"Feed the corn to mice before you fed it to humans, sure.  But then, besides asking 'What precautions should I take?', one should perhaps first ask, 'What exactly could go wrong in the first place?'  What could potentially go wrong with trying to create a new strain of corn?  How could there be a disaster, not just a minor stumble, from trying to create a new strain of corn?"

Otolmens wishes she had not been REMINDED of that.  Those were not GOOD TIMES.

This is a question Chelish wizards are spectacularly good at answering. 

"It happens to be really good for a certain kind of pest, and they grow to ten times their usual size and eat everyone in the village."

"It smells irresistible to dragons."

"It angers the fae."

"It's so much more fertile than all other corn that it gets carried away on the wind and grows everywhere, blotting out all other life, until nothing grows anywhere on the continent but corn."

"It's great for a couple years but it's sucking all the vitality out of the soil and leaves only sand behind."

"It lures maneating rats from the Underdark and then the infestation is impossible to root out."

"It grows six hundred feet in height and angers the aerial dragons."

"Locusts that lay their eggs in it have an unnaturally high survival rate and so instead of occasional clouds of locusts we have constant clouds of locusts and they blot out the sun."

"It's addictive and once you've eaten it you can't eat anything else."

"It disrupts the flow of magical energies through the land beneath its roots and remaps all the ley lines in Cheliax, which causes a bunch of adjustment hurricanes and strands half the towns on the royal line."

"It develops impossible geometry - the kind where looking at it gives you a headache - and anyone who wanders into the field come harvest time is lost forever."

"It requires so much water that it sucks up water for hundreds of miles around, turning half of central Cheliax to desert."

"It's actually just mediocre corn but with mind-control to make you think it's really great corn, and we're convinced we succeeded and plant it everywhere at which point it's powerful enough to enslave the whole country."

The little mortals really have NO idea, do they.

"Does that sort of thing happen a lot around here -"

" - I mean, those would be unusually bad outcomes."

"Usually interactions between the natural world and surrounding magical geography are fairly bounded, there's only a couple documented incidents of ley-lines moving because of ecological changes alone."

"Something going horribly wrong with pests is pretty likely but that's what adventurers make a living handling."

"Plants are the category of living thing least likely to spontaneously develop spell-like abilities or sapience, there are only about a hundred known kinds that have."

"No one has the slightest idea what angers the fae, planting a new corn crop might but not planting a new corn crop might too."

"And if the dragons are mad they'll probably tell us. - burn some villages first, but then tell us."

"I am suddenly concerned about whether I have accidentally fallen into the trap of thinking I am a story protagonist, instead of applying the principle of mediocrity.  When I landed near the Worldwound, was I otherwise-improbably being placed near the most important present disaster in the world, or actually is most of Golarion like that outside the protected housing and the Worldwound was just number twelve on the list of worst disasters from the previous week?"

Keltham's first-order uncorrected intuitive probability he'll want to have kids here has now fallen to 65%, not so much from this update about Golarion's nature, per se, as what it implies about a predictable string of first-order updates that have all been in the same direction so far.

"...I think the Worldwound is a reasonable candidate for the worst problem in the world," Meritxell says. "There are a few other planar rifts but they're much smaller. Cheliax has most of our military deployed at the Worldwound, we wouldn't do that if there were ten things as likely as it to destroy the world. But there are a lot of places that are horrible on a scale that won't destroy the world."

"Well, at least that's a slightly pleasant surprise about how nice a place Golarion is to live, compared to where I'd just sent my second-order estimate."

Keltham genuinely is relieved about this; to be repeatedly surprised by the same observation indicates that the machinery making up yourself is not properly reflecting the idea of 'surprise'.  A stable meta-rule 'no matter how bad you imagine Golarion is, it is actually worse' implies some defect in you and not just the planet.

A lot of people, it has occurred to Carissa, think Cheliax is the worst problem in the entire world. That's because they're dumb and get really worked up about a little bit of torture, though.

"So if this was dath ilan, I'd know how probable or in a sense typical the disasters you named actually were, but I have very little sense of that here.  Being able to imagine disasters at all is step one.  Being able to refine your sense of which disasters are actually likely to hit you is step two.  Then, usually, after that, you need a step three where people realize that even if they don't like a disaster and start hurling insults at it about it being super incredibly improbable, there's a difference between the kind of disaster that ignores insults like that and gets you anyways, versus the kind of disaster that really actually goes against the character of reality and almost certainly won't happen.  I cannot guide you through developing that sense for probabilities using realistic examples, because, I am realizing, I haven't a butt's notion of where any of what you said falls on that spectrum."

