Our Outdoors: (A)I’d Rather Be Fishing
By Nick Simonson
Winter weather has a way of worsening cabin fever. In our screen-filled world, the glow of a blue and white backdrop with black letters rolling out across it is a panacea of sorts, numbing my mind to the increasing white in the world outside, and creating hopes from whole cloth of a shift in the conditions and a changeover to spring’s greenery. At the very least, I can come up with some good “Top 10 Outdoors” tip lists while waiting for winter’s end.
Beyond the spillage of text from the gray keys to the white digital sheet that often gets me through the latter stages of the season are new technologies and distractions, and they are far greater than what I could have imagined 22 years ago, when a copy of the now-outdated Quark page maker software provided the starting point of this side-career. These new technologies are, in some cases, far more terrifying than the clunky editing software which would inexplicably shut down just before deadline.
First disclaimer, no, ChatGPT – the OpenAI darling that every columnist of every sort is saying will replace them – did not write this column. I’m still too proud for that. However, when I requested a 700 word article for spring fishing tips in the upper Midwest, I was floored by not only how well it was written by the bot, but the detail with which it was. It certainly didn’t have my over-wordiness or the flowery language I’m known and criticized by my editors for; but from timing weather patterns, to adjusting location based on water temperature and spawning movement, to recommending a number of baits, including those in my own brand-loyalty zone and others I hadn’t heard of, but quickly added to my online shopping cart, it was spot on. Short of casting a line and detailing its experience, the accuracies were startling and would be about the same as I’d provide any new angler for the ups and downs of the season to come.
To further test the artificial intelligence’s experience with angling, I asked what the five best springtime tactics for smallmouth bass were, and the response was again on or near the money in the handful of categories it provided, with the weakest being the final one covering topwater angling tactics, which I don’t often employ until the cusp of summer, but I doubt the digital brain on the other end of the page was considering calendar and unofficial cutoffs in its response. The fact that both samples hit so close to the mark and were concise and helpful didn’t concern me about my pending replacement. Knowing that AI draws from the millions of articles scattered across the online repository of human consciousness, I felt that perhaps in some way, the last two decades of doing this contributed to the knowledge the machine on the other end (the world’s fifth largest supercomputer apparently) was drawing from my text stashed away on the web somewhere. There’s that pride again. So, there was nothing really to fear. That is until I tried an equivalent AI image generator.
If you would have seen the picture of the largemouth bass eating the crayfish I requested from Craiyon AI – a “lite” version of the highly touted DALL-E image creator in the same vein as ChatGPT – you would have tossed it back immediately. In fact, you would have flung it halfway across the lake in terror. Strange crab legs grew from its mouth, and they were not on the tail of a crayfish or any freshwater arthropod I had ever seen. It had multiple eyes and a strange jaw, and while the tail looked like that of the familiar summertime fish I had requested, the rest of its body changed pattern and grew farther and farther away from recognizable. In some of the samples provided, the supercomputer placed multiple extra human fingers on one hand of the angler with some growing from the fish, and some bending in unnatural directions. It was disgusting, and my guess is the collection of processors on the other end of the connection is going to need a few more years of fine tuning to get it right.
Afterall, a computer will never know the sting of winter wind on exposed cheeks during another shoveling session and the relief that comes with the first warm day afloat in spring with a rod in hand. And it will probably never feel the sandpaper grit of the bottom lip of a four pound bass, and the dozens that come after it on a fast day of fishing, leaving a white patch on its rubbed-raw thumb. Beyond telling us all that is already known and stored in the brain of the internet of things, it’ll never experience something new in the field or on the water, and that will thankfully remain with us to share with it and each other…in our outdoors.