The first AI that should get money is not AI
The first and most important "hardware" we need right now does not come from nVidia.
AI, e/acc, transhumanism, longtermism... Never mind them. There is only one reality, the one in which we are physical flesh that needs physical food, coming from physical matter from physical land. But as things stand now, unless we deploy the "AgI" that is Agriculture Improvement soon and fairly, sooner or later all we humans alive today will endure fairly substantial food supply shocks, none of which could matter less for hypothetical quadrillions of future superhumans. Agriculture Improvement is the first AgI that must become real, avoiding certain traps, by using the right, already existing technology and social innovation.
Agriculture needs help
Farmers urgently need help against poverty, from the European ones now riding tractor protests across Europe, to the ones going for suicide from the US to India.
Agriculture urgently needs help against globalization that makes food stupidly jump all around the planet before it reaches any store, or starves people with wars or troubles... happening half the planet away. Like the Ghanaians experiencing a food crisis because of a coup in Niger, or all other Africans suffering grain shortages because of war in Ukraine, or India export restrictions on onions, rice, wheat and sugar "wreaking havoc globally".
Agriculture needs lots of help against procustean, one-size-fits-all solutions and attitudes at all levels, from individual diets to land usage policies. Going vegetarian, not to mention vegan, is good but hardly a worldwide solution on a planet that has already passed ‘Peak Agricultural Land". Getting even more vegetables out of that land would require even more intensive farming, and much of the remaining land could feed meaningful number of humans only indirectly, through animals. In Australia, for example (but the same is true in many parts of Italy, and surely in other countries too) cattle or sheep grazing on rangelands is the only way humans can get substantial nutrients from 70% of the land.
And of course, agriculture also needs help to cope with climate change: after a century spent optimizing crops to not bear unreliable heat and water, we have situations like the one in UK, where wheat yields have stagnated just because temperatures increased a degree and rainfall declined by nearly a third.
Finally, agriculture also needs to be liberated, on one side by elite overproduction that makes young people loss skills and interest in it; on the other, by dreams, both rural and urban ones, about how agriculture should actually feed 8 or 9 billion people in the world and societies that actually exist now.
Every human being has the same dignity and right to live as you or your favorite pop or tech star. But expecting everyone of us, the actual human beings alive today, to go back to self production in Hobbiton-like rural paradises is not a strategy. It is just what will be naturally left after most of those people will have starved (and even the prepping survivors wouldn't last much more anyway).
For a thousand unescapable reasons, most people alive today just cannot "grow their own food", and the only places and ways in which they can live in is cities that, even if they were rebuilt as they should before 2070, may have too much soil contaminated by illegal dumping, paint chips, and leaded gasoline to be self-sufficient. Feeding those masses, which very likely include every reader of this post, is always going to need some "mass" form of agriculture outside cities, as this journalist personally discovered. Or, as I recently read on Twitter, "organize locally... sounds nice n all but ‘tis impossible" in the world that will actually exist at least until everybody living today is around.
Agriculture also needs help, with social and environmental conditions changing so rapidly, to innovate meaningfully, without being trapped by top-down monopolists on one side and uninformed nostalgia on the other. Ancient wheat cultivars, for example, were more beautiful than current ones because they were tall. Fact is, their successors were deliberately shortened to protect them from wheat lodging). And while the John Deere's of the world should stop "selling" the wrong kind of tractors now, it's machines like those tractors that let millions of farmers get an education by freeing them from "spending 30 hours to harvest every quintal of wheat".
Solutions?
Solutions to these huge problems must be both individual and collective,that is: political. At the personal level, the first thing to do is to get rid of silly ideas and avoid falling in the traps outlined above. "Returns to land" are possible only for a minority of people, and green technology, as great as it is, remains either useless or pointless without political and economic change.
Second, let's all eat better: that is, as most scientists would surely agree, enable everybody to "Eat food. Not much. Mostly plants", instead of wasting money on Ozempic. Making of us current adults the LAST generation to eat so stupidly as we do would go a long way towards making agriculture face the challenges of our time.
System-wise, the first thing to go for is probably to free agriculture from monopolies, both physical (land-grabbing) and immaterial, namely "intellectual property" that is everything but innovation and really free markets. Personally, I don't know if improving the soil is always "just colonization, with a new label". But big business in agriculture fails almost by definition to address issues like land-access, local community health and food justice, by creating thousands of "nutty demands" worldwide like that of scarce water for almonds in California.
Even cutting ruminants who are main producers of natural soils can make places "more dependent on Big Ag". It's essential, instead, to demand and enable in every place the policies, land uses and diets that are really feasible and effective for that place, without starving others. If quinoa or fonio are good but just cannot grow in Europe or North America without further climate change OK, let's import them... after making really sure that it doesn't create more harm than good to their first consumers.
Come to think about it, the more one is obsessed with race, replacement and "foreign invaders", the more money he should spend NOW to make poorer countries food-independent, shouldn't he?
At the same time, while it is crucial to free small farmers from "free markets" that exploit both them and their final customers, it is equally crucial also to restore their social standing. A really advanced, really civilized society is one where there are enough local, young farmers, who stayed in the fields or went back to them because they actually can and want to do it, maybe after getting some degree for the right reasons, not to "get a career".
Finally... yes, technology!
There are a lot of ways in which appropriate, properly used technology can and should improve agriculture, the sooner the better, at all levels. Back in 2016, I co-authored a paper just about “Digital DIY for self-sustainability of rural areas”.
Today, stuff like edible microchips inside parmesan cheese wheels seem more another source of e-waste than a sensible food certification solution to me. But digital platforms can help farmers worldwide to share useful information directly. Research and analyses can do a lot of good in lots of ways, from development of dry-farming to reduction of pesticides that eventually just produce pesticide-resistant pests, to mainstreaming (again: where and when it DOES make sense) really efficient new foods, from crickets to pythons and rice big in protein, but small in carbon footprint.
Solar farms can be built in ways that provide important ecosystem services. Data gathering by drones may greatly enhance field trial research that would take forever for humans to do. And "artificial intelligence" can help all such research a lot, as long as it stops being just an actual clone of today's Tech Bros.
All this tech could really innovate agriculture as it sorely needs, if it's done and deployed correctly. Even in cities, why not? Do we have thousands of empty office buildings that would cost too much to convert to homes, or not? If yes, why don't convert their insides (via hydroponics), or even just their walls to vegetable gardens? Why not mow city public parks with sheep and goats that would also provide meat and cheese, as a Major of Rome actually suggested a few years ago?
Future Agriculture must be more technological but also much less flashy and much more concrete than Big Business paints it. That is, it must be much more like Star Wars, just without child slavery, or general life conditions that make one want to run in space. The way to go is return to what even John Deere itself used to do, that is companies actively cooperating with farmers, instead of locking them in.
The most innovative tractors these days, for example, are the old, really "ownable" ones. Farmers can create virtual fences for their cattle, and "move" them every day from their porch, to preserve pastures and even make cheese taste better, with very simple, solar-powered GPS collars. Even robots that can weed fields 48 times faster than a human and make farmers spend much less on pesticides are really good when they can be repaired, if not directly manufactured, by their actual end users and owners.
Conclusion: first AIs first
The problem, with agriculture and everything else, is not technological innovation, that is new machines or new plants. It's who controls it, and controlling it only for short-term profit. Real, but really open and distributed innovation, Skunk Works style, may be more needed in agriculture than in any other field. Artificial Intelligence is great, but THIS AI is more important. Besides, unlike that AI, almost everything you just read is technology that is much cheaper, already exists and could save, rather than deplete, land and above all water, and water, and water. Please think about it.