Under the Agridome

Editor's Note: Today DTN is sharing content by that is normally reserved for paying customers. Not a subscriber? Check here for a trial: https://www.dtn.com/…

**

On Thursday you could hear a little bit of a sigh of relief across southwestern Ontario farm country. In many ways it's a bit like grasping at straws, because we've had such a tough year of commodity pricing. At the present time, we have what they call a bull flag pattern in soybeans, which often sees a big breakout something dramatic. However, I'd say we're still grasping at straws. These crops are huge in the United States and maybe the 20-cent rise Thursday reflected a few dry days coming up in the Eastern Corn Belt.

Yes, big supply continues to win, but it is also hard to know. Or to put it more succinctly, I'm seeing almost a growth industry within the farm community and the farm business community of people not wanting to believe what is right in front of them -- one of the biggest crops ever. Is it being aided by some of these private crop tours putting more colour to the bare bones of USDA estimates.

At the same time, I'm getting farmers contacting me saying so-and-so said this and so-and-so said that about where commodity prices are going. Keep in mind, when you look at that, you should keep an eye on basis levels in the United States as well as future spreads. It is true that future spreads have narrowed especially in soybeans during the last couple of days, but it's so hard to tell. You would think in 2025 we would get to the point where it would be easy to predict crop yields. NDVI images from the USDA were supposed to do that. However, it seems that everybody is a skeptic.

In many ways, in 2025 we are at a bit of a fork in the road with regards to where we get our information. It has become the norm for younger people especially to avoid Google and go to their favorite AI client, whether that is ChatGPT or something else. I do some of these artificial intelligence algorithms and I've been impressed with them. However, I would give them a seven out of 10 as a mark. In other words, I'd say they get about 30% of things wrong even the better ones might get 10% wrong. My question is do we call that success?

Case in point is something called AI slop. AI slop is content created in a poor way, mass produced and generated by AI. It is increasingly flooding the internet making the internet less reliable.

Increasingly, we're getting AI-generated books clogging up eBook sites. We have AI-generated music clogging up streaming services. We have AI-generated images and videos clogging up all kinds of things on the internet producing images and videos that look real but are not. Then there is a plethora of AI-generating bots on social media sites exchanging information on each other and creating more disinformation. It is becoming a very difficult to ignore.

At the same time, farmers are using artificial intelligence in their farming operations. In many ways, it is a good thing, but it is likely to come with the limits. For instance, I remember last year I was able to use an artificial intelligence client to find out the specifics on how to repair the air conditioner in my tractor.

However, I still couldn't do it myself. The next thing that will be coming along via queries to your loyal scribe is if we can use artificial intelligence to tell us when to sell our corn soybeans and wheat.

You surely can use it. However, you better be ready to give that AI client the right prompts. Of course, therein lies the problem. If you don't know the right questions to ask, you're unlikely to get the right answers. Grain marketing is never just about predicting prices; it's about spreads, basis, crop conditions, geopolitics, and weather patterns. An AI model can summarize what's already known, but it cannot tell you when a black swan is coming.

The bigger issue is that AI is always looking backward, trained on history to advise about the future. If you ask when to sell soybeans, chances are you'll get an answer shaped by past USDA reports, not tomorrow's surprise. In that sense, it may look polished but can still amount to AI slop (bots) dressed up as analysis.

That doesn't mean AI has no place. Just like GPS and yield monitors, it can help organize information and even save time. But it isn't a marketing plan. It's simply another tool, which you yourself can scrutinize.

In 2025, marketing grain still comes down to discipline: knowing your costs, setting profit targets, and acting when opportunities arise. Daily market intelligence is key. Hopefully that 20-cent rise in soybeans Thursday was like ether to a Massey tractor. We need to get something new on the marketing horizon started.

**

The views expressed are those of the individual author and not necessarily those of DTN, its management or employees.

Philip Shaw can be reached at philip@philipshaw.ca

Follow him on social platform X @Agridome

(c) Copyright 2025 DTN, LLC. All rights reserved.