AI tips, tricks, faults, and failures

[Khunter]

I was extremely disappointed with ChatGPT when I tried to use it for researching where to apply for tags this year, it gave me a lot of completely wrong information.

As of now, GoHunt is still worth the $99 or whatever over the free chatbots as far as I can tell.

Edited to add: I won three decent public land tag draw this year, all the 3 tags I picked based off odds I found on GoHunt.

If I hadn't used GoHunt and instead just relied on the free chatbots, I may have gotten nothing in the draws.

I was surprised how well chatGPT carved up multiple years of CPW recap reports to answer my questions and suggest which hunt I had better chance to draw that included multi year “trend” . Mine is a paid version, not free.

I also used Chatgpt to generate illustrations and help draft text for a 12 page fable I distributed at a symposium i plan for 340 folks each year. Just a fun no risk project to add a little levity that was well received. Loosely framed on goldilocks and the three bears concepts.

This fun project also served as a great learning experience for how to effectively harness AI via clear constraints and instructions so it does not go rogue and waste your time with useless iterations.








One place AI works well for my work is distilling and organizing key passages of BLM resource management plans, EAs, and EIS. Those can be massive docs. To aid framing up issues of concern or support for my agency to provide comments on. Requires careful checking to ensure things are not missed of course. But definite time saver to find and extract certain relevant passages including page citations for each.

Another is state oil and gas rulemakings. Feed AI two dozen prehearing position statments from other parties and set it to the task of identifying parties who align or do not align with our positions and it does well to carve through all of that and make a nice table outlining all of it. Not ever perfect but big time saver.

In short, feeding AI a mountain of text with clear instruction of what to look for and how to organize results can be very effective. Over reliance on said result without due diligence would be a recipe for failure and embarrassment in public proceedings. Have seen some doozies where a party clearly used AI as their ‘brain’ instead of using own and it showed.

 

[Forkyfinder]

Ive found ai useful - but like others i find the prompts a major driver in the success. Ive heard "you tell ai what to do" comparing to "asking google a question."

Theres already a company (wont name) that is tailoring ai for hunters to completely eliminate e scouting. Kind of crazy to go from doing it entirely boots on the ground, to doing it on the shitter, to having an AI service do it for you for less than a trail camera.

I dont do a lot of data analysis or manipulation at work, but i have found it useful for asking difficult technical questions - but it is limited there as well - because what you understand from the output is critical for your next deeper input. In other words - if you cant have an informed conversation about a topic - its hard to manipulate the prompts in the first place. Ive used it for draw odds successfully and comparing some harvest stats.

One thing ive noticed - if you are deliberate with the intent its more compliant with your request. Sometimes - it will complete a prompt better when clarifying its sattire. For example:

 

[TOGIE]

Just had Claude ingest some quotes for residential concrete work at our house. I was extremely skeptical of the lowest bid, because of how much lower (significantly lower) it was than the other two which were nearly identical. Specs were basically identical in all of them - square footage, estimated cubic yards, etc but i was certain they must've missed something when punching the numbers.

Didn't want to bring it up to the contractor yet to avoid accidentally triggering an unnecessary price hike if it was legit.

After ingesting the quotes and researching local typical prices it determined the other two quotes were actually well above local average and the lowest one was right in line.