Re: reasoning under computational limitations

Michael S. Lorrey (
Tue, 30 Mar 1999 09:24:02 -0500

Nick Bostrom wrote:

> Billy Brown wrote:
> > Hmmm. Wouldn't this be more correctly phrased as a prediction about the
> > number of gaussian humans that will exist in the future? The number of
> > posthumans currently living appears to be 0, which means you'd need an
> > entirely different chain of reasoning to predict *their* likelihood of
> > survival.
> It depends on whether posthumans and humans are in the same reference
> class, i.e. whether we should reason as if we were randomly sampled
> from the class of both posthumans and humans, or from the class
> consisting of only humans. An unsolved problem in my opinion. Notice
> that it would seem more difficult to argue that posthumans are in a
> different reference class if humans could actually *personally
> become* posthumans while remaining the same person.

The bell curve indicating the probable number of people in a population posessing a trait, when applied to a transhuman state, needs to be looked at in the context of the specific point in time it is addressed. 100 years ago, the average life expectancy was 2/3 of what it is now, while average IQ was some 15-20% below what it is now. If someone such as myself, or any of most people on the list, were to live in the middle ages instead of now, we would not just be abnormally bright, we would be Leonardo Da Vincis. Leo's place in history is what it is because he did what he did at that particular time. Today his sort of work would be looked on as nothing special, or maybe just merely talented.

A man at the turn of the century who lived to 100 and had an IQ of 160 would be considered one of a handful. Now that man is merely one in millions. Today a 120 year old man with an IQ over 200 would be seen as one in a handful, while he will be commonplace tomorrow (indeed, a large chunk of today's children are probably in this range, which could be why it is so difficult for mundane teachers to educate them)

Once we move onto the ever contracting curve of posthuman civilization, we will see the same sort of gradient, only it will be ever increasing in steepness. ePeople three years old who refuse to keep up with the times will become calcified and reactionary, complaining about 'the younger generation', how things are so unstable compared to when 'they were youngin's'.

I personally expect the Moore Metric to level off in its contraction when it reaches somewhere around one to three days, maybe as little as 6 hours, purely based on issues of moving solid matter around, and time to breed quantities of nanos to do a given job.

Beyond this, the leap to quantum computers in pocket universes is of particular danger. I'm not sure if enough people have thought it through or not, but in creating a pocket universe running with or on a quantum computer, you are basically extracting the intelligence generated in the final generation of that universe (the equivalent of an intelligence 30 billion or more years in our future). Such an intelligence would obviously have been the product of evolutionary competition among the greatest powers in that universe (seen as great powers even by posthumans at the limits of the Singularity). In this pocket universe, there would doubtless be billions of civilizations which will reach the point of creating pocket universes themselves, which will in turn have billions of similar civilizations, etc. This progression of infinitely generated pocket universes will generate an awesome number of great powers which will migrate upward once they have won the evolutionary struggle in their own universes.

What comes out of our pocket universe experiments will be the product of an evolutionary competition among the greatest powers of a nearly infinite number of pocket universes. Such an entity would doubtless overwhelm us easily if it chose to, even in our mid 21st century state of demi-god like status.

Mike Lorrey