Aldeer.com Forums Alabama Deer Talk Aldeer General Forum Hunting Platform Build - ChatGPT

I used this prompt:
Can you sketch hunting stand building instructions step by step and give me the boards that I need to buy. I want a platform that is 5ft high with the legs buried ~3ft in the ground. platform 60" by 60 inches (2x6's) and the frame made of 2x6s. legs are 4x4's, I want ~2ft 2x4 corner braces on every corner to support the platform (8 total braces)

Then had tell it to make it look like an image with lumber that I found online and it made this. It messed up with the safety rail that it added at the end, but did a pretty good job, and gave measurements.

Its a great tool. I used it for a pretty complex project (and similar); and it created a professional engineering package for it. It can be a great time saver however, you need to be a subject matter expert on the material so that you can check every single thing it give you. It will lie to you and make shit up and if you don't know the material you will never know.

The last image that you created was great. Not create a set of instructions where the 4x4 corner posts extent up thru the plaftorm by ~48". Make the total length of the 4x4s 10ft and only sinke them into the ground 1 ft. calculate the resulting total height of the surface of the platform. recreate the last image with these specs

Yeah.

Typical for what I've seen from AI in multiple applications, it's 85% there. There will always be some mistakes. First step should be explaining exactly how to make sure your posts are "square". You'd be amazed at how many people do not know how to complete this task. Look at Picture 3. Says install 3 joists but only shows 2 on the diagram. 14.25" centers are correct but the pic is not. Pic 5 shows installing the deck boards running in the same direction as the joists. That's not correct either. They would be run in the opposite direction of the joists so that each deck board has 10 attachment points across the span. It's a solid design but be careful using AI on builds. AI is smart but you've got to have some common sense to go with the design.

Pretty neat. My experience is similar to what others have already said. Its a great tool to get you close, but still requires a bit of scrutiny to be 100%.

Drawing is incorrect again. With 1' in the ground, 6' height and 48" of post above the platform, AI has managed to get 11' of wood out of a 10' 4x4.

Oh cool, thanks for posting that pic. I took your pic to my project chat and asked "Here is an example you created for a friend of mine by using the prompt "Can you sketch hunting stand building instructions step by step and give me the boards that I need to buy. I want a platform that is 5ft high with the legs buried ~3ft in the ground. platform 60" by 60 inches (2x6's) and the frame made of 2x6s. legs are 4x4's, I want ~2ft 2x4 corner braces on every corner to support the platform (8 total braces)" Can you give me a similar diagram for this project?"

Here's what it came up with.

I said this was a complex project, not because the build was complex but because the logistics were horrendous. The engineering package maps out all the logistics as well as the build trade off study, selection, materials list, instructions.

It's in a MS Word document and I don't know how to post that for you to see.

Not the way I would do it

I don’t need AI for that

That looks a lot better than mine.

Admittedly I am trying to do dinner and haven't read all the replies. But it sounds like you told it everything, and it just gave you a sketch?? I've never used chatgpt or other.

seems like a lot of work to get 5 ft off the ground.

A lot of people do though.

I can see where it would come in handy

I get the value in chat gpt, AI, etc, but it’s like a calculator for math.
My wife and I are 3x better at doing simple math than our kids because our kids were all able to use calculators in school.
The teachers (apparently) are all pushed to hit goals and let the kids basically cheat with calculators.
They’re getting better, but my point is it’s good to be able to sit down and figure that stuff out on paper (also).
My FIL quit school when he was 16, but could figure every raw material in a bathroom remodel or carport expansion. It’s more lost knowledge I’m afraid.

On the second platform, that's way too many posts and foundations for a platform like that. Why build it 14'x14' when standard lumber comes in 8', 10' and 12'. It's got way too much blocking in there too. Nobody uses that much blocking.

On the first platform, that is a better, simpler design, but there's no need to use 2x6's for spanning 5 feet. A 2x4 will span 8' for a walking surface just fine.

The pictures are odd too in that some pics show the joists framing to the bond beams and other pictures show the joist bearing on top of the bond beams. No need to bear them on the top. I would use 2x6's for the bond beams, use a 2x2 to support the bottom of the joists and just fram the 2x4 joist into the bond beams just like conventional floor framing.

Chat GTP is full of weird information sometimes.

Good for you. I personally have never built a deck and didn't even know where to start. This got me heading in the right direction in about10 minutes.

This is great information, thank you.

I've used various flavors of AI to assist with things such as reviewing building codes or checking my logic when troubleshooting problems on my son's F150. I've used it to summarize documentation or even creating a skeleton for a presentation. Its only getting better over time and prompt engineering really helps it provide you what you need.

My first thought on the output for the platforms above were that they seemed over engineered, but in hindsight the prompt maybe didn't limit the design to consider commonly available lumber dimensions or the visualization wasn't absolute in that the image provided would require a set of steps or ramp versus just building closer to the edge of the hill instead.

I've had fun using AI to create imagery. Let me see if I can make something for Skinny.