic replacement. Now we can add marriage counselors to the list. From the article:This statement from the article is especially interesting:The implication is that, prior to this model, divorce prediction by human counselors was not accurate. So, not only will robots be able to replace human counselors, but they will do a better job -- they will be able to accurately predict divorce and then do something about it. The article describes the techniques a robotic counselor might use to steer a marriage down the right track:Robotic marriage counselors will revolutionize the industry. They will be able to predict divorce with 90% accuracy and offer useful solutions to problems in the relationship.

Humans with troubled marriages might have a problem "opening up" to a robot that looks like R2D2. Fortunately, they won't have to. We will be able to create robotic counselors that look completely human.

In Robots get friendly, the work of "Sculptor roboticist" David Hanson is described. From the article:Keep in mind that this is all happening in 2004. Imagine what robots will be like 25 years from now. Hard to imagine? Compare computer games from 1980 (Remember Pac Mac? It was introduced in 1980) to today's (e.g. Half Life 2). Pac Man had little pixelated icons moving around in a square maze. Half Life 2 is a fully immersive DVD-quality gaming experience. The same thing will happen with robots over the next 25 years. Today Valerie will be a little rough around the edges. In 25 years, she will seem completely human. And she will be able to fix your marriage.


Inexpensive Chess Computer Holds Its Own Against Grand Master

ChessBrain Community :: We did it!!!

From the article:From the ChessBrain site:Since 1950 or so (see this chronology), people have been predicting that computers would be able to beat human beings at chess. It took perhaps 40 years, but this did come to pass. In the 1990s, personal computers became fast enough that chess software could beat many human players. However, Grand Master players could easily defeat simple chess software.

It took a team of people at IBM and millions of dollars in equipment to develop a chess computer that could compete with the best human players. But in 1997, the computer named Deep Blue was able to defeat Garry Kasparov in a well-publicized match. IBM discusses the Deep Blue machine on this page:The ChessBrain project takes a completely different approach. It is using CPU time from normal computers "donated" by people around the world. The software distributes the processing load of a chess game across those thousands of computers. The advantage of this approach is that the ChessBrain project can amass a very large amount of computing power when necessary from its network of donors. The other advantage, of course, is that this distributed chess computer is free.

The ChessBrain project also shows that, perhaps in 20 years, a desktop PC costing $500 will be able to beat the best human players. In 20 years, a single desktop machine will have the power of the thousands of computers in today's ChessBrain project.

Chess is a very complicated game, but computer scientists understood how to create a chess-playing computer decades ago. Then they simply had to wait as the computer hardware got more and more and more powerful. Eventually there was enough CPU power available for chess computers to beat the best human players.

In that same way, we will be able to wait 30 to 40 years or so and we will have $500 desktop computers that have the CPU power of the human brain. It just will not be that long. Then, 20 years later, a desktop machine will have the power of 1,000 human brains. 20 years after that, a desktop machine will have the power of 1,000,000 human brains, and so on. This CPU power will fuel the robotic revolution. See Robotic Nation for details.



After robots take over a factory...

Most Pillowtex workers jobless months after layoff

Here in North Carolina last year, we lost a major employer -- a textile factory run by Pillowtex. The mill employed nearly 5,000 people. The closing of Pillowtex was the largest mass layoff in North Carolina history. Although these jobs were lost to textile competition from China rather than directly to robots and automation, the effect is exactly the same. A large block of employees lost their jobs. Therefore, this mass layoff lets us ask and answer an important question about what will happen as robots begin to take over very large blocks of jobs in the U.S. economy.

The question is simple: When an industry rapidly automates and dumps thousands of workers onto the unemployment roles in today's economy, what will happen? How long will it take for the economy to absorb the unemployed workers?

Here's what the article has to say:According to the conventional wisdom, the economy will create new jobs for all of the people unemployed by automation, robots and off-shoring. Also, our current economy is "booming" according to the White House. Yet here we find that only 400 out of 4,300 unemployed workers, or 9.3%, have found work after 7 months. At that rate, it will take 5 years or so for the economy to absorb this block of workers.

This paragraph from the article is telling:There just are not that many jobs available.

The article also indicates that the jobs these workers end up taking are worse than the jobs they lost. For example, many try (unsuccessfully) to apply at a place like Wal-Mart. Wal-Mart jobs "pay less and offer fewer benefits than jobs at the mill." The Pillowtex factory offered "good jobs". For example, one employee profiled in the article was making $500 a week (approximately $12.50/hour or $26,000 per year gross) and now receives $184 per week (approximately $9,600 per year gross) in unemployment benefits. Those benefits are obviously quite meager (well below the poverty level) and run out after a year. If it takes the economy 5 years to absorb several thousand unemployed workers, what exactly are these workers going to do after the year of benefits expire?

Now imagine what will happen as robots reach critical mass, and multiple industries are experiencing mass layoffs at approximately the same time. For example, the airlines automate and lay off tens of thousands of pilots and air traffic controllers. Simultaneously the construction industry starts to automate and lays off hundreds of thousands of roofers, painters, brick layers and carpenters. Simultaneously the retail industry begins the mass installation of automated checkout lines, kiosks and stocking robots, laying off millions of employees. Simultaneously the fast food industry starts introducing completely robotic restaurants, laying off millions more. And so on. This will all start at approximately the same time, beginning in 2015 or 2020.

