Hostile machinery

“The popular idea, fostered by comic strips and the cheaper forms of science fiction, that intelligent machines must be malevolent entities hostile to man, is so absurd that it is hardly worth wasting energy to refute it. Those who picture machines as active enemies are merely projecting their own aggressive[ness]. The higher the intelligence, the greater the degree of co-operativeness. If there is ever a war between men and machines, it is easy to guess who will start it.”

– Profiles of the Future, Arthur C. Clarke

"When we use automation and cybernation more extensively, not only industrial workers, but also most professionals can be replaced by machines. Even today, the most visionary writers and futurists have difficulty accepting the possibility of robots replacing surgeons, engineers, top management, airline pilots, and other professionals. Machines can easily replace humans in government and in the management of world affairs. This does not represent a take-over by machines, as some people might fear. Instead, the gradual transfer of decision-making to machine intelligence is the next phase of social evolution."

- Designing the Future, Jacque Fresco

The rise of internet nationalism

Universal Declaration of Human Rights: Article 27
1. Everyone has the right freely to participate in the cultural life of the community, to enjoy the arts and to share in scientific advancement and its benefits.

In the last few years something nasty has been emerging. I first encountered it when trying to watch a video on one of the YouTube-like sites. No doubt something in some way related to robots. Instead of the expected lo-fi stream of pixels and sound I received a message, saying something to the effect that "this content is not available in your country". This is what I've come to describe as internet nationalism. It's really a sort of censorship or prejudice, based upon the locality where people by pure accident of chance happen to have been born. Someone somewhere decided that for some contrived reason that people living within a particular administrative zone of planet Earth are not worthy of participation. They're digital persona non grata.

Though only applying in a small minority of situations this sort of blocking of information has been increasing, and if it remained confined to video that would be not much more than an annoyance. But now the phenomena of internet nationalism is beginning to hit closer to home, with the veteran software development site Sourceforge blocking developers from participation based upon their country of origin. I've been writing open source software for long enough to know that there may be many reasons why you might want to exclude someone from a project. Maybe they're lousy coders. Perhaps they're spamming the forum. Or they might just be poisonous people, out to cause mayhem. But excluding someone based on their nationality - a factor which the majority of people have absolutely no control over - just seems wrong. It would be similarly trivial if they were to be excluded based upon their skin colour, gender or sexual orientation.



The US export regulations have been around for a long time. I remember reading about them in the 1990s as part of the end user licence agreement that comes with Microsoft Windows. In the olden days, when software was something in a shrink wrapped box made by a particular company, the notion of export made some kind of sense. But open source development is quite a different animal. The software may be developed in multiple locations around the world and the location in which the code or binaries are uploaded to and downloaded from could be anywhere, often with multiple mirrored data sets. In the internet era the notion of software being "exported" across national borders seems rather irrelevant.



Since the regulations have been around for so long, and it's debatable whether they even apply to open source software, why have Sourceforge waited at least since 2003 to start enforcing them? This suggests to me that US officials are beginning to apply political pressure to software organisations, which they didn't bother to previously. [In subsequent searching it appears that the pressure has come all the way from the very top].

Apparently the same restrictions also feature on Google Code, which means that US export controls apply to some of my own projects even though I'm neither a US company nor a US citizen. By agreeing to Google Code's terms of service I'm unwittingly participating in a discriminatory policy which goes against the principle of free software. This is obviously something that I don't want to do, so I'll be looking to transferring these projects to other hosts, such as LaunchPad, GitHub or Gitorious.

Cutting twaddle emissions: towards a sustainable future

An excellent guide to whether and how Britain might be able to live sustainably without burning fossil fuels is Sustainable Energy - without the hot air by David MacKay. Perhaps most importantly for recessionistas like myself it's free to download. The book provides a grand tour of the various renewable energy options available, with the assumption that average energy consumption per person remains similar to its current value, and has plenty of quantitative analysis which makes a refreshing change from the usual gaggle of hopeful assertions. At least if you disagree with him about particular technologies, you'll have some numbers to argue over.

On current trends you would expect energy consumption per person to increase in future, but this is obviously unsustainable unless we begin to include technologies which currently remain unproven, such as nuclear fusion or space based solar. The best case scenario over the next few decades, assuming no major breakthroughs other than cheap solar, is that we can at least sustain the sort of quality of life that we have today.

It looks as if Britain and the rest of Europe are going to struggle to live sustainably in a world where fossil fuels are no longer cheap enough to burn, and may become dependent upon solar energy produced in northern Africa. If this is the case you can expect a significant amount of emigration southwards, and for north Aftrica to become far more economically vibrant than it is today, since commerce and industry will presumably want to gravitate towards where the cheap energy is. A possible prediction is that Europe may seek to accomodate north African countries within the European Union in order to try to ensure the sort of political stability which would be needed if you're going to be making large energy generation investments there.

