The seed of reason

Breakthroughes III

Yet more possible (or current) breakthroughes in technology together with a very basic description, critique, likelihood and timescale.

Development: High capacity hard drives >10 terrabites
Likelihood: High
Importance: All the (English alphabet) text in all the books in the British library would fit on a drive with 10 terrabites of storage.  With this technology there would be the potential to give billions of people access to almost everything that has ever been written.  The pace of scientific, technological discovery as well as the literary prospects are staggering.
Drawbacks: Copyright infringement.
Time scale: A decade.

Development: Thin film dye solar cells.
Likelihood: Currently being commercialized.
Importance: By reducing the cost of a solar panel by as much as ten times these devices could make solar power competitive with coal.  The potential to reduce our dependence on fossil fuels will make preventing catastrophic climate change that much easier.
Drawbacks: The dyes used degrade too fast under sunlight.  The MIT scientists working on this think the rate of degradation can be sufficiently reduced so that the solar cells would have decades of useful life.
Time scale: 3-5 years.

Moral progress: moral knowledge

This post is part of my series on the belief in progress.  This blog post deals with the evidence for a positive trend in the existence of moral understanding.  In this post I am not arguing that moral understanding leads to good moral behavior, that an increase in moral understanding is inevitable or that moral understanding has not increased for non moral  reasons (slavery in the US (and arguably in the UK) for example was abolished at the time it was for economic reasons).  I am merely pointing to an extraordinary trend which I think requires explanation.

The world is improving in moral understanding:

10 Billion years ago: There was no life so there could be no moral understanding.

Clearly there can be no moral understanding before there are entities capable of decision making.  Hence the beginnings of moral understanding can't have been earlier than around 1.4 billion years ago when animals originated.

1 Billion years ago: The emergence of animals allows decision making.  Basic elements of morality such as helping your relations and some forms of cooperation would have started then.

1 Million years ago:  Until humans we have little knowledge of the state of moral understanding.  However, we can make shrewd guesses as to the origin of morality.  Various other organisms engage in reciprocal altruism (a relationship not unlike friendship) and there is evidence in various primates of individuals being punished by third parties for misdemeanors.  These abilities seem to depend largely on the intelligence of the animal concerned and hence their brain development.  As brains probably have got more developed over time this would suggest that these capacities for proto moralities would have developed with them.

10,000 years ago: With modern humans came crude forms of morality.  Most of our modern conceptions of social morality have analogues in this time. 

100 years ago: Slavery, looting and pillaging are now generally regarded as wrong.

Now: In the last 100 years things have got considerably better in some of the world and stayed the same or got slightly better in much of the rest of the world.  Clearly the West has had most of the benefits but it should be emphasized that this still means on average that things have got better.

In much of the world expanded suffrage is recognized as good.  Freedom of the press and individual liberties are now also regarded as worth possessing.

In addition in the last 50 years there have been further significant developments.  Racism, sexism and homophobia are now considered wrong in much of Europe (Racism and sexism elsewhere too).  New movements such as vegetarianism & fair trade have just burst on to the scene and have made significant advances in the last 20 years.

I will address the question of moral behavior in another blog post.  I shall also be trying my best to answer the question "What causes this trend?"

Computation

The advent of computers have spawned many new areas of research including computational complexity theory, chaos theory and of course AI.  I shall give several examples of where a computational framework for understanding may be fruitfully applied.  I will write further blog posts on particular interesting aspects of this framework and I will list the topics I intend to cover.

Modern economics:  I have already argued that a world with AIs would have a radically differing economy to that of the present day.  I believe that various trends can already be observed whose direction is towards such an economy.

The mind:  If the mind can be understood as a Turing computer then this implies that it must have certain limitations.  Carefully analyzing these may help us understand neuroscience and also help us achieve AI.

Ethics and Politics:  I shall argue that minds must be understood as highly parallel computers.  This has profound ethical and political implications.  I shall spell out what these are and how one would argue for them.

Mass creative collaborations

I have been interested for some time in the idea of mass creative collaborations.  By this I mean hundreds or thousands of people working together on a work of art, a book, a film or a piece of music.

Now I'm not talking about an artistic endeavor controlled by one or two individuals but involving many more.  Rather I'm speaking of a much more egalitarian undertaking with no individual controlling the whole process and at least 50% the work the work of the 80% least prolific contributers.

I know of no good collaborative creative work (there's plenty of collaborative work outside of the arts though).  I shall list a few reasons why I think this is:

1) The creative process often involves developing material which is not part of the finished work (like creative scaffolding).  Existing collaborative projects do no seem to have space set aside for this sort of development.

