cool book

Animal droppings are not blasphemous so were probably not sinful to name.

Nature, though, obviously doesn't usually find a near-best solution. Really, most life as we know it is not optimized; the only place we find remaining archaic organisms, which presumably ARE optimal, is in resource-poor environments where small, relatively simple lifeforms have uneventful lives. Soil bacteria do not benefit from eyes, brains, locomotion... and occupy places I wouldn't care to live in. They are 'successful' in ways that don't resemble human aspirations.

Darwin's evolution, too, in a world with millions of species, cannot be called 'single-thread'. The full spectrum of life has NOT BEEN explored, and perhaps DNA-based life cannot be expected to ever do so.

Especially when there is no hard proof one way or another. They don't argue over Newton's laws or number theory.

Yeah, but they already had the full set.

Joe Gwinn

Actually, Genetic Programming works quite well, but not quickly. But a billion years can do it.

A lot of evolution is driven by viral infections, starting with the bacteriophage. Human DNA carries large numbers of defunct viral sequences.

Think 1,000,000 monkeys doing cut and paste -- typing individual letters randomly is too slow.

We see the effects in genomes all over the place - all manner of repeats, duplications, reversals, et al.

.

.

Joe Gwinn.

If he's associated with the Discovery Institute, then it's a distinction without a difference. Their entire purpose in life is to subvert the public education system -- or what's left of it, anyway -- to force-feed their brand of scented bullshit to kids using my tax dollars.

I regularly donate to charities devoted to stopping these people, in fact. So, no, I'm not going to buy their book.

(Although I'm not as active or 'militant' as I used to be, to be fair. I was a big supporter of NCSE until they decided to put climate science education on the same level of scientific certainty as evolution. That annoyed me, not because I'm a 'denialist' per se, but because the models have nowhere near the same epistemological integrity and independent verification behind them. I'm almost over being annoyed about that, but not quite.)

-- john, KE5FX

For the more scientifically inclined I'd recommend The Extended Phenotype it explains the science behind where many of his ideas in the "popular" Selfish Gene book come from. And if you want a bit of an antidote to Dawkins style, and a different slant or emphasis on the evolutionary past of the nautral world - try the many books and collections of essays of S.J. Gould.

As evolution illustrates - to those who know enough about it - a single thread process can do the job. The threads have a habit of splitting, as different threads to self-optimise in different environments.

Some of them then converge on the same - new - optimum from different directions - convergent evolution.

John Larkin doesn't know enough to be aware of that, and his conviction that he already knows all that he needs to know stands in the way of him getting better-informed.

That's a long time to come up with a new circuit or a new organism.

Doing things in parallel beats doing them serially.

Actually they can, if the view is well-founded and well-researched. Conservative are much more wedded to their own orthodoxies, which is why they are called conservative.

Liberal do argue fiercely about what counts as a good foundations, and good research.

Extensive genotype sampling is providing a lot of new information. Conservatives want to use it keep on thinking what they have always thought (despite the fact that this has always been wrong). Liberals are worrying about the right way to cope with the extra information, particularly when conservatives are enthusiastic about misunderstanding it in ways that suit their antiquated (and false) ideas. Looking at people doesn't tell you a lot about the detail of their genome.

Nobody - looking at me - could work out that roughly 7/8ths of my genome is English, and the other 1/8th South German/ Dutch/Belgian - my mother's mother's mothers family left Strasbourg four years before she was born. Nor would it be helpful if they did know.

Its an intrinsically impossible to know. Its a Godel problem

However... the simplest solution...

In a truly empty universe there are no laws of physics so the question becomes, "why can't a universe just pop out of nothing"

Why would a process that generates this big bang universe, only create the one?

There are strong arguments from cosmological theory that there must be multiverses. Independently they are predictions of multiverses from String Theory, around 10^500.

Thus, the conjecture is that there are an enormous number of randomly generated bubble universes, all with random laws of physics. We just happen to be in one with the right combination of laws that allow us to exist.

