How AI will make “the little guy” a better fisherman

Since I was 5-years old I have spent most of my summers either fishing or finding a way to go fishing. Fishing is a generational sport. I learned how to fish with my grandfather. My father-in-law and I taught my son to fish. Now my son and I enjoy spending our summers teaching my grandson how to fish. “The little guy” is off to a good start, as you can see.

Perhaps unexpected to non-fishers is how much technology affects both how we fish and how many fish we catch. Fishing has always inspired the innovator in me. There are always new places to fish and techniques to try to catch that wily old bass. But technology is at the heart of fishing, particularly fly fishing. I first learned to fly fish from my grandfather using a tubular-steel fly rod with a simple spool reel and braided silk line. Today we use specially coated polyvinyl chloride lines with fluorocarbon tippets to throw flies two to three times bigger, twice as far as my grandpa ever could. Our fly rods are made from graphite and boron and the reels are precision machined out of blocks of aluminum in forms that defy common sense. Technology increases both the speed and precision with which we can put our favorite flies over the target area.

Target area is a keyword in fishing. “90% of the fish are in 10% of the water” my grandpa once told me. He bought one of the first depth finders and I learned the value of sonar in finding fish using that simple below the boat sensor. But I never imagined that five 16-inch “fish finder” screens running on full PC-like computers with fish sizing capabilities would become the technological norm for the sport.

But if I have learned anything from my predecessors it’s that papas need to stay up to date with all the latest techniques. So this year, given all the ChatGPT news, I thought I better consider how AI might help my grandson catch more fish.

AI comes in two basic forms today: generative and operational. Generative AI is defined by the likes of ChatGPT and DALL-E – algorithms that can generate new content of many types. In the case of fishing this already means new fishing lure designs. The figures at right come from tackle.net which “imagines new lures with the help of AI.” I am not sure how or how long it takes to test all the designs to learn what works best, but like most generative AI efforts the output is entertaining. Generative AI is still somewhat risky because the data upon which it is trained is often unknown and the algorithms appear to put more emphasis on answering the question than answering it correctly. See the “South Dakota Governors” test as scary example of generative AI. AI designed lures may not catch more fish, but I guarantee they will “catch” more fishermen and fisherwomen. We can’t pass up flashy tech things in the bait store.

The other form of AI is what we build at Tamarack, Operational AI. Operational AI helps improve operations by building models using historical data that predict future outcomes. Good outcomes are pursued more diligently, and bad ones - the ones that contain risk - are avoided. The methods of developing and using operational AI are more transparent and easier to validate. The commercial fishing industry is already using operational AI in smarter “fish finders” that remember past successes and recommend, based on that data, the best place to fish. It’s even more true today than it was in the 1960’s – “90% of the fish are in 10% of the water.” AI will be very good at finding that 10% of the water.

AI for fishing, like most every application, is inevitable. Hopefully it will also help my grandson avoid those snags that will take his new AI-generated lures. They look expensive.