Catching halibut with ChatGPT

OpenAI's latest customer story ditches the usual enterprise-software backdrop for something saltier: a halibut boat. The piece follows a fisherman who started leaning on ChatGPT not to write code or draft emails, but to sharpen the unglamorous mechanics of commercial fishing — reading regulations, working out bait ratios, and making sense of the shifting weather and tide data that decide whether a trip is worth the fuel.

What stands out is how mundane the use case is, and that's the point. Halibut fishing runs on quotas, permits, and depth charts that change season to season and region to region. Instead of digging through PDFs from fisheries agencies, the fisherman describes typing questions straight into ChatGPT and getting a plain-language breakdown of what's allowed, what gear works best at a given depth, and how to plan around currents that can wreck a day's haul in minutes.

OpenAI frames this as proof that generative AI isn't just for coders and marketers. A tool built on scraping the internet's text ends up useful for someone standing on a deck in cold water, trying to decide whether to reset lines or head back to port. That's a long way from the chatbot-as-productivity-app pitch that dominates most AI marketing.

There's an obvious skepticism to raise here: a language model has no idea what the tide actually did today, and it can be confidently wrong about local rules that vary by nautical mile. The story doesn't pretend ChatGPT replaces judgment or experience — it's positioned as a faster way to get oriented before making a call, not the call itself. Still, the image of an AI chatbot factoring into whether someone catches dinner, or a paycheck, says something about how far this technology has wandered from Silicon Valley.