Peter Gill, Managing Director, Social:Farmers
In 1927 my grandfather received a business enquiry made on the behalf of Henry Ford:
‘Would you kindly be able to provide a quotation for the dismantling and packing for shipping of Fairbottom Bobs?’
Henry Ford was the most influential man in the world. He had rarely ever left America but had crossed the Atlantic to get to the little Lancashire village of Bardsley to make a personal inspection of the long-abandoned relic that was Fairbottom Bobs. It was one of four surviving Newcomen steam engines and, if all checked out to his satisfaction, he dearly wanted to acquire it for his museum.
For Ford the great significance of Thomas Newcomen’s invention of 1712 was simple. It was the world’s first practical engine; a starting gun, if you will, for the industrial revolution.
Less whimsically named than Fairbottom Bobs is the H100 Tensor Core GPU computer chip, the main muscle of the AI Revolution. Estimates say that since 2023 Jensen Huang’s Nvidia will by the close of 2024 have shipped 2.5 million H100 chips at around 25,000 pounds a pop.
The tectonic change effecting this sudden tsunamic flow of capital is the belief that the AI Revolution is set to rival or even eclipse the industrial revolution.
Agritech
Search for “agritech” and the search results will include precision agriculture, GPS guidance, variable rate nutrient application & drilling, robotic machines, sexed semen, see ‘n’ spray, genetic modification, carbon flows, remote sensing, bio-things, ‘sustainability’, data-driven decision-making and so on and so forth.
Two inherent threads have always linked most agritech innovations. One is that they cleave to Jonathan Swift’s ancient observation that growing two ears of corn where only one grew before is always a good thing for humankind. And secondly they usually involve the farmer in having to make a capital investment or concomitant extra annual financial input.
Yet, and supremely counterintuitive as it may seem, increasing farm production and lifting farm profitability clearly didn’t read the email about ideally being in lockstep.
Indubitable UK national success in achieving higher farm yields via decades of bringing in new agritechs has serially failed to translate into ascending farmers’ profits. In short, there is an agricultural productivity paradox.
So, if concentrating on higher yields is a flawed business goal then today’s question must be:- how can AI deliver on achieving greater farm profits?
Farming Co-Intelligence
Highlighting the productivity paradox is one of the best - if not the actual best - pieces of research into farm profitability entitled The Characteristics of High Performing Farms in the UK, conducted in 2018 by Andersons with the AHDB.
Using gold-standard DEFRA Farm Business Survey data farms were paired into two sample groups sizeable enough to give statistically meaningful results. Thus for each most profitable farm a similar ‘twin’ farm (not the farmer) was identified in the less profitable group.
This brilliant concept meant that, without any other muddying factors, the relationship between farm management practices and farm profitability lay exposed in sharp focus.
It’s a really fascinating read, and most important of all is the revelation that for achieving higher profitability all-round farm decision-making skills and the resultant lower overall costs are significantly much more important than the production of higher yields.
The report states that over 70% of the difference between top and bottom quartile farmers is because of different decisions made by the farmer; logically pointing to the best solution for raising farm profitability being an agritech that will improve farm decision-making.
AI is brilliant for decision-making. This can either be working solo or as a joint input ‘co-intelligence’ with the user. For strategic, technical, tactical, large or small decisions using AI will increase the likelihood of reaching more good farming decisions. Using free or almost-free AI every farmer can discreetly have the decision-making skills of top-quartile farmers at their personal disposal. So far, we have identified around 250 branches of farm decision-making for which using AI as Farm Co-Intelligence will either confirm or improve farmers’ own decision-making.
AI is the best agritech ever invented, and that is right here, right now. The plasma-hot AI developers’ race is bringing yet more. Generative AI just celebrated its second birthday party, but this kid is a Mozart. What AI can do is fast-tracking on myriad disparate knowledge frontiers and one frontier, pardon the jargon, is called ‘agentic AI’.
Agentic AI means being able to act as a user would while sat at their computer or on their phone: email reading and sending, filling forms, researching and ordering products, using apps, requesting quotations, chasing invoices, and chatting with colleagues about the weather or the rugby or Max Verstappen’s latest display of quite plainly reckless driving in the F1.
Thinking at the uppermost level, top-quartile farm decision-making and agentic AI mean a time is approaching when AI could be running a farm - decisions-wise - entirely on its tod.
So if AI isn’t already on your desktop or phone or both I would respectfully suggest not dawdling before making pals with AI. Put your AI into voice mode and just strike up a conversation on any subject of your pleasing - be as deeply technical or as frivolously light-hearted as you like.
‘Wow, just wow’ is a normal first reaction on experiencing this new altered reality, where the ‘person’ with whom we are having a totally normal conversation is but a supercluster of H100 computer chips. Truly it is amazing: why not give it a go right now?
Footnotes:
[a] Fairbottom Bobs successfully made it to America, was restored, and may be inspected at The Henry Ford Museum in Dearborn, Michigan.
[b] For the rest of his life my grandfather always drove a Ford.