Just bringing this one back to the top. It's pretty crazy how fast it's improving month to month. I used it to analyze my favorite western state's 2024 draw results to anticipate what strategy I might want to consider next year and it did a pretty good job. The truck is in the prompts you use which is actually pretty intuitive once you get the hang of it. For example, if I want to talk to it about a particular unit, I begin my prompt with, "Assume the role of a wildlife biologist or a Forester based in unit X with a B.S. in wildlife biology, a M.S. in GIS, and a M.S. in statistics. I also started getting a lot more out of it once I stopped using it like an intelligent search engine and started talking to it like it's a real person. Learning to talk to it is kinda like learning how to enter search keywords in the 90's. I used it to identify 3 potential areas to hunt whitetails in a new to me state a month ago or so after giving it a half dozen factors that were most important to me. I fact checked it by researching those areas and talking to biologists for those factors after and while one of them is definitely a no go, two of the three are worth checking out in person.
It's also interesting to see the strengths and weaknesses of the big 3 options- chat gpt, Gemini and Claude. Each has things it's better at than the others and once you get a feel for that, you can get better results for different applications by bouncing to the better AI for what you're after.
It's still got some bad info but it knows this, too, and will help you filter through the BS. For example, I have it set to only share info it has greater than 90% confidence in which is also based on double verified source info, and to cite sources. That kind of stuff made a huge improvement in accuracy for me.
My prediction, hive it 6-12 months and I think services like GoHunt will be nearly obsolete. Maybe those services can get a little extra life by incorporating AI, but at some point, the value just won't be there anymore vs a general purpose large language model and they'll start losing subscribers. I also think early adopters will get an edge for a little while until the masses figure out what they've been missing and then it'll be a bunch of whining about AI ruining the draw odds everywhere.
We use it at work on a daily basis. We're a startup manufacturing company with pretty limited people resources and it's doing a pretty good job doing a lot of tasks normally done by roles we don't have yet, and will likely never need to add to our org chart now. I tried it for something yesterday I was sure it was going to flop on, but it came back with an A+ solution, and it did it in about 3 seconds. A lot of white collar jobs are in real trouble. When I have this conversation with friends who don't want to accept it, and therefore aren't learning how to use it cause they don't like what's coming or it's not perfect yet, I just shake my head. It's no different than when the internet became commonly used. There were business owners I knew who insisted on sticking to the yellow pages and those that built websites and adopted e-commerce, and we all know which of the two eventually had to play catch up.
