I would never say "Wake up and smell the coffee" what the fuck

I think the thing stopping forums from getting flooded with bots is that they aren't great to get money from or anything so it's not worthwhile. If everyone were on forums, it'd be like social media again regardless of the format. Not that forums will take off anyway, they make for poor mobile experiences, and aren't fast paced. But that's fine, it's nice to have places online that have few people, there's already plenty of places to go if I want many.
I remember seeing this image filter being developed by researchers which exploits 'blind spots' in current AI image generation. You run your image through, it makes subtle changes that are barely perceptible at most, and then (supposedly) if your image is used by the program, it will cause artefacts to come out in the image, rendering the result useless. I wonder if there is any similar defective points in ChatGPT? I wanted to explore that. All the useless shits in my uni wont shut up about how ChatGPT writes the code for their assignments, so I decided to see if it's any good at lifting anything heavier than beginner Java. And it absolutely shit its pants.
I said to it "
Describe the ASL instruction from the MOS 6502 CPU"
ChatGPT
6502.org
Mind you, the MOS 6502 isn't niche or anything, it was used in the Apple I, Apple II, NES, the Commodore PET, and because of its small instruction set it's a go-to for hobby projects. It would probably be one of the most well documented retro CPU's online, yet ChatGPT cannot give even 10% accurate information about a particular instruction. So it seems like accurately reporting technical information is a massive blind spot, maybe that could be used to create some sort of a Captcha?