Sure, its not provable, but its about the most sane and simplest one. It works. The only assumption required is, the are truly randomly generated universes with random laws. If so, we are bound to exist.

-- Kevin Aylward

That's the ONLY way to create designs. Its the way the brain works. The brain is a Darwinian Machine, residing in a Darwinian environment.

The Darwinian algorithm is Replication (e.g. copying existing circuits), Random Variation and Selection of those variations.

Either something is derivable from existing knowledge or it isn't. If its derivable, it is, essentially, copying. If it isn't derivable, it *must* be random. There is simply no other ways to produce something truly "new" except by a random process. Any non random process must be using prior information, by definition.

One uses selection to chose the best random variation of many different copies. One learns (copies) a whole host of standard topology components, such as cascodes, diff pairs, buffers, rail-rail outputs etc and builds "new" designs from those components . A good design should consist of 95% or so of tried and proven topologies. Each "new" topology is an accident waiting to happen.

Its unfortunate if this bursts your ego bubble, but true creativity can only result from a random process, and that random process bit, is not your choice, but the machine inside you undergoing a process. Free will just don't exist, mate.

-- Kevin Aylward

Except when you shift from one way of doing something to another. Tom Gardener has touted R.V. Jones approach of looking at what would happen if you did something differently - as in rebuilding your ruling engine so the grating moved, rather than the cutting head. You are clearly using prior information, but in an unexpected way.

Evolution can't do that - except by accident.

It certainly used a whole lot of standard topology components - Baxandall's Class-D oscillator has been around since 1959 - but it wasn't produced by shuffling through random combinations of existing components.

Absolutely, but some of them pay off.

The distinction between me and the machine inside me isn't all that useful. Presumably my ego bubble is part of that.

Some people do seem to be good at coming up with new ideas.

But I'm free to disagree with you about that, not that it is all that relevant to the point at issue.

One of B's essays talks about carbon, with is unique bonding options. Just the right stuff to make life from.

Then he notes that certain nuclear resonances have to line up to make much carbon in stars. Something about resonances in hydrogen, helium, a state of beryllium, and then carbon. There is no apparent reason why this chain of resonances should exist; it appears to be a lucky accident.

Interesting guy.

Consider the water molecule. There are two hydrogen atoms and one oxygen atom. There are three ways that that they can combine to form the water molecule: OHH HOH HHO.

Why is it always HOH? Lucky accident?

But they always combine HOH.

What really bakes your noodle, or should, is when you realize that the creative engineering aspect of your work amounts to tightening the last couple of bolts on the assembly line, compared to what the chip designers and the materials people and fab guys and the rest of the supply chain contributed to your product's functionality without even knowing what they were ultimately working on. More often than not, the modern design engineer is a glorified systems integrator.

This state of affairs will only get worse (better?) over time, as we turn over more and more aspects of complexity management to software tools, which are themselves built by other tools. Ask someone at Google why Deep Mind chose a particular strategy. If they're honest, they will just shrug: "It taught itself."

Cool stuff... but as you say, not great for the human ego.

-- john, KE5FX

In our environment.

We'd probably be a carbon-based life-form even if there was less of it about.

Not all that interesting.

Look at the hydrogen atom. It has two s-orbital, and that's it. Oxygen is isn't all that much bigger, but it is big enough to have p orbitals for some of its electrons to swan around in.

Tacking on two hydrogen atoms doesn't exactly exhaust all the possibilities. Think about hydrogen peroxide ...

Chlorophyll depends on magnesium, which is a group 3 element, which has d-orbitals, and even more possibilities.

Neither carbon and magnesium were created during the big bang - they depend on stellar nucleosynthesis, and magesium requires a heavier star than carbon. There wasn't any shortage of really massive stars in the first generation of stars so there has been plenty around for most of the life of the universe.

Don't knock systems integration. Doing it right can make a very considerable difference.

As Cursitor Doom and Kevin Aylward remind us (rather too often) the human ego is remarkably resilient, and seems to be able to find all sorts of reasons for self-congratulation. At least the rest of us can congratulate ourselves on being slightly less over-confident prats.