Has anyone done statistics on whether horses today are faster than eg in the 19th century (when records ought to be good)? Is it increasing linearly or asymptotically? Are horses running (ha) into fundamental physical problems preventing further evolution?

Asymptotically, I believe. There were a some people a few decades ago who tried to use training methods which had improved human running times, only to discover that horses' cardiovascular systems were not the same bottleneck that humans' had been. (bottleneck is with tendons and bones; racing breeds are already morphologically quite different from the wild types)

that seems to beg the question then if we only get good simulated results with guidance, then that implies we get good biological results due to some sort of directedness

it seems dawkins example undermines his argument

You are making two different points here. Natural selection is guidance, but guidance without any will or preconceived notion of good and bad. In this super simple simulated experience, you provide the guidance and depending on your guidance, you can get what fits your world. The second point is around form following function and you are inverting cause and effect. So, under natural selection, certain functions are selected for, and as form follows functions, those forms procreate and proliferate. In this simple simulated world, you choose directly for form (there is no function to choose for) and forms that are pleasing to you will procreate and proliferate. If you like whorls and I like straight lines, our created worlds will be populated by different creatures. Finally, guidance does not mean direction. A boolean guidance (live or die) does not provide direction.

great clarification of wording!

natural selects without forethought for form, yet gets fantastical forms, yet in biomorphs that is precisely how i select, and that is how i get fantastical forms. this would seem to imply natural does indeed somehow select for form

i really don't get why this doesn't undermine dawkins argumemt

I think the key point here is that selection for form need not be intentional. Taller plants are able to collect more sunlight, and collecting more sunlight means they are more likely to have many descendants. This does not happen because someone had the intention for the plants to be taller, but because of the physical fact that sunlight will hit the leaves of tall plants before it has a chance to reach the leaves of short ones growing below. Natural selection only cares about form insofar as it affects reproduction, but since form absolutely does affect reproduction, the result is that natural selection acts on form.

but is that sufficient to generate the forms we have? our intuition is intelligent intention is necessary, and dawkins argues it is not, but then demonstrates with a simulation that requires our intelligent intention

The whole argument is circular, yes. It's a metaphysical problem. One can't conclusively proove anything with physical evidence.

One can use the word "natural" to mean "unguided", but that position is just begging the question. It assumes nature and its laws are inherently unguided. Why did virus A proliferate and not virus A'? Why did gene B mutate and not gene B`? Is that "random"? Well, we don't have evidence either way, do we? We can assume, I suppose.

what does 'random' mean, even?

Sexual selection is the term for form for form sake.

Many people mistakenly take 'natural selection' and 'sexual selection' to be interchangeable terms but 'sexual selection' is the term for the mechanism that explains why you find features in some species that seem to have astatic value but not survivability value. For example peacock tails.

where did the aesthetic come from originally?

Say there is some actual adaptive advantage like upper body strength. Genes that cause greater upper body strength tend to survive, and because of this genes in the opposite sex that can detect upper body strength in a potential mate tend to survive. The population of individuals of the opposite sex, tend towards a preference for upper body strength.

What is the limit? Well the answer is very little, while more upper body strength may have little additional value above some level there is no environmental influence that selects against it until it becomes so pronounced as to become an impairment. If there are mutations in genes of the opposite sex that prefer evidence of upper body strength it is unlikely they will be removed, while genes that disregard upper body strength will occasionally fail to distinguish good quality mates and will therefore become less prevalent. Further, the appearance of upper body strength may become so preferred that fake upper body strength is also a more successful trait at least to the degree that it doesn't impair survivability.

Because of all of this, sexual selection is a mechanism in which a small astatic for actual survivability traits becomes a large exaggeration of traits that suggest survivability but may or may not actually improve survivability.

In the case of the peacock tail, and birds in general it probably stems from the fact that sick birds often have dirty tails, therefor females that prefer male birds with not dirty tails (or better) are selected for.

What is a fantastical form and how can you determine whether something is objectively fantastical instead of merely fulfilling the preconceptions of whatever creature has the subjective experience of deeming something fantastical because deeming something fantastical is advantageous for the survival of that creature?

Dawkins defines 'fantastical' as great complexity that fits an a priori definition.