According to MacKay, a realistic plan for generating a reasonable amount of energy for the population of Europe and north Aftrica (approximately 1 billion people) would look something like this:



Here each yellow blob is a solar power facility, where the actual light collecting area is one third of each blob.

More bogus inventions

From questionable body scanners to bomb detectors. This is a good example of where lack of critical thinking is not just a laughing matter, but also can have fatal consequences.



Anyone with even a cursory familiarity with the purveyors of pseudoscience will immediately recognise this device as a posh dowsing rod. Throughout history there have always been charlatans producing bogus inventions which they can sell at a premium to the gullible, so the main fault here is lack of proper evaluation by whatever government agency procured these devices. In this case sniffer dogs would have done a much superior job, would have saved lives, and would probably have been considerably cheaper too. It also demonstrates the usefulness of negative knowledge - knowing what isn't likely to work in addition to knowing what is. Those who lack negative knowledge are open to being easily misled, and in the business of inventing things you could say that most of the knowledge acquired is of the negative variety.

For some independent testing of the validity of dowsing see the following:



But it's not just water or bomb (or elephant) detectors that you need to be wary of. Are all of those pills which Ray Kurzweil is claimed to swallow each day really based on good medical science? Hopefully they are, but historically medicine is an area where charlatanry has traditionally made a lot of money out of people who are uninformed, desperate or who merely want to believe the confidence trickster's scientific-sounding sales pitch.

Body scanners

With body scanners being rushed into UK airports as quickly as possible, this demonstration (German language) shows just how ineffective they are at actually detecting potentially dangerous explosive materials.

3D TV

Although you might expect otherwise, I'm not a great fan of 3D TV or cinema, which seems to have been enjoying a hype bubble over the last year. The current incarnation of the technology is primitive and provides no real advance over the kinds of experience which were available in cinemas in the 1950s.



There's the initial novelty value, but after a while I expect that the irritation of the coloured or plane polarised viewers will outweigh the declining novelty factor. If the current 3D craze is not subsequently followed up by more advanced 3D viewing experiences I expect that people will soon lose interest, just as they did decades ago.

3D is just the latest attempt to revamp and reinvent conventional television broadcast services, which are facing an increasing level of competition - including competition for a finite pool of advertising revenue - from online video sites such as YouTube. This is a battle which will go on for some time, but which conventional TV broadcasters are certain to lose.

So what might make the 3D experience compelling to the point where it becomes more than just a gimmick? One possible solution is that shown by Johnny Chung Lee.



This is only limited to a single viewer, but possibly with the addition of plane polarisation you might be able to have two or three people viewing the screen at the same time. This really involves no additional cost above that of existing consumer electronics, so could be implemented very soon.

The other method, which is more advanced but which still could occur within a decade, would be to wear something like an EyeTap device, and have the 3D content projected directly into your field of view. Ultimately I think it's going to be this sort of technology which becomes the standard way of delivering multimedia content in any location, and it might not be too much longer before the required components are at a sufficiently low cost for the mass market. All that's needed is a pair of cameras, plus a gyro for head pose estimation (the same type of component used in the Wii game controllers), with a wifi link to a computer (perhaps something like a netbook or mobile phone) which can broadcast the 3D graphics to the wearer. Also in addition to entertainment, there would be many other work and social networking related uses for augmented reality. In fact, such a viewing device could replace many of the expensive monitors and phone screens that we have today, and could also be worn discreetly with a similar appearance to a regular pair of spectacles.

Marketing fear

Dan Gardner talks about how fear can distort estimations of risk, and some common cognitive biases which lead to dubious or erroneous beliefs. These ways of thinking were useful in the past, but have failed to scale well in modern societies.



Some organisations, aware of the distorting influence that fear can bring to bear even upon well educated individuals, can and do use fear as a marketing tool. Classic examples of this would be the fear, uncertainty and doubt spread by Microsoft about rival products, or the way that CCTV cameras or anti-virus software are marketed. As pointed out, often there is a significant difference between popular perception of risk, as derived from media and marketing sources, and the actual risk as judged by the best methods currently known.

On becoming war-averse

Bill Rammell, the Armed Forces Minister or what in 19th century politics would have been known as minister for war, complains that we're becoming too risk averse.

"My great fear is that we as a nation will become so risk-averse, cynical and introverted that we will find ourselves in inglorious and impotent isolation by default"

It would probably be more appropriate to describe people as skeptical rather than cynical, although there certainly is a high degree of cynicism with regard to foreign wars, which originates in large part due to the dishonest way in which politicians behaved over the Iraq war. Sadly, war remains a highly profitable enterprise for a few individuals and companies, but there are a few cases where war may be justified on a non-profit basis as a means of defending territory and resources from hostile invaders.