2) The creative process often requires a great deal of time and effort getting yourself acquainted with your own work.  Collaborative projects are often structured in a way that does not encourage more than a passing contribution.  Most projects are not commercial and so contributors cannot expect financial reward (it is difficult to guess how rewards should be apportioned anyhow).

3) Any project needs some sort of filter as to who may contribute.  Even assuming everyone contributing to a project is creative they may well have inconsistent ideas for a project.  Projects need filters which effectively select people to work on specific projects.

4) Creative projects base themselves around traditional art forms when it may well be the case that there are new forms to be discovered which lend themselves to collaborative creation.

I suspect that interesting collaborations are not only possible but will end up producing stunning creative works.  However, we still need to find the right ways to structure projects of this type.

Breakthroughes II

Yet more possible (or current) breakthroughes in technology together with a very basic description, critique, likelihood and timescale.

Development: Directional speakers
Likelihood: Already made
Importance: Allows sound to be to directed to a localized area.  Useful for communication, to reduce sound pollution and for luxury applications.
Drawbacks: These are ultrasound devices and may interfere with bat sonar amongst other things.
Time scale: Next decade.

Development: Multi touch sensitive screens
Likelihood: Already on the market
Importance: Allows intuitive multi user applications.  Allows users to interact with computers in an entirely new and natural way.
Drawbacks:  Requires much more energy to use a computer.  Users get tired arms much quicker than when working at a normal PC.  On the other hand its probably more healthy!
Time scale: 2-5 years before technology becomes affordable for the general public.

Why vegetarians should want in vitro meat

As a vegetarian I find people often ask me whether I think that vegetarianism will ever become the norm or even whether killing animals for food may ever be made illegal.

I always answer: "Yes I do believe that in a century or so almost everyone will be vegetarian* and that it will be illegal to eat most animals (not insects) for food."

In this blog post I shall give some reasons for these opinions.  I warn you now that the explanation I will give is incredibly cynical.  But that is precisely the reason that I am confident in my argument.

I think that major political changes occur for economic reasons.  I observe that most people will only consider something to be immoral if that thing does not benefit them significantly (assuming there is some other reason to think it is immoral).

So what is the economics of meat production?  Meat is not an effective way of producing protein, vitamins or calories.  90% of the energy that is present in the food cattle eat is lost before it becomes a steak.  Modern agricultural techniques reduce this loss but not eradicate it (of course those techniques are horribly immoral).  So meat is not an economic way of feeding people.  This makes meat a luxury good.  It is a drain of the productivity of the economy.  But we can expect meat from animals to be consumed until there is a luxury good to replace it.

In vitro meat is that luxury good.  In vitro meat is meat grown in vats with no brain or nervous tissue and hence without the moral problems of animal flesh.  The technology is still in development and may take decades to mature but when it does we will wake up to a world in which there is meat that is (a) cheaper, (b) (eventually) more safe and (c) without moral problems.

When this happens economics will encourage most people to stop eating slaughterhouse meat.  Once this has happened I expect to see the moral zeitgeist move on.

Is this realistic?  Can such a massive change really be effected?  I believe so.  But first we must see vegetarian alternatives reach a high degree of sophistication.  Probably the best way now to support the vegetarian cause is to fund the science behind these in vitro meats.

* By vegetarian I mean moral vegetarian:  Someone who does not support economically the slaughter of animals primarily by not buying slaughterhouse products.

Biofuels

Biofuels have recently been implicated in the dramatic food price increases.

Many current biofuels are made from food products such as sugar cane or corn.  The food is broken down (crushed and strained) before being fermented to make ethanol.  This ethanol can then be used as an energy dense fuel.

However, the greater the demand for these biofuels the greater the demand for key foods.  Together with plagues of crop diseases, droughts and trade barriers (ironically erected to protect national food supplies) this has pushed up the price of rice and wheat by more than 3 times.

The obvious alternative, according to many, is to take plant material that cannot be eaten by us and convert this into sugars before fermenting it to make ethanol.  Various biotech firms are currently working on this technology.  Craig Ventor (of gene sequencing fame) is one of those involved in this research.  The idea is that agricultural waste that is currently just composted could now be exploited to make environmentally friendly fuel.

This all seams like a great idea to me, neatly sidestepping the problem and increasing the potential production capacity simultaneously.

