Posted as science, because it makes vibrantly clear how important it is to measure precisely.
H/T to DrewM. over at AoSHQ

Posted by: leoncaruthers at 07:27 PM | No Comments | Add Comment
Post contains 30 words, total size 1 kb.

Brian Wang has an excellent post up about new funding for diamondoid mechanosynthesis viability. If you've never heard of it, it's the centerpiece of Dr. Eric Drexler's classic book Engines of Creation (highly recommended, very fun read, and it's free for download here). His follow-up, far more technical tome Nanosystems is also recommended. The basic idea is simple: with the right tools at the molecular scale, you could build working machines at that scale, molecule-by-molecule. The deliberate fabrication of these machines through mechanical action is why it's called mechanosynthesis, and "diamondoid" refers to the fact that nearly all such products will be given structure by carbon atoms in diamond configuration.

The ramifications of just getting that far are world-shaking: machines could operate on cells, build processors and electronic components with molecule-wide feature sizes. The real hope of the method, though, is to create tiny, multipurpose machine shops (i.e. nanofactories), that are capable of building copies of themselves. This is where the "grey goo", hollywood-nightmare version of nanotechnology gets its inspiration (it's also stupidly easy to prevent by proper nanofactory design). Nanofactories capable of self-replication would constitute a physical instantiation of a kinematic self-replicating machine, which has so far just been a neat thought experiment.

The experiments Wang describes are to explore the viability of the essential reactions of diamondoid mechanosynthesis. These have been the source of a great controversy and intense debate, particularly between Drexler and the late Dr. Richard Smalley.  I admit a certain bias in this debate, mostly because I found Smalley's tone to be positively trollish (personal attacks, appeals to emotion, flatly-inappropriate metaphors).  Nonetheless, his principal argument wasn't without merit, namely that we've never seen these sorts of reactions done by anything but living cells using wet chemistry and aqueous enzymes.  That's part of what these experiments will be trying to put to rest.  We already have dry enzymes for certain reactions, and we've done single-atom manipulations with scanning probe microscopes (SPM), both of which Smalley argued were impossibilities in the course of his debate with Drexler.

Posted by: leoncaruthers at 08:27 PM | No Comments | Add Comment
Post contains 341 words, total size 3 kb.

Jatropha, sometimes referred to as the diesel tree, there's a fair amount of investment going into improving the yields.  You get the oil by way of seed processing.

Strangely, another species, copaifera_langsdorfii also shares the "diesel tree" moniker in many articles, but there are some big differences. The coolest being that you can get diesel fuel basically the same way you'd get sap from a maple tree: just tap it.

What sort of scares me about these trees is the wildfire danger.  Anyone who's ever had to put out a grease fire knows how bad they can be.  Imagine a plantation of grease fires the size of an apple orchard.  Considering that both species seem to favor tropical temperatures and aren't particular on soil quantity, I'm guessing there's a good chance we'll see a large farm of them on almost-desert soil go up in flames at some point.  It might only take one bad fire to blunt the interest in the crops.

Posted by: leoncaruthers at 07:43 PM | No Comments | Add Comment
Post contains 165 words, total size 1 kb.

As any computer scientist can tell you, evolution is awesome. Properly implemented as an algorithm, it can design aircraft, ships, trusses, antennas, just about anything that can be described with a "chromosome". So long as you have some notion of what a "good" design is, and a way to test for it, the genetic algorithm can usually make substantial progress in finding one, through many generations of reproduction, crossover, and mutation.

As an aside, genetic programming is an idealized version of the way we think biological evolution works. It's probably pretty close, and it can provide valuable insight into the evolution of species, but it's not the same thing. On more than one occasion, I've seen people point to GP as "proof" of biological evolution. It isn't, any more than a flight simulator is proof that airplanes can fly, and claiming it as such is a disservice to the work of archaeologists and evolutionary biologists.

Anyhow, in computer science terms, GP is a method of search. Though better than many methods at escaping local minima, it isn't perfect.  In a program, this could mean a set of populations that never produces a very good design, even after many generations.  There are parallels in biology.  Consider that the efficiency of plant photosynthesis tops out at just over 6%, and most plants are lower than that.  That amount of utilization is enough to power all the plant life on the planet, and it hasn't changed much in a long time.  Current solar panels already have that beat (10% to 20%). A plant species that could utilize solar energy at those rates would have a strong advantage over other plants in it's niche, and possibly beyond it. Whether the lower efficiency of plants it a fundamental limitation of DNA-based biology or simply "good enough" isn't the point; plants aren't as good at as we're becoming.

Humanity, near as we can tell, has a unique approach to competition among species. Intelligence, situational memory, and planning are only part of that approach, the merely physical and biological strengths that propelled us to early success. Symbolic communication -- facilitated by our intelligence -- led us to culture, to memories that survive individuals, and thus to complex technologies. With them, we became more than just hairless apes with cool hands and bad backs. We became, quite simply, better at evolving than any other complex animal.

Technology lets mankind be nearly mythological in power. We can communicate with each other across the globe. We have durable "bodies" that we can inhabit as needed that are stronger and faster than any animal that's ever lived, and allow us to explore any environment, even places where no other life may tread, largely with impunity. We've even created machines to do the boring parts of the very cognition that's made us so successful, so that we can do even more of it. As our powers have grown, so have our appetites. All the miracles we employ require energy, far more than our tiny bodies could possibly metabolize. So we built engines and motors, artificial metabolic organs capable of ingesting otherwise indigestible "foods" with incredible energy density, first coal and petroleum, later uranium and thorium.

All the "foods" we've found to feed our miracles, however, have limits and costs. They aren't depleted, by any means, but they aren't forever, and we can foresee a time when they may become so scarce as to be effectively gone. One possible, long term way out of this (barring some lucky breakthrough in fusion) is to become plants.

I don't mean that we should turn ourselves green (physically or metaphorically). We became what we are by being better than any other animal at literally everything they do, using technology to do so. To continue on our path, we have to find ways to be better than any plant at everything they do. Solar cells will be our leaves, spread out to catch the radiance of Sol. Batteries or other storage will be our sugars and fats.  Will it be enough power?  For a while, surely, and it will get us further along the way to whatever comes next.

I'd much, much prefer fusion, by the way.  We should still beat plants at their own game along the way, but fusion is nearly a necessity if we ever want to leave the solar system.

Posted by: leoncaruthers at 10:12 PM | No Comments | Add Comment
Post contains 729 words, total size 5 kb.