Welcome to The AI Brief.

Every week, you get clear, honest breakdown of what’s really happening in AI. No recycled press releases pretending to be insight. Just the developments that actually matter.

Every AI newsletter covers the same three industries. Healthcare, where AI is going to revolutionize diagnostics any year now. Legal, where the billable hour is going to change. Finance, where risk models are going to be transformed. The results in all three are still mostly theoretical, announced with fanfare, with actual deployment still somewhere on the horizon.

Meanwhile, in an industry most technology writers have never thought about, AI has become essential infrastructure. Not a pilot. Not a proof of concept. Actual deployed systems doing actual work, with results that already show up in measurable costs and efficiency numbers that have changed how the industry operates.

This issue names that industry and explains why it got there first. Let's get into it.

Bloomberg's Marcus Ashworth wrote plainly recently: "No more reliable safe havens."

After all, the S&P fell over 7% from the February peak. Bonds, even with less risk, are barely keeping pace with inflation.

So-called "diversified" portfolios have gotten hit from multiple directions.

Meanwhile, the world’s wealthiest have been setting records in another asset class.

Circumstances are always unique, but after the dot-com bust, it grew roughly 24% annually for a decade. After 2008, roughly 11% annually for 12 years.

Blue-chip art.

Why? It trades globally in multiple currencies, has scarce supply, and has shown near-zero correlation to equities since 1995.*

With Masterworks, 70,000+ investors allocated $1.3B fractionally across 500+ artworks featuring Banksy, Basquiat, and Picasso.

Accredited? You can invest in a diversified portfolio of postwar and contemporary art alongside two other real assets. From 2017-2025, the mix would’ve beat the S&P 500 by 3.1x.

See if you can improve your portfolio performance all in one diversified strategy.

Our subscribers skip the waitlist

^{*According to Masterworks data. Investing involves risk. Past performance is not indicative of future returns. Important Reg A disclosures: masterworks.com/cd}

The industry is agriculture, and the reason you have not read much about it in AI coverage is that agriculture does not hold conferences in San Francisco, does not produce founders who give TED talks, and does not fit the narrative of white-collar work being disrupted by chatbots.

What it does is produce measurable results, and that is increasingly where AI is most comfortable.

John Deere's See & Spray system uses computer vision mounted on a sprayer to identify individual plants in a field, distinguish crops from weeds in real time, and apply herbicide only where it is needed. The result is a reduction in herbicide use of up to 77 percent per pass. This is not a pilot or a proof of concept. It is a product deployed at commercial scale on farms across the US and sold globally. The technology works because the problem is specific, the feedback is immediate, and the economic incentive for a farmer is direct: less chemical spend per acre, visible in the same season.

Satellite imagery companies are giving farmers field-level crop health monitoring that would have required a team of agronomists and weeks of manual scouting a decade ago. Services analyse multispectral imagery to identify stress, disease, and water issues across entire farms and surface the problem areas before a farmer would spot them on foot. AI-driven disease detection has made this even more accessible: a farmer photographs a leaf on a smartphone, the system identifies the disease and severity in seconds, and returns a treatment recommendation. This matters most in regions where agricultural extension services are limited and expert diagnosis was never fast or affordable.

What separates agriculture from every other industry that AI is supposedly going to change is structural, and understanding the structure explains why the timing worked here when it has not worked elsewhere.

The problems are well-defined. Is this plant a weed or a crop? What is the predicted yield given current soil and weather data? Where exactly is the water stress on this field? These are not open-ended questions. They have correct answers that can be verified quickly. Healthcare AI faces questions where the correct answer takes years of trials to establish. Agriculture does not.

The feedback loops are fast. A growing season is months. A farmer who tries an AI-driven planting recommendation sees the result before the year is out. Medical AI faces decade-long regulatory approval processes. Agriculture does not.

The data is abundant. Decades of satellite imagery, soil sensor networks, and yield records give AI systems real signal to learn from. The inputs are measurable and the outputs are measurable and the connection between them is something machine learning handles well.

The economic incentive is direct. Every input cost reduced goes straight to margin for farmers operating on thin returns. There is no procurement committee, no change management program, and no multi-year implementation project. If it works, you use it again next season.

Healthcare AI is going to be significant when it arrives. Legal AI will change how documents are written and reviewed. But those timelines involve regulators, liability frameworks, and institutional inertia that are years from being resolved. Agriculture did not wait. It found the tools, tested them in a season, saw the numbers, and kept going.

The industry nobody writes about is the one actually running.

The AI industry has spent three years explaining how artificial intelligence is going to transform healthcare, law, and finance. Meanwhile the most widely deployed, results-verified, commercially scaled AI application in the world is a camera on a tractor that has learned to tell a soybean from a weed. The soybean did not need to sign a consent form or wait for regulatory clearance.

Healthcare, legal, and finance dominate AI coverage because the writers covering AI live in cities where those industries are visible and the founders of AI startups in those sectors are skilled at talking to press. Agriculture gets ignored for the same structural reasons that technology coverage has always undercounted the industries that actually feed, build, and move things.

The pattern of where AI actually works is becoming clear. It works on problems that are specific, data-rich, and have fast feedback loops. It works where the economic incentive to adopt is direct and the path to deployment is short. Agriculture meets every one of those criteria and has been quietly running AI at scale for years without anyone in the mainstream AI conversation noticing.

The broader point worth taking from this is not that agriculture is special. It is that the AI applications most likely to be working right now are probably not the ones getting the most press coverage. They are in the industries that solve specific problems, move fast, and do not need a keynote to prove the results.

Agriculture was probably not the answer you expected, which is exactly why it is the right one to write about. The industries that are actually running AI tend to be the ones that nobody predicted.

Reply and tell me which industry you thought the answer was going to be before you read it. Those replies are going to be interesting, and the best ones will end up in a future issue.