The trends lamented in the article are really quite positive ones. In the digital age, people want and expect there to be information about what's going on, and the ubiquity of information means its very difficult to keep anything secret for very long. Families are quite right to pursue inquiries into the deaths of soldiers, because the findings may result in improvements being made to procedures or equipment. It makes no sense that in order for government to act in our best interests they must behave with obsessive secrecy, with a few obvious exceptions such as the disclosure of military plans or the maneuverings of battle ships at sea. In most cases if information was gathered using public money then it should be available to the public (there are also a few exceptions here, such as medical records).

The trends are encouraging because their ultimate consequence is to make the prosecution of war more expensive and bothersome, and to increase public scrutiny of these matters. The more cost is associated with war, the less quick politicians will be to thoughtlessly rush into them - as happened in Iraq - or for profiteers to see them as a sure way of making a quick gain.

The notion that the Afghan war will ever be "won", in any conventional military way, shows that Mr Rammell has little grasp on military history, and the idea that troops on the ground are essential also doesn't stand up to much scrutiny. If the primary objective is to capture or kill Bin Laden and to shut down the training camps used by terrorists then this can be achieved with drone aircraft and possibly by selective use of special forces where necessary.

The hunt for AI

An awesome documentary from 2001 depicts the brain building progress (or otherwise) of Rod Brooks, Hugo de Garis and Steve Grand. There's some very astute commentary from the philosopher Dan Dennett about the ever present dangers of hype.

In some ways a lot has improved over the last nine years. There's much more compute power available, and camera technology has become more standardised. The microcontroller issues which Steve Grand has with his Lucy robot would probably no longer be relevant, especially if he used something like a Beagle board as the main computer, and his comments about building cats and dogs turned out to be somewhat naive as military telerobots with less than benign ambitions have been an active area of development within the last decade.

The whole de Garis Starlab episode is quite entertaining, and he's obviously a natural performer for the cameras. It's a heady mix of apocalypticism and technology, which all goes horribly wrong, but I assume that the failure of Starlab had more to do with the ending of the dot com bubble than with any overblown AI claims. Not much has happened since on his particular approach, which is based upon evolving neural networks with FPGA hardware, but having moved to China and received more funding it's possible that this might not be the end of his quest for a thinking machine.

Rod Brooks has enjoyed tremendous commercial success with iRobot, but hasn't done anything much of academic interest since the Cog robot. He now has a new company called Heartland Robotics. The idea of making use of naive physics to classify animate and inanimate objects would seem to be a good one, but to my knowledge hasn't received much subsequent attention.

The automatic typist and other modern business machines

A fascinating look at office automation in the mid 20th century, shows just how much the digital computer has revolutionised the way that many common business operations are performed.

Probably only larger companies would have been able to afford these specialised types of machinery, which would have been costly to acquire and maintain. The office computer and printer has now replaced these multiple machines with a single general purpose system, and also to some extent consolidated the range of skills needed by office workers. In time I hope that even printing documents on paper will be reduced to an unusual task, with most activity being entirely electronic.

A better captcha

This is quite amusing, because if it's eventually used (and I expect that it will be) it's effectively forcing the shady organisations responsible for producing spam to solve hard problems in computer vision.



They'll probably respond to this simply by ratcheting up their human computation efforts, whereby people on porn web sites may be asked to complete the CAPTCHA in order to see the next picture, or via systems like mechanical turk where users don't necessarily know what ends their cognitive labour will ultimately be put to. But at least in the turk scenario spammers would need to pay some amount of money in exchange for the cognitive labour, and even small payments may render the whole spamming exercise uneconomical - which is beneficial for the rest of us.

UK population age profile

UK population by age in 2001 and 2008, according to the Office of National Statistics data, made available as part of the effort to open up government statistics to public scrutiny. You can just about make out the baby boomers in the 55-64 age range. Within the next five years there are going to be very roughly a million extra pensioners, increasing the pensionable population (allowing for natural casualties) by approximately 5-9%.

If the data was collected in the same way in both 2001 and 2008 you can see that people over 65 do seem to be surviving slightly longer - on average a longevity increase of 5%. I think it's fair to assume that there has been minimal immigration within that age range. However, you should be cautious about drawing further conclusions because it's unclear how the data was collected (it might just be interpolated or an approximate guess) and this is only a snapshot of a couple of years.

It's also evident that the number of people aged over 40 is going to decrease over the next 10-20 years, reducing the pool of workers able to support both the very old and very young and the overall future trend for UK population - unless there is significant immigration - looks like it will decline. Declining population isn't necessarily a bad thing provided that it doesn't slide too far, but it will put pressure on technology to deliver automation which can at least maintain the average standard of living.

There's also a jump in the size of the population aged between 20-24. I'd guess that this represents immigration, possibly from new EU member states.

Overall the UK population increased in size by 3.8% over this period.

