vetta project

Imagine the world: all the countries, all the towns, all the cities, all the mega cities, and all the 6 billion people living in them.  Think about all the places you’ve been too, and all the places you haven’t been to, and just try to get a sense of the vastness of this for a moment…ok?

Now, try to imagine 20 times this scale.  A street with 100 people on it now has 2,000 people.  Regular two story family homes are now 40 story apartment blocks.  Big skyscrapers now have 2,000 floors and are 10 km high, taller than Mt. Everest.  Or if you prefer to go out, rather than up, take New York and drop another 19 of them around the United States.  Then do the same for Los Angeles… and so on for every city and every town, and continue this way across the whole world.  The planet would be burried under a seething mass of humanity.

That’s 120 billion people, approximately the number of neurons in your brain.

Now give each of these 120 billion people a cell phone, and load each one up with something like 5,000 phone numbers, mostly of people who live in the individual’s area.  Get everybody on this planet to start sending messages to each other.  Some only slowly send messages, others are busy sending 200 messages a second to all 5,000 people that they know.  Now let this system start to adapt in order to control which messages go where and when…

That’s about the scale of your brain.

When people say they can’t believe that the human brain is “just a machine”, I suspect they are suffering from a lack of imagination — have they seriously tried to wrap their mind around how unbelievably profoundly gigantic this machine actually is?

Yudkowsky has been posting a lot on Overcoming Bias recently about his theory of metaethics.  Today he posted a summary of sorts.  Essentially he seems to be saying that morality is a big complex function computed by our brain that doesn’t derive from any single unifying principle.  Rather, this function is a mishmash of things and even we don’t really know what our own function is, in the sense that we are unable to write down an exact and complete formulation.  It’s just something that we intuitively use.

I’m not convinced that ethics can’t be derived from some deeper unifying principle.  I’m also not convinced that it can, lest you misunderstand me.  What I do accept is that if this is possible then finding such a principle and convincingly arguing for it is likely to be difficult in the extreme, and probably not something that is likely to happen before the singularity.  Nevertheless, I haven’t yet seen any argument so devastating to this possibility that I’m willing to move it from being extremely difficult to certainly impossible.  Any system of ethics that does derive from some unifying metaethical principle is almost certainly going to be different to our present (western?) ethical notions.  I think some degree of this is acceptable, given that our ethical ideas do change a bit over time.  Furthermore, no matter how human we try to make the ethical system of a powerful AGI, post-singularity we are still going to be faced with ethical challenges that our pre-singularity ethics were never set up to deal with.  Thus, our ethics are going to have to be modified and updated in order to remain somewhat consistent and viable, otherwise we’ll end up with this kind of nonsense.

Anyway, let’s assume that this unifying principle either does not exist, or at least can’t be found.  How can we tell if an AGI is ethical given that we can’t explicitly and completely specify what this means?  This seems like the problem Turing faced when trying to determine whether a machine is intelligent or not.  He figured that he couldn’t explicitly and completely say what intelligence is, unlike the research by Hutter and myself, and thus he tried to dodge the issue in the obvious way by setting up an imitation game that doesn’t require an explicit description of intelligence.

Here we can do something similar: set up a group of people and the AGI and ask them ethical questions from a panel of expert judges.  If the judges cannot tell which the machine is, then it passes.  Given that the morality function varies between people, and that we can’t say explicitly and completely what our own function is, this seems to be about the best we could hope for.  Naturally, this doesn’t prove that the AGI, or indeed any of the humans participating, are “good”.  An evil genius could probably pass such a test.  Rather, it is simply designed to test whether the AGI is at least able to compute a version of the human morality function which is sufficiently similar to ours that it is able to pass as being human.  Whether the AGI (or human) actually takes its human-passable morality function and reliably and consistently seeks to follow it into the future is a whole other set of problems.  Thus, passing such a test is perhaps a necessary, but certainly not a sufficient condition for having an ethical AGI.

