Draft dragons - a question

Graydon wrote:

cralki wrote:[did Shan-wei's crews alter dragons?]
They'd have to figure out the native DNA-equivalent to be able to engineer anything at all, including the moss.

And their tech may very well have been able to reliably predict what an alteration would do without need for trial and error.

Figuring out the native DNA -- how it's coded and transcribed -- doesn't get you anything about what it means, though,

Oh, sure -- it's much much easier to e.g. sequence a genome than to actually figure out what each part does.

I can see people with experience of several different biospheres (and thus understanding the differences/similarities and knowing what to look for) and essentially unlimited computing power being able to work that stuff out astonishingly quickly, though.

EDIT: Anyway, they must have figured it out given that they altered the Fleming moss. (At least on Earth, plants aren't really that much less complex than animals; the really big jumps are prokaryote/eukaryote and unicellular/multicellular.)

It's very unlikely they can skip the trial-and-error step entirely, a lot of questions of expression are environmental, DNA doesn't and there are strong reasons to expect the Safeholdian equivalent doesn't, work as prescriptive rules, it's going to be a bunch of constraints on responses to the environment. It's pretty much impossible to simulate that, you have to do the actual experiments. (Rather like you stop being able to predict billiard ball bounces after low integer numbers of bounces; too much lurking chaos in the real system.)

It's true that a lot of development is not exactly and strictly controlled by DNA, but I'm not sure that's counter to what I'm talking about -- I'm not talking about dramatic anatomical changes but tweaking gut flora and maybe digestive enzymes, stuff that might actually be fairly simple.

Silver fox results go by generations; they were at generation twenty to get ~1/3 actually domestic (or at least, dog-equivalent) behaviour. That would take a long time with dragons! When you're starting with a creature you don't know very much about simply because you haven't been observing it for very long.

Yeah, sorry, I wasn't being very clear. What I was talking about was how they seemed to get a lot of "dog-like/domestic" traits that weren't specifically being selected for... which seems to me to suggest that tweaking a few 'control' genes might get you dramatic domestication-like results. (IIRC neoteny - retention of immature traits into the mature animal - is thought to be involved?)

cralki wrote:Well, consider trees; there are maple trees, various fruit trees, etc. In a hundred years, you haven't even gone through one generation for some tree species. Certainly a terran-style forest hasn't undergone one complete successional cycle. You don't really know these things are going to reproduce OK locally, that you got everything right so there isn't going to be an outbreak of insect pests or a collapse in the apple population due to a mis-timing with the pollinator life cycle. (Or that the horribly flammable native understory persists as roots, and somewhere around year twenty, foom!)

This is way beyond what we could do; not only are ecologies massive collections of moving parts, the parts all modify themselves and each other.

Sure, but I think a lot of it might have been a bit more brute force than that. Take an island or region, kill everything native down to bacterial spores in the soil, introduce Terran species. It seems to be either mostly-Terran with a few hardy Safeholdian species or mostly-Safeholdian (in unterraformed land), not really a single ecosystem created from both (which really would be insanely difficult).

Also, it might not have been anywhere near perfect. Humans are really adaptable even at low tech - more forest fires and stuff wouldn't have been fatal to humanity on Safehold. It might not have been nearly as stable at the beginning as it is at the time of the books, almost a millennium later.

(Also they might have been really really good at ecological modeling.)

cralki wrote:*Biotech is IMO mostly being kept from taking off the way computer-tech did because it is so hard to market altered crops etc. [snip]

Well, there are three big problems with biotech.

Politically, it gets used as a way to establish control and extract rents forever; farmers are understandably less than pleased about that.

Yeah, sure, but that's tied to the stupid patent laws I mentioned. Throw out those and the problem disappears. That's actually an argument on my side

It gets used to do really stupid things, often in context of trying to extract those rents; "Roundup-Ready" lasts for about twenty years before the genes start leaking into the weed population, at which point the business model collapses in a welter of bad outcomes. (Much tougher, including physically tougher, weeds are not a net win! Neither is having dumped herbicide everywhere for a couple decades.)