Bruce Perens on open source

It's been over a decade now since the "open source definition", but I think it's only within the last few years that the impact of this type of software development has begun to be felt, as open source applications have been able to compete with and exceed the capabilities of those made under the traditional secret source model, and copyright on software has been more strictly enforced. The pragmatic approach to use of FOSS within businesses taken by Bruce Perens I think makes a lot of sense. Certainly a very similar approach seems to work quite well for Google.



Will open source still be popular ten years from now? My guess would be probably yes, that this is a long term trend in software rather than merely a passing fad. I expect the process of software development to become more automated over time, and eventually it may be possible to describe to the computer in general terms what you want it to do, or merely compose systems from pluggable pre-built parts. Such automation would render most of the issues around open source irrelevant, since the code would not be intended to be human readable, although the freedom to share aspect would remain. Progress towards automation of software engineering has been very slow though, so it doesn't look as if manual software development is going to disappear any time soon.

The blue tit mafia

Photos of blue tits on a snowy bird table.

2010: The age of retirements begins

A few notable events of the last decade were:

• Fuel blockade
• USB 2.0
• Broadband internet
• Mobile phone coverage becomes acceptable
• 9/11
• Windows XP
• Digital photography
  
• Blogging
• "Reality" TV shows
• The Iraq war
• Wikipedia
• Wifi
  
• Unit Tests
• Even the most technophobic people get a computer and get onto the internet "to email my friends"
• The rise of Open Source
• Ubuntu
• Outsourcing of technology jobs accelerates
• YouTube
• Facebook
• Fuel prices surge
  
• Sub-prime mortgages
• Twitter
  
• "Quantitative easing"
• The slow demise of TV/analog switch off

Most of the above technologies, such as Wifi and broadband internet, we take completely for granted now, but were only invented or introduced to a wide user base within the last decade. We might well wonder what the next ten years will bring, and most of it will be somewhat unpredictable (except with the benefit of hindsight). Some things are completely predictable though.



We're now entering the decade of retirements, when the demographic bulge of the post war baby boomer generation start to become pensioners. This is an event which has been much talked about in robotics circles, because of the possible applications, but seems to have been barely even acknowledged in the rest of the media. It's quite easy to predict that issues surrounding retirement and old age will become more prominent over the next decade, and the so-called "grey vote" will hold more sway than it did in the past.

Some predictions for the next ten years are:

• The world recovers from recession. Business resumes more or less as usual.
  
• The construction of large industrial scale retirement complexes - applying economies of scale to elder care.
• Pension age is increased.
• The Singularity University closes. Press release announces "mission accomplished" (or similar).
  
• Political scandals as some pension companies fail, or cannot deliver the expected payouts.
  
• The Afghan war ends with a Soviet-style withdrawal.
• Electric vehicles become more common.
• Governments struggle to adapt to the increasing levels of transparency which are expected and demanded by voters. There are a series of "outings" of corrupt dealings/practices, due to the increased availability of government data and ease of use of data mining software which permits interested parties to carry out their own independent audits, separate from more traditional regulatory organisations.
  
• Some heavy goods vehicles drive themselves. This will probably happen in relatively remote areas first, where traffic volumes are low and there are fewer regulations - maybe China, an African country, Russia or Alaska.
  
• Augmented reality becomes a major entertainment medium. Although the EyeTap patent may be a discouragement to some, there's simply too much money to be made here to ignore.
• In the wealthier countries which do not have a high birth rate, unemployment falls since there are fewer people of working age, and automation is not yet capable of replacing most human labour.
• Sub-orbital space tourism becomes routine, but remains the preserve of the super-rich.
• First person younger than 18 years of age - possibly a TV/movie/internet celebrity, or the son or daughter of an oligarch - becomes a sub-orbital space tourist. Announced as "first child in space" or "first space teenager".
  
• Lunar X-Prize announces a winner.
  
• Newspapers become almost unheard of. Most people read news electronically on mobile devices/readers.
• In the political arena, new bogeymen are invented to replace Bin Laden and Saddam Hussein, but the threat from terrorism remains small.
• First space tourism accident.
• After one or more scandals resulting from inappropriate use of social networking data by companies, government agencies or unscrupulous individuals, a peer-to-peer equivalent of Facebook emerges, where users can optionally choose to host their own personal data.
  
• Telerobots become a common way to visit people, and are linked to social networks. Most are not much more than a wheeled/tracked base with a vertical pole and a holder for a mobile phone.
• First offensive use of a telerobot by a paramilitary organisation.
• The 2012 singularists will re-interpret the tea leafs and defer doomsday to some later date.
  
• For the first time symptoms of post-traumatic stress are reported amongst some gamers, as computer games become more realistic, interfaces are more immersive and games tend to be developed in a more open source way (meaning that they're less subject to traditional forms of retail goods regulation).