An AI Lesson From the Farm: It's All About the Usage
2024.01.16
As the promise of artificial intelligence washes across the National Retail Federation (NRF) Big Show in New York as it did last week in Las Vegas at CES, we’re wise to keep our eyes not on the technologists and their star turns, but on potential users – and the dependencies of AI adoption.
Less Sam Altman, less Davos, less big stage keynotes in white-soled shoes.
More quiet decision-makers – lips pursed, in a cubicle or small meeting room – attempting to find a way to sustainably save a percentage here or add a percentage there.
Because the ultimate value of the technology will depend less on parameter counts and technical benchmarks, and much more on how, when, and where vertical industries will use AI to do more with less.
This blog will review and explore AI usage through the year. A good place to start is this thought-provoking essay on the economics of technology, published recently in The Economist.
A discussion, not algorithms nor datasets, but of the humble farm tractor.
Allow me, with full acknowledgement of The Economist’s authorship, to provide this summary:
The The Economist tractor, for an industry that in the first decades of the 20th century employed roughly one-third of American workers, was the ChatGPT of its day. It promised revolution, a “new epoch in farming,” and that farm work would be faster, cheaper, smarter, and much easier.
Looking back now, that promise has been fulfilled. But it was fulfilled gradually. Inch by inch. With, as the essay noted, “a whimper, not a bang.”
Per The Economist: starting around 1910, opportunistic business types “piled into the tractor-making business, hoping to make a quick buck (just as every second tech firm in Silicon Valley now describes itself as AI-first.”) But despite publicity splashed across the industry (rave reviews, for instance, in Prairie Farmer) adoption was slow. Many early tractor brands withered and died. By 1920, just 4% of American farms owned a tractor. By 1940 (and with a crushing Depression a raw and recent memory), only 23% did.
Why such slow adoption? It’s not that farmers were technology adverse, Luddites in overalls. Many tractor-less farmers owned automobiles. And other agricultural innovations were in broad use.
If the tractor was so good, why did farmers not buy them more quickly?
The essay pointed to three reasons – all with relevance to today’s AI boom:
1. Early versions of the technology were less useful than originally envisioned – and were revised and improved only through painful, in-field failures (Some early tractor models were simply too big or couldn’t maneuver through mud due to their metal wheels.)
2. Adoption required changes in labor markets, which took time. (For years it was less expensive to feed horses and hire human help than invest in an expensive machine with new, expensive and ongoing maintenance.)
3. Farms needed to transform themselves – go through a corporate restructuring. Tractors made far better financial sense on large farms than small.
In summary, and quoting directly:
The history of the tractor hints at how quickly generative AI may take over. At present most AI models still have metal wheels, not rubber tyres; they are insufficiently fast, powerful, or reliable to be used in commercial settings. Over the past two years real wages have hardly grown as inflation has jumped, limiting companies’ incentives to find alternatives to labour. And companies have not yet embraced the full-scale reorganization of their businesses, and in-house data, necessary to make the most of AI models.
No matter how good a new technology may be, society needs a long, long time to adjust.
To be continued.
I’m Jon Stine, 35+ years in business and technology.
I read, I write, I advise.
Jcstine1995@gmail.com, +1 503 449 4628.