"But if you have a sense of something like - what is a typical disaster for Golarion that might happen to you personally while breeding a new crop; versus things that happen often enough that you hear about them, but rarely enough that they haven't happened to you or anyone you know; versus possibilities that can't really get at you?  From a selfish perspective, people are mainly incentivized to guard against common disasters that hurt themselves; if you look at it from the perspective of Chelish Governance, it's their job to make sure anyone who's allowed to experiment inside their country needs to be guarding against country-injuring rare disasters; and both of these agents will be falling down on their jobs if they spend all their resources guarding against imagination-capturing disasters that are genuinely out of character for Reality and not just being insulted by being called names like 'impossible'.  So - common-level disasters, rare-level disasters?"

"Ruins the soil is common," Tonia says. "Attracts new pests that grow to unusual size is common. Grows too well and takes over your other fields too is common. All of those've happened to my father or my grandfather. Angers the fae is - happens to someone's cousin in another village."

"The usual rule is that as you go up the tech ladder, the danger levels go up because the power levels go up, meaning that smaller missteps can have larger effects.  I do not get the impression that this is the tiniest bit untrue of magic and Golarion - but, just to check -?"

"Seems right. The Worldwound was caused by a fight among gods. Most other really bad things I can think of were caused by epic wizards."

"As you get more clever at breeding plants, you can, to some degree, even in dath ilan without magic, manage to do a bit more damage than if you were less clever.  If you successfully use focused breeding to create corn that is incredibly resistant to the most common diseases around, more disease-resistant than any corn has ever been before, it may be more likely than regular corn to take over all your other fields by accident."

"If you breed ultra-fast-growing plants and plant them repeatedly on the same land, year after year, they will suck key nutrients out of the soil, unless you figure out what those nutrients are and take extra steps to replenish them.  Even if you figure out how to provide the plants with the nutrients that they need to keep growing, the fast-grown plants may end up less nutritious for people, or animals, who eat those plants - unless you replenish aspects of the soil that aren't as obvious.  Subtle deficiencies that people may not notice at first, especially if people are eating the older crops too, for a while.  If you figure out how to fix the short-term problem of your fast-growing crops dying, by replenishing the aspects of soil that just the plants need, you may not replenish enough of key tiny nutrients that people need.  Praying for guidance sure would be helpful for that sort of thing, if it worked perfectly reliably.  But even leaving aside the point that apparently 'prophecy is broken' now, it seems wiser for you to try to develop the skill that we needed in godless dath ilan - like, at least write down in advance what you predict the guidance will be, before you pray for guidance."

"You rotate crops," says Tonia. "And maybe...take soil samples, to see how other older crops grow in them?"

"Rotating crops helps some, doesn't fix the whole problem because there's some things that almost every crop takes from the soil.  How much is it helping?  Feed your rotated fast-grown crops and non-rotated fast-grown crops and slowly grown crops to mice, or other animals that grow in even faster generations, and see how the three groups do health-wise relative to each other.  Keep your eyes open, don't wait for problems to materialize before you start looking, imagine things that might go wrong and look for them early, maybe you'll catch something you didn't imagine while you're looking for some possible problem you did imagine, the important thing is to keep your eyes open wider."

"Besides soil depletion if you figure out how to grow more crops faster, there's one other problem that's very predictable, that happens because of how well you succeeded at plant breeding, and it leads into another one of those larger points.  Suppose you produce a single field of corn, all of one strain, that's the best corn you've ever seen.  It grows fast, it's resistant to disease and insects and strong winds, it's tastier than the previous corn, you feed it to mice and the mice do fine.  You find - well, you probably don't have replicators - you find somebody reliable to verify your reports, and you take that corn strain and sell it to farmers all over Cheliax.  A year later, it's the most profitable corn anyone has ever seen, and only a complete fool of a farmer would grow anything else the next year.  Two years later, it's the only strain of corn that anyone still grows in Cheliax, and it's starting to displace other crops that are less profitable to grow.  All massively replicated out of that one original field."

"What happens next?"

"If there's a blight it'll take out half our food crop," says Tonia. "Because everything will have the exact same vulnerability."