Sure, and this is unwise and problematic. I think the advantages of it used well would greatly outweigh these disadvantages, though. The potentials are just colossal (fixing much of third world malnutrition; multiplying farm crops enormously and thus minimizing land used for farming, putting tons of land 'back to nature'; making medically etc. useful chemicals; etc.)

Biology means things change themselves. This is much worse than computer code, where at least if something in the code is horrible someone put it there. Biology, not so much, and you're not specifying outcomes, you're constraining developmental processes that interact with the organism's whole environment in often surprising ways and certainly in very complex ways. There's no way you fully understand how that works when you make the genetic changes.

Sure, but so what? Nature is a giant cauldron of evolutionary change happening everywhere all the time. This isn't like climate/CO2 where we can genuinely tip the balance by introducing something 'from outside the cycle'. The purported dangers just aren't really an issue, and certainly aren't worth crippling the technology by regulation the way it currently is.

cralkhi wrote:I can see people with experience of several different biospheres (and thus understanding the differences/similarities and knowing what to look for) and essentially unlimited computing power being able to work that stuff out astonishingly quickly, though.

EDIT: Anyway, they must have figured it out given that they altered the Fleming moss. (At least on Earth, plants aren't really that much less complex than animals; the really big jumps are prokaryote/eukaryote and unicellular/multicellular.)

True, though plants are more forgiving; splicing a firefly protein into plants to make them glow is easier than doing the same thing to pigs.

Though I suppose the fleet would have had a great deal of AI with it, and been packed with experts. (Which raises some dark ethical questions about what happened to them.)

cralki wrote:It's true that a lot of development is not exactly and strictly controlled by DNA, but I'm not sure that's counter to what I'm talking about -- I'm not talking about dramatic anatomical changes but tweaking gut flora and maybe digestive enzymes, stuff that might actually be fairly simple.

Whereas I'm thinking of "this thing smells like nothing on my planet and I'm supposed to admit it to the top of my social hierarchy like a good domesticate?"

cralki wrote:

Graydon wrote:Silver fox results go by generations; they were at generation twenty to get ~1/3 actually domestic (or at least, dog-equivalent) behaviour. That would take a long time with dragons! When you're starting with a creature you don't know very much about simply because you haven't been observing it for very long.

Yeah, sorry, I wasn't being very clear. What I was talking about was how they seemed to get a lot of "dog-like/domestic" traits that weren't specifically being selected for... which seems to me to suggest that tweaking a few 'control' genes might get you dramatic domestication-like results. (IIRC neoteny - retention of immature traits into the mature animal - is thought to be involved?)

They did, but foxes are canids, and one variety of canids practically domesticated themselves.

All critters aren't that easy; people have been trying to domesticate zebra for the last couple hundred years, and it doesn't work. Some taming successes, but a domestic strain isn't happening. Cheetah are incredibly tamable but until the last 20 years or so no one knew how to get them to breed in captivity, so they couldn't be considered a proper domesticate. Despite our ancestor's success with aurochs, buffalo -- North American bison -- aren't especially practical as a domestic species. Horses, which have enormous genetic variety -- there are breeds of horses with different vertebra counts! -- also appear to have one and only one Y chromosome. The implication is that all those indo-european legends about how Our Culture Hero tamed the Father of Horses have a kernel of truth; there may have been one and only one domesticable stallion. Elephants are a borderline case; hippos are Right Out.

The odds of getting a naturally domesticable Safeholdian anything seem low; maybe they lucked out utterly on dragons, or maybe dragons are a native species that Shan-wei's team gave the brains of St. Bernards.

cralki wrote:Sure, but I think a lot of it might have been a bit more brute force than that. Take an island or region, kill everything native down to bacterial spores in the soil, introduce Terran species. It seems to be either mostly-Terran with a few hardy Safeholdian species or mostly-Safeholdian (in unterraformed land), not really a single ecosystem created from both (which really would be insanely difficult).

Initially, I would certainly expect that.