However, I wanted to make a point about the effects, which large scale production of cellulose based ethanol would have.  Much of the food stock for animals in the meat industry comes in the form of cellulose.  Any large scale production of cellulose based biofuels would therefore imho dramatically increase the cost of meat.

Now I am in favor of this as a vegetarian.  However, I do wonder if the general public will be happy for this squeeze on their lifestyle choice.

Breakthrough technologies

We live in interesting times.  Huge changes are making themselves felt in all areas of life, the economy, science and technology.  Like a magpie I have been collecting interesting potential developments in technology.  Now I bring all of these together and in a series of postings will describe, critique and rate their likelihood.

In each post I shall give two possible developments, my estimate of their likelihoods, why they would be important, what draw backs they have and how long before we can expect to see them.

Development: Fusion power
Likelihood: Very low
Importance: Important because it would open up a huge and so far untapped energy source.  Hydrogen is readily available in the form of water.
Drawbacks: A fusion power plant creates radioactive waste.  The plant would almost certainly require intricate and costly maintenance.  The cost of this  maintenance may be difficult to estimate before the power stations come online.
Time scale: At least 40 years away.  Fusion power always seems to be 40 years away!

Development: Room temperature superconductors
Likelihood: Low but feasible
Importance: Important for many reasons.  Room temperature superconductors would increase efficiency in power transmission, allow power to be moved effectively between continents, decrease the size of many household appliances and bring us practical levitation technology decreasing the cost of moving freight on land.  Room temperature superconductors would also hugely speed up computers and decrease the cost of various peaces of medical equipment.
Drawbacks: Would probably lead to new types of weapon.
Time scale: Not for another 10 years at least.  A room temperature superconductor could be discovered tomorrow but it would take a decade to mass produce it.

Progress and physics

This post is part of my series on the belief in progress.  This blog post deals with the evidence for a positive trend in the existence of physical understanding.

This one is a fairly clear and unambiguous trend and as usual I will take the trend back as far as it can be discerned.  In doing this I shall be thinking in information theoretic terms.  I am looking therefore for systems in which useful information about physics can be found.  Some of this information is stored in strange ways.

Strictly speaking a cell cannot 'understand' the laws of thermodynamics in the sense most people mean by the word.  However, a cell may contain implicit information about the laws of physics.  Before humanity I shall look for information stored in this way.

10 Billion years ago:  There was no life so there could not be any understanding even of an evolutionary character.

3 Billion years ago:  Rudimentary organisms now exist.  Within these primitive cells enzymes exist which exploit various biochemical laws.  Those enzymes represent a store of useful information concerning biochemistry that will help in the development of new species of life.

0.5 Billion years ago:  Organisms now exist which possess skeletons.  The growth of skeletons in vertebrates is controlled by a class of homeobox genes.  Slight changes to these genes (and associated regulators) allow a space of skeletal forms to be accessed which will hold an animals form up against gravity.  The homeobox genes contain information concerning structural design which will encourage the development of new forms of life.  In addition genetic programs that produce various organs of the body encode information tacitally about the laws of fluid mechanics, diffusion, biochemical laws, optical laws and energy conservation.

1 Million years ago:   A species exists (quite possibly more than one) which has learnt to tame fire.  The knowledge of fire is a basic but never the less useful level.

5,000 years ago:  Boats are now common.  Their building requires a degree of cultural knowledge about gravity, buoyancy and the structure of wood.

100 years ago:  Mathematics has been extensively used to create highly accurate theories of the world.  The two most successful theories of modern times (relativity and quantum mechanics) are just about to burst forth into the world.

Now:  Our understanding of nanoscale properties of materials, of emergent physical systems and of chaotic systems is still in its infancy but growing fast.

We are finding out that our knowledge of cosmology and of the standard model of particle physics is less certain than we believed.  Further significant developments are surely inevitable.

3D circuits, nanotech and AI

This blog post follows on from my last.

Today I shall give my reasons for predicting that computational capacity will continue growing for a while after 2020 (and why 3D chips will allow this).  I shall also be explaining what that means for the development of AI and what other barriers may emerge for increasing computational capacity.  It will all be back of the envelope calculations but I hope they won't be too misleading.

Making chips 3D will allow us to get around the communication barrier to increasing performance as the memory of a computer will start to move onto the processor itself.  The number of transistors, the speed of the chip and the energy efficiency should go up too because building in 3D means more space, very short wires and

Of course the current technology of stacking chips on top of each other isn't great because it increases the cost of production and once you're stacking 20-30 chips its getting very difficult to do so accurately enough (not to mention the difficulty of keeping your chips cool).  So what might be the next technology to take over after this?