"Not if.  There will be a blight.  It's not just that there's a corn blight and Cheliax is growing too much corn.  It's that blight itself is a form of life and it reproduces and blight that targets this exact strain of corn will reproduce faster on this exact strain of corn and then reproduce just as fast when it jumps to the next ear of corn that all came through the same bottleneck and all has the same genetic information inside it.  Everything still alive has an internal system that counterattacks and resists disease, and since everything has slightly different tiny spirals, all the internal systems use slightly different counterattacks and methods of resistance.  If all the ears of corn are too similar to each other, if you copied too much of the tiny spirals too fast and made too many organisms out of them, because they were such great tiny spirals and such great organisms, the disease that's mutating and reproducing and targeting those exact disease-fighting systems will get too good at targeting those exact disease-fighting methods and wipe them all out."

"The same thing would happen if we produced a kid with INT 20 and there was magic for copying kids and somebody got the bright idea that Cheliax needed a million kids like that.  There's a million tiny variants of even minor diseases, one of those variants would happen to be really strong against that exact form for a disease-fighting system, and then instead of the disease just killing that one kid, he'd sneeze and that variant would jump to the copy of the kid, and then the next copy, and the next, and that variant would be just as effective against all of them."

"That's one reason why dath ilan doesn't take the thousand brightest men in the whole world and try to have them each get ten thousand women pregnant.  It's copying too much of the heritage information too fast."

"....also sounds logistically difficult," someone mutters.

"Hey, if doing that sort of thing wouldn't kill everyone, jumping 2 Intelligence points in a generation would be worth a few logistical difficulties."

His audience, which is definitely interpreting 'a few logistical difficulties' as 'you'd have to have the men under an exceptionally powerful Dominate Person with a team of dedicated clerics healing them and keeping them under', nods seriously.

Keltham, who is definitely interpreting 'a few logistical difficulties' as some mix of 'men mastering the partial ejaculation technique' if they're doing it the fun way, and otherwise 'divide up the sperm quickly so it's still healthy during the mass insemination process', continues.

"One reason I'm giving you this caution, obviously, is so that, if you do start getting results from more directed and clever heritage-optimization, you don't push your luck; the disease counterattack is close to a universal hazard if you start deriving too many children from too few parents."

"The larger point is that variation, itself, is a kind of resource.  It doesn't just apply to variation of disease-fighting systems, although that sure is one of the clearest cases of it.  If you're tackling a difficult mental problem, and you've got five people on your team, adding a sixth person who thinks in a different way from the first five people is often a larger boost than adding somebody's previous acquaintance from a previous job who had a lot of similar life experience.  There are also benefits to people knowing each other, to be clear, but the longer you've hammered on a problem without solving it, the more likely it is that you need somebody new."

"The variation of your crops is a kind of resource that plant species has, making it more likely that at least some of it survives when it gets challenged with a new kind of weather, a new kind of pest.  When you apply powerful breeding pressures to a crop and squeeze it through narrow bottlenecks of parentage, you lessen that variation as a side effect, and make the crop probably less resilient in some dimensions, even if you're improving it in others.  Variation is a kind of resource for heritage optimization, and the process of heritage optimization uses some of it up."

"This ties into something that, in dath ilan, is seen as a central dichotomy of all life's existence - a dichotomy between -"  He needs to be careful in speaking here, so that the direct spell-supplied translation from the dath ilani terms into Taldane doesn't give away his point too early.  "- diversity, and optimality.  After all, if there's a best way to do things, wouldn't doing it any different way, necessarily be doing it worse?"

"Think about the logic I showed you earlier today, the one which can derive exactly all of the actual consequences of the premises and no non-consequences of the premises.  If you had a logic that was meaningfully different from that one, wouldn't it have to be, in some sense, worse?"

"I pose to you this question, then, which I have not yet told you how to answer:  Is diversity only ever valuable in places where we haven't found the best strategy?  Is there necessarily some optimal disease-fighting system an organism could have, which could fight off every different form of disease that exists, and if an organism had that, it could be duplicated a million times without worry?"

"it's like war. There's not a single best military strategy that defeats all other military strategies. There are things that work out best for a range of possible things your opponents might be doing, and you can't be engaged in the best possible tactic against anything they might be doing, there are tradeoffs. 'best disease-fighting system' sounds less ridiculous than 'best war-fighting system' but I think only because we know how to fight wars so the tradeoffs are obvious."

"Well, perhaps, perhaps.  But let's consider some much simpler case than complicated oppositional games.  Do you have locks, here, which go by knowledge?  Say, somebody has to punch in a series of numbers, or spell out a sequence of words, to open the lock?"