Can AI Get EQ + Judgement + Context and Displace Agronomists?

Index:

1. Tacit Knowledge and the Future of Work: A Perspective on Agriculture

2. Context is Key and Requires Tacit Knowledge

3. Emotional intelligence and Human Connection

4. Judgement and Ergodicity

5. What’s the cost of being wrong?

6. Final Thoughts

Tacit Knowledge and the Future of Work: A Perspective on Agriculture - Elliott Grant

There is a constant discussion about what jobs AI will displace— including whether agronomists will be displaced.

The topic is interesting because if agronomists/agronomic advisors are displaced (or their value to farmers declines) and GenAI tools become the default spot for farmers to ask questions, it shifts value chain dynamics in multiple ways— product information dispersal approaches, sales rep focus, marketing budget allocation, trial/tour initiatives, and more. All of this has implications for input manufacturers, input retails all the way to equipment dealerships.

A useful way to consider implications is through frameworks. I referenced The Job Bundle in Mindware: 33 Mental Models for The Modern Agribusiness Leader as one consideration.

In Tacit Knowledge and the Future of Work: A Perspective on Agriculture, Elliott Grant shares an insightful framework that is useful for thinking about AI and job displacement:

I encourage you to read the full article, but here are the related quotes that stood out to me:

1. “One way to determine whether a task involves tacit knowledge is to ask the human performing it “how did you make that judgment”.

2. “AI does not learn efficiently like humans. Both of the examples — playing Go and riding a bike — were trained through millions of iterations in a simulation. The methodology for training today’s AI requires vast amounts of data and energy-intensive learning by trial-and-error to optimize model weights.² That is not how the human brain works. Humans learn quickly from very sparse data. This implies that AI will struggle to gain hidden knowledge about tasks in domains that are both hard to simulate (such as biology) and difficult to quickly run millions of real-world experiments.” (emphasis mine)

3. “There’s no reason to assume that the total area of knowledge remains fixed. More likely is that leveraging AI tools as ‘co-intelligence’ will expand the size of the pie — creating new breakthroughs and unlocking more human intuition.”

The last one, and to me the one I have been thinking on lately because of a couple conversations and articles I read (more below) this week was this one:

The last quote encapsulates what I have written in Disrupting Agronomy? Mental Models for Why the Future of Agronomy is Ai-Augmented not Automated and The Shortcomings and Opportunities of Large Language Models in Agronomy.

Which brings me to another article I read this week on AI by Dan Maycock: Harvesting ROI from AI on the Farm and a statement from venture capitalist at Voyager Capital, Erik Benson:

The reason I highlight the article from Dan Maycock is I agree with much of what he states, and I think several examples show why farmers themselves will be challenged to forego their trusted advisors.

Regarding Erik’s comment, the time cap is always useful to consider (on a long time horizon, maybe it’s true), but I’d speculate that much less than 10% of crop advisors are displaced by AI in the next 5-7 years.

My rationale is effectively three-fold, drawing upon the previous Elliott Grant article and previous Upstream writings:

Subscribe to Upstream Ag Professional to read the rest.