At some point in the next 20-30 years I expect we shall start to produce 3D chips as solid (although probably porous) lumps of semiconducter.  I have two suggestions as to how this might be achieved together with some difficulties each face:

X) A self assembling nanoscale crystal.  The idea here is to create nanoscale particles (for instance single stranded DNA chemically attached to bits of semiconductor) which link up like a 3D tesselating pattern.  The pattern is your 3d circuit.

Defects in the crystals growth are inevitable.  Care must be taken to ensure that errors in the crystal's growth are self correcting and do not effect the final circuit.  The nanoscale particles can be mass produced using today's technology.  However, predicting the 3D structure a selection of nanoparticles will create requires great computing power.

Firstly you must be able to predict how the DNA strands will fold.  The algorithms for this are still under-developed but once they are worked out we can expect them to take at most 10,000 processor hours (circa 2008) of CPU time (about £200 of energy) per folding.  My source for that is folding@home.  The figure may be much better as its unclear when the page was written.

Secondly you must be able to predict from that the crystaline structure grown and its electronic properties.  Thats probably very hard although I'm guessing of about the same difficulty as simulating that amount of DNA folding.

In order to determine a 3D chip structure with millions of bits worths of information to describe it you'd need to tailor design millions of these nanoparticles and determine their growing properties and grow the resulting crystal at a very slow rate to reduce errors.  If say you use a chip structure with a million bit description, and each DNA strand needs 10,000 folding operations to design then you'd be looking at a capital outlay of about £200 * 10,000 *1,000,000 or £2 trillion.  This is an upper bound on the cost of design of such a chip (once the simulation technologies for crystal growth and DNA folding are available).  The cost could be reduced by advances in computation efficiency, computer hardware or by reducing the complexity of the 3D chip.

Y) A 3D lithographic process.  Cancer therapies routinely target precise volumes in a person's body to destroy tumours.  Similar technology might allow us to burn a pattern of 'wires' and transistors into a 3D block of semiconductor.

There are at least two limitations here.  Firstly we need a substrate which will respond appropriately when heated.  Secondly we need the chip to be largely transparent to the energy beam (electromagnetic or electron stream possibly).  Thirdly we need a technique that can deliver highly complex patterns to be burnt into the 3D substrate.

Whichever way you cut it its difficult to get the information into the substrate.  If you want to specify the structure of the 3D chip down to 12nm then you'd need to get 508 quadrillion bits or about 72,000 terrabytes into the device somehow.  Its probably best to opt for a chip design which is massively repetitive to avoid these fluxes of data.

Never the less there would be a huge advantage to specifying the structure precisely at that scale.  Chips designed for a specific purpose are often between 10 and 1,000 times better at their task than multipurpose chips (The chess playing computer deep blue is an example).  The amount of data moving into a 1cm cube of substrate can be approximated if we know how accurately we can focus the energy beam (probably around 12nm is feasible) and how quickly we can modulate the beam.  In order to write the information to the chip in under a second (remember this things must be mass produced) we'd have to be able to modulate the beem at about 0.3 million times per second.  Doing this and still accurately creating the chip amount to a formidable challenge.  But it is not obviously impossible.

Now I shall address how these technologies effect the prospects for AI.  The human brain is very memory demanding.  If an AI we create is similar to the human brain (a big if) it will probably need as much memory.  Even if it isn't it will probably be much easier to design an AI which has access to a large amount of memory.  So 3D chips are going in the right direction.

Landauer's principle implies that the human brain operates at about 3% efficiency (assuming that it does not use reversible computation).  However, I am not sure whether to trust this 'law' as it has recently been challenged.  Interestingly it implies that the capacity of the human brain cannot be as high as some estimates put it.  Clearly something has to give.  If the law does hold it should be possible to match and exceed the human brain in energy efficiency.  Actually for AI to be economical it wouldn't be necessary to match the energy costs as human consume much more energy than their brain needs.

How good do we need to be at making 3D circuits to match the human brain's capacity for reasonable outlay.  Assuming Hans Moravec's estimate of the human brain's computational complexity is accurate then I calculate that even an accuracy of a few hundred nanometers would be enough.  At 100nm we'd have both the memory and the processor capacity.  As the human brain has relatively few types of neuron (<1,000,000) the complexity of the circuits won't have to be too high so the considerations about delivering design information when creating the circuits doesn't apply.

All in all it seems feasible to make 3D circuits which have the same capacity as the human brain.  This should make us considerably more confident that it will happen and more so that it will happen soon.