Leaders in Tech and Ecommerce

The episode contains several genuinely useful ideas for supply chain operators - fat-tail yield risk being drastically underpriced, banks starting to use satellite crop intelligence for supply-chain assessment, and strategic sourcing relocation as a response to persistent climate failure - but these are interspersed with repeated COVID non-linearity analogies, generic AI-is-improving commentary, and vague encouragement to 'think differently.' The signal-to-noise ratio is moderate.

“a lot of these models say, oh, maybe there's going to be like a 5% reduction in yield or maybe a 4%. But actually a lot of the models, when you look, look at them, there's a fat tail risk of things like 30% reduction, 50% reduction”

“we are overall underpricing risk quite significantly”

The framing of a Treefera agent sourcing answers from satellite observations rather than published text is a genuinely fresh angle on the AI-agent concept, and the observation that traditional S&D models systematically underweight tail scenarios is well-made. However, much of the episode recycles familiar territory: AI is improving fast, climate change is non-linear like COVID, supply chains need better data - none of this is contrarian or first-principles for a B2B audience already following this space.

“What Tree Ferrers Agents is doing, it's going off and making those observations or gathering the results of previous observations from satellite and scientifically robust models. So it's sourcing its information from a completely different domain, a different realm”

“If you look at Hungary and corn growing in Hungary, a lot of Hungary, you know, used to 20 years ago, corn, you know, there's a huge amount of corn produced out of Hungary. It's far less viable than it used to be”

Jonathan Horn is a credible practitioner: 20 years across Citigroup and JP Morgan in risk and balance-sheet AI, a farming upbringing, and a physics research background - all directly relevant. He has a named blue-chip client (JP Morgan) and is running a Series B company solving a real problem. He is not a career podcaster or pure thought-leader, though the company's youth and the episode's promotional framing limit how deeply operational his insights get.

“15 years at Citigroup and five years at JP Morgan. And that was mostly in the kind of space of risk and control”

“I can name this plant because it's part of their website. So that's straightforward. So we provide data to JP Morgan as an example”

The episode offers a solid level of concrete specificity for a podcast: 4.7 million fields modeled in a weekend, a 7-year crop history, beef prices up 20 - 33%, JP Morgan named as a client, Eastern Europe-to-Sub-Saharan-Africa sourcing shift, 150 candidate locations for a procurement decision, and 47 Singapore meetings. These ground the claims meaningfully. What's missing is harder evidence: no published studies, no precision on model accuracy, and client names beyond JP Morgan are withheld.

“On Monday morning we had a clear view going back seven years of 4.7 million fields in the US”

“cattle are you know, kind of finished animals are ah, 20% more expensive than they were kind uh, of last year and the year before. And for certain sectors that goes up to 33%”

The host provides useful scene-setting context (the Zurich event, the chocolate-producer anecdote) that nudges the conversation toward concrete terrain, which is above average for this format. However, he never pushes back on a single claim, never asks about competitive differentiation, model accuracy, pricing, or failure cases, and closes with a standard 'what advice would you give' question. The conversation is facilitative rather than interrogative.

“I will give you two examples that came across in an, in an event we've done recently in Zurich”

“I was talking some months ago with one of the larger chocolate producers in the world and they were so, and I was talking with the person who was looking after procurement analytics”