Proof of Corn

What counts as “AI growing corn”?

• Many argue the project mostly proves an AI can hire humans, not that it can “grow corn.”
• Some say that’s still meaningful: a farm manager or investor already orchestrates work rather than doing field labor, so replacing that role is non‑trivial.
• Others insist a stricter bar: one prompt like “grow 500 bushels by October” with no further human steering; anything else is just a fancy search/assistant.

Human role vs AI orchestration

• Repeated concern that the human initiator is doing core work: choosing prompts, interpreting outputs, handling money, and manually emailing or fixing issues.
• Critics note this is closer to “AI‑enabled hobby farmer” than autonomous agent.
• Some suggest making every email, decision, and inbox interaction AI‑driven, with humans obeying instructions exactly, to make it a real experiment.

Practical farming and scale concerns

• Multiple commenters with farm backgrounds say 5 acres is “a garden,” too small to interest custom operators at realistic prices.
• They highlight missing or naive elements: seed choice and ordering windows, fertilizer plans, fungicide decisions, moisture vs drying costs, harvest timing, and local knowledge.
• Existing precision ag and autosteer already automate much of tractor work; the remaining value lies in nuanced judgment and on‑site monitoring.

Budget, location, and business viability

• Budget page is widely viewed as unrealistic: underpriced labor, missing machinery, seed, irrigation and risk costs.
• Polk County, Iowa is flagged as a questionable choice due to development pressure and land economics.
• Several predict the project will lose money regardless of AI quality, given commodity margins and recent bumper crops.

Autonomy, architecture, and missing pieces

• People ask where prompts, logs, and Claude API calls are; without full transcripts, claims of “every decision logged” feel hollow.
• LLM limitations raised: recency bias, lack of sensors, weak world models, tendency to “wobble” rather than commit, and difficulty adapting mid‑project.

Ethics and externalities

• Some dislike experiments that send convincing but non‑committal inquiries to real businesses, calling it AI‑driven spam and time‑wasting.
• Others counter that land‑leasing firms exist to field such requests.

Broader implications and mood

• Thread frequently pivots to AI as future managers/CEOs coordinating human labor, evoking dystopian “AI boss, human serfs” scenarios.
• Enthusiasts see the project as an early test of AI‑run business processes; skeptics call it hype, “reverse centaur” theater, or a bar bet likely to fail.