Hi from New York. Farmers are showing that AI isn't just for the office. But first... Three things you need to know today: • Palantir wins a $100 million contract for AI-targeting technology • Amazon launches an AI assistant for the merchants on its website • GM will resume robotaxi service in California after a delay of almost a year Some of the first technologies humans made, at least after the ones for killing things and chopping them up, were for farming. Plows, hoes, etc. — crops themselves are technologies of a kind, bred from wild cousins. Still, agriculture isn't something we've solved. It requires lots of land and fossil-fuel derived inputs; its runoff mucks up the water; and it still relies heavily on the capriciousness of the weather. Climate change is making all of this harder, as a world population set to keep growing for awhile means more mouths to feed. Maybe artificial intelligence can help. Recent years have seen the introduction of various agricultural tools powered by AI. Many of them focus on the specific problem of herbicide application. Deere & Co., the maker of John Deere-brand equipment, now sells a special spraying boom equipped with dozens of cameras and a graphics processor trained to identify, say, Palmer amaranth, a fast-growing, nutritious flowering plant native to North America — or, as it's known to soybean farmers, a weed. The boom, as it's towed through the field, zeroes in on and sprays stubborn interlopers like these rather than the cash crops around them. A Canadian startup called Precision AI is using AI-powered drones to do something similar. In a statement this week, Deere said results from the 2024 growing season showed its "See & Spray" technology cut down on herbicide use by an average of nearly 59%. Still, the technology has yet to be widely adopted, points out Christopher Ciolino, a Bloomberg Intelligence analyst who covers US machinery companies. "I don't think there's enough data out there yet for us to draw any meaningful conclusions," he said. Deere seems set on changing that this year with a new aggressive pricing strategy. The company and its competitors are hoping to drive — and profit from — a change from the last few decades in industrial agriculture in which "targeting" was determined chemically: soy, corn, cotton and other cash crops genetically engineered to be resistant to particular herbicides. Those altered corps allowed (and, critics would argue, encouraged) farmers to spray with abandon. It remains to be seen how the AI technology works at scale in the real world, and if farmers will trust it enough to lay off the spraying. But farmers have shown a willingness to embrace new things like drones and, for that matter, genetically modified seeds. Farms, like factory floors, are increasingly automated — higher-end tractors already basically drive themselves. In this case, the economic logic is clear — the cost of herbicide adds up. Using less of it also would mean less runoff and less collateral damage to plant and animal life outside farms, and less use of chemicals that, depending on which jury you believe, may or may not have given people cancer. Those are all good things. It's a reminder that many of the most promising applications for AI don't involve ChatGPT-style large language models. And they aren't the things that get the most publicity. People may continue to garden, but increasingly it's the machines that will farm.—Drake Bennett Thousands of pagers that exploded across Lebanon have raised global alarm about securing the supply chain for technology. OpenAI is finishing its $6.5 billion funding round with more prospective investors than it needs. A former Google executive's startup joins a crowded field in trying to use AI to improve weather forecasting. |
No comments:
Post a Comment