The implication of the book of the Writ that's about "preparing unconsecrated land" -- and what is the theological justification for there being unconsecrated land? the War of the Angels interrupted things? -- is that you can do it with muscle power, though, which implies that it's surprisingly easy to do.

cralki wrote:(Also they might have been really really good at ecological modeling.)

I think we have to suppose they are spectacularly good at it.

cralki wrote:[genetically engineering crops]
Sure, but so what? Nature is a giant cauldron of evolutionary change happening everywhere all the time. This isn't like climate/CO2 where we can genuinely tip the balance by introducing something 'from outside the cycle'. The purported dangers just aren't really an issue, and certainly aren't worth crippling the technology by regulation the way it currently is.

It's not hard to imagine a case where cross-pollination from the engineered crop wipes out the non-engineered crop; this is a big deal in some of the Hawaiian fruit crops, where about half have to be engineered to resist the problematic fungus -- grown in blocks -- but the main market intensely doesn't want the engineered version. (That's one of the really good examples; open development of the modification, it's all public and published. Total crop value around 20 MUSD, which may explain how that happened.)

It's similarly not hard to imagine that someone could go Too Far and produce an interesting food crop that was horribly invasive; someone deciding to make kudzu tasty and able to tolerate a broader range of climate zones would be something of a mistake.

And of course engineered microbes, which is where a lot of the work has been done, are much more unnerving than anything on the macro-scale.

I think the potential benefits are enormous but I wouldn't want to get rid of regulation and I definitely want a publish-and-replicate requirement. No secrets, and at least some check on optimism.

cralkhi wrote:Oh, sure -- it's much much easier to e.g. sequence a genome than to actually figure out what each part does.

I can see people with experience of several different biospheres (and thus understanding the differences/similarities and knowing what to look for) and essentially unlimited computing power being able to work that stuff out astonishingly quickly, though.

EDIT: Anyway, they must have figured it out given that they altered the Fleming moss. (At least on Earth, plants aren't really that much less complex than animals; the really big jumps are prokaryote/eukaryote and unicellular/multicellular.)

This assumes that it was engineered at all - it could have just been stumbled upon, just like penicillin was, as a contaminant in another experiment looking for potential disease organisms

Silver fox results go by generations; they were at generation twenty to get ~1/3 actually domestic (or at least, dog-equivalent) behaviour. That would take a long time with dragons! When you're starting with a creature you don't know very much about simply because you haven't been observing it for very long.

Again pure dumb luck could be involved - kid with odd pet perhaps. Foxes aren't the first choice for pets to most people but natural selection can be very powerful and if the dragons breed at a couple of years like most wild animals then a century should give a make or break call if the dragons were useful as herd stock (hides perhaps, even if they can't be eaten).

cralki wrote:Well, consider trees; there are maple trees, various fruit trees, etc. In a hundred years, you haven't even gone through one generation for some tree species.

With fruit trees a generation is the time to produce viable fruit not the life of the tree. This can be as little as two years. So in a thousand years I can see breeds of food plants appearing utterly unlike anything on earth. (Most of the fruits we eat now are breeds less than a century old and in many cases would be unfamiliar to our great grandparents).

For timber trees the issue is identifying useful traits for breeding like rapid growth and good quality timber - which may take longer. As for fires - parts of Australia and North America get regular bush fires but the plants are able to survive without difficulty (and some require fires to germinate their seed). The problems happen with too frequent fires before the trees set seed or too infrequent fires which prevent seeding growth.

Graydon wrote:Whereas I'm thinking of "this thing smells like nothing on my planet and I'm supposed to admit it to the top of my social hierarchy like a good domesticate?"

The bacteria/enzymes thing was about dragons being practical to feed, not their ability to be domesticated in the first place.

All critters aren't that easy; people have been trying to domesticate zebra for the last couple hundred years, and it doesn't work. Some taming successes, but a domestic strain isn't happening.

I'd heard "zebras can't be domesticated" before, but never read anything about any serious attempts to do so (taming attempts, yes, but not a serious multi-generational breeding program). Do you have any links or book references?