I’m sure somebody must have proposed this idea before, but at least my half hearted attempt to find the idea on Google didn’t turn up anything.  I should also point out that in order for this test to work you’d probably want the AGI to pass a more general Turing test first so that it doesn’t get singled out by the judges for various other reasons.  Only then should you bring in a group of expert ethicists to try to judge which of the test subjects was ethically inhuman.  We would also want to include in the test subjects a few very nice people and a couple of professional ethicists as we wouldn’t want the AGI to be able to “fail” for being too nice or consistently ethical.

Remember when I was raving about nVidia’s new GTX 280 graphics card that crunches 1 Tera FLOPS?

Yeah, well, that was 3 weeks ago.

Today, Radeon’s new HD 4870 X2 graphics card has 1600 stream processors that crunch 2.4 Tera FLOPS.

While The Register isn’t the most reliable source, the time frame and specs look about right so I figure this article probably isn’t too far off.

Name: IBM Blue Waters
Peak performance: 10¹⁶ FLOPS
RAM: 620 TB
Location: University of Illinois
Delivery: 2011

For every synapse in an adult human brain this machine will have around 6 bytes of RAM and 100 floating point calculations per second.

So you’ve finished writing your thesis, your pièce de résistance, your magnum opus. Next step, get it printed as a book for the world to admire. Of course, being the misunderstood genius that you are, no professional publisher will want to touch your great achievement.  Never fear, the internet age is here! So you gather up your LaTeX  files and head off to lulu.com, only to find that lulu has no idea about LaTeX. A search of the lulu help system literal returns no results.  Google returns fragmented and in many cases possibly out of date suggestions. If this sounds like you, read on.

[Read more →]

My thesis is now available at lulu.com.  As promised, it’s at cost, which works out at $18 plus shipping.  It’s all under a creative commons licence and in a few months I’ll put the pdf online for free.  I’ll also write a post shortly on all the tricks involved in publishing on lulu.com with LaTeX, in case you plan on doing something similar.

Table of Contents

[Read more →]

Many people seem to have questions about Kolmogorov complexity, Solomonoff induction, algorithmic probability theory, AIXI, the universal intelligence measure and so on.  I don’t always have time to watch all the email lists where these things get discussed, but if you do have any questions, concerns, etc. that you’d like to put to me, feel free to post a question below and I’ll try to answer it.

In 2006, a computer capable of 3 T FLOPS was enough to get onto the list of the top 500 super computers in the world.

Two years later…

This PC has 3 nVidia GTX 280 graphics cards, costs about $3,000 and is rated at 3 T FLOPS.

This morning I received the wonderful news that I’ve won the Singularity Institute for Artificial Intelligence - Canada Academic Prize for 2008!

The award is in “recognition of [my] efforts to improve AI theory” and is worth CAD $10,000.  This will certainly help my budget over the next two years while I study at the Gatsby Unit in London.  So, thank you to SIAI Canada, and to all the Canadians whose donations made this money available!

Speaking of my research, after a long weekend of final edits, corrections, formatting, indexing, embedding fonts and other complexity (I’ll write a blog post about what I had to do at some point), I’ve finally uploaded my thesis “Machine Super Intelligence” to lulu.com and have ordered a test copy.  Once I’ve checked that everything is ok I’ll let you know where copies can be ordered.  Copies should be USD $18 plus shipping for a 200 page casewrap hardcover.  Probably about in a month…

This morning I passed my PhD defence… making me now Dr. Legg.  Soon I’ll upload my thesis “Machine Super Intelligence” to lulu.com where you will be able to pick up a printed copy at cost if you’re interested.

And in other news, I’ve been awarded a grant to study machine learning and theoretical neuroscience at the Gatsby Unit, University College London.  I should start that in January and be there for two years.  For the meantime I’ll remain at the Swiss Finance Institute working on portfolio choice models.