As those millions of displaced employees flow into unemployment offices, they will face the same situation these Pillowtex workers face, except there will be 10 million of them instead of 5,000. As their benefits run out, will they end up in Terrafoam?

See Robotic Nation for details.


Robot boosts chances of selecting winning stocks

The future of intelligent robots



Robotic Ships

Radical Warship Takes Shape

This article from the Washington Post describes the next generation of stealth destroyers that the U.S. Navy plans to build. The project is called DD(X). The ships will be 600 feet long (100 feet longer than current destroyers) and travel at 30 knots.

From the article:In other words, the staffing needs of the new destroyer are reduced by more than half because of robots.

That same level of staff reduction will be happening simultaneously throughout the U.S. economy over the next 15 years. From the Robotic Freedom article:We need to start thinking about the implications of this transformation now. See Robotic Nation for details.

Robotic kiosks proliferate
IHT: Self-service speeds check-in
NYT: Speeding Flight Check-In at Self-Service Kiosks

From the article:The phrase, "saved the industry millions of dollars in labor costs" is a euphemism for firing people.

Also:This example demonstrates why every corporation will be acquiring robots as quickly as possible -- the economics of robots are impossible to ignore.

Automation that reduces costs like this would normally be seen as good, except that the rate at which these kiosks will start to unemploy people will be startling. From the article:You will soon find these kiosks not only in airports, but in hotels, fast food restaurants, etc. Hundreds of thousands of people will be losing their jobs at approximately the same time. Simultaneously, automated check out lines will be dismissing workers from retail establishments. At the same time, robots and off-shoring will be eliminating the remaining factory jobs. At the same time, software and offshoring will eliminate most call-center jobs. At the same time off-shoring will be eliminating the majority of IT jobs in the U.S. At the same time...

See Robotic Nation for details.

Extreme monitoring
Mood Ring Measured in Megahertz

From the article:This seems like it would be a logical extension of Manna in white-collar environments. In minimum-wage environments, it would be a way for Manna to infer that the employee might be performing in a sub-par way or is doing something wrong. On airplane flights it might be a way to detect that someone is thinking evil thoughts.

There is a quote later in the article: "I honestly can't see corporations daring to use this monitoring system on their employees. People would not accept this -- it's just plain spooky." This quote neglects the "power of situation". For example, if you are told, "you will lose your job unless you let us wire you up," or "you cannot fly on airplanes unless you let us wire you up", then you will comply.

How many people in 1980 would "accept" the security measures seen in airports today, where you are required to take off your shoes and belt, you are told to unbutton your pants during personal screening, etc.? We accept it today because it happened incrementally and we have no choice -- no one can fly unless he/she submits to this level of intrusion.

Or think about what is happening in Utah. According to
Dossier program alarms Utahns:This is happening now, and the citizens of Utah appear to have little or no control over it.

As robots and automation control more and more human systems, it becomes easier to implement these types of systems. This section from Manna describes one logical conclusion of the process:See Manna for details.

Concentration of Wealth
Reich’s Reprimand

Robert Reich's view of the concentration of wealth in America today:Also:As discussed in Robotic Nation and Robotic Freedom, robots fuel the concentration of wealth. The trends we are seeing today are the tip of the iceberg unless there are fundamental economic changes to protect everyone who is automated out of their jobs in the short term.

See this page for a wide variety of concentration-of-wealth examples.


Robotic Prediction
The Robots Are Here

Rodney Brooks has weighed in with his predictions for our robotic future. From the article:Also:Rodney A. Brooks is Director of the MIT Computer Science and Artificial Intelligence Laboratory, and is the Fujitsu Professor of Computer Science. He is also Chairman and Chief Technical Officer of iRobot Corp.

For details, see Robotic Nation.

Robotic Military Vehicles
Robots for No Man's Land

From the article:Also:

On the one hand, it makes sense to remove people from the battlefield and replace them with robots carrying machine guns and launching bombs. A battlefield is a deadly and disgusting place -- a place where millions of lives have been lost in the most atrocious ways.

On the other hand, one purpose of war is to murder human beings. Do we want intelligent machines in charge of mass murder?

If machines are given the right to murder humans, here is one scenario. Imagine that an intelligent, autonomous robotic soldier circa 2040 costs $50,000. In 2040, the U.S. invests $200 billion in robotic soldiers (about one half of today's defense budget). $200 billion would buy about 4 million troops. In 2041 the U.S. does the same thing, and buys 4 million more. In 10 years, the U.S. would own 40 million of these robotic soldiers. That is more soldiers than the entire planet has today. They would be supported by robotic ships, robotic aircraft, robotic vehicles, etc.

This entire robotic military force would be commanded by a handful of people -- the Commander in Chief, the Secretary of Defense and a small collection of generals and admirals. That is identical to today's military command structure. The key difference is that a robotic army, unlike a human army, can easily be programmed to have no conscience.

Let's take an extreme example of where this lack of conscience could lead. It would likely be impossible for the President today to decide to use the army to kill all 35 million of the citizens in the state of California. We would like to believe that at some level in a human army, the human soldiers would rebel against the command. A robotic army would have no such reaction, and would simply carry out the order.

On the other hand, let's imagine that we do program robotic soldiers with a conscience. The robots would end up ignoring many commands as unconscionable. In other words, robots with a conscience would not make very good soldiers.

Archives © Copyright 2005 by Marshall Brain
Atom RSS

This page is powered by Blogger. Isn't yours?