Despite our ancestor's success with aurochs, buffalo -- North American bison -- aren't especially practical as a domestic species.

Do we know that? The Native American cultures that lived around them the most weren't sedentary (since without steel plows you couldn't practically farm the Great Plains - the prairie grass roots are too difficult), so I'm not sure they ever tried.

And when Europeans met bison, they already had cows.

And there are bison ranches. I think if there wasn't an interest in keeping bison more or less as-they-are for conservation purposes, they might well end up domesticated.

Bison are aggressive, but so were aurochs, so I'm not at all convinced that bison are in fact non-domesticatable.

There was a Soviet experiment in domesticating moose, but I don't know how successful it was (the farm - Kostroma Moose Farm - is still in existence and produces moose milk, but I don't know how much the farm moose's behavior differs from wild moose.)

I think it's at least as likely that there was just not much point in investing the time/effort in domesticating other things once we had cows, horses, pigs etc. developed.

The implication of the book of the Writ that's about "preparing unconsecrated land" -- and what is the theological justification for there being unconsecrated land? the War of the Angels interrupted things? -- is that you can do it with muscle power, though, which implies that it's surprisingly easy to do.

It is supposed to be labor intensive though. My guess?

-Scrape off the topsoil (this is the hard part);
- Replace it with fallen leaves, crop waste (wheat/rice stalks/leaves etc.), manure, maybe kitchen scraps etc;
- Add a bit of "terraformed" soil to introduce the right soil biota;
- Wait a couple of years to compost the waste.

cralki wrote:[genetically engineering crops]
Sure, but so what? Nature is a giant cauldron of evolutionary change happening everywhere all the time. This isn't like climate/CO2 where we can genuinely tip the balance by introducing something 'from outside the cycle'. The purported dangers just aren't really an issue, and certainly aren't worth crippling the technology by regulation the way it currently is.

It's not hard to imagine a case where cross-pollination from the engineered crop wipes out the non-engineered crop;

Even if the engineered form has a strong selective advantage in wild conditions, this doesn't really qualify as "wiping it out"; you're replacing a couple of genes, not an entire species.

It's similarly not hard to imagine that someone could go Too Far and produce an interesting food crop that was horribly invasive; someone deciding to make kudzu tasty and able to tolerate a broader range of climate zones would be something of a mistake.

I believe creating something more invasive than current invasive plants is well beyond current tech.

If it were that easy to produce something that was "better at everything", we'd see a lot less diversity in nature. Specialists generally do better than generalists in their own niches.

The processes of genetic engineering are not fundamentally different from what exists in nature. (Even "transgenic" species happen naturally, but the process is called "horizontal gene transfer". There is a sea slug - Elysia chlorotica - with algae genes. Some bacteria can incorporate loose DNA from the environment. Indeed, a vaguely similar event seems to be the origin of all eukaryotes, where bacteria became incorporated as mitochondria into the proto-eukaryote cell. Our mitochondrial DNA is not at all "related" to our cell nucleus DNA. Chloroplasts work the same way.)

And of course engineered microbes, which is where a lot of the work has been done, are much more unnerving than anything on the macro-scale.

Work with potentially pathogenic microbes would need serious regulation, yes. (Not all microbes, by a long shot, fall into that category).

I think the potential benefits are enormous but I wouldn't want to get rid of regulation and I definitely want a publish-and-replicate requirement. No secrets, and at least some check on optimism.

I currently think the regulations, by creating a high barrier to entry, are actually benefiting the "bad" users of the tech like Monsanto versus more beneficial and generally less funded uses, so I would rather see it gone (for crops, not pathogens!)

But I may be a bit unusual on this sort of issue - it's my firm opinion that FDA regulation of medicines kills or ruins the lives of far more people by delaying the availability of life- or health-saving treatments than would be harmed by "bad" medicines even in an utterly unregulated environment (which I actually think wouldn't be a very common problem - doctors still would know what they were doing, and medical information is wildly more available to regular people now than when the FDA got started).