AI forum post slop has to stop

I am not against AI. I use it in projects, for drafts, for debugging, and for the ordinary private work where a second pair of eyes is useful. That is not the problem.

The problem is posts that the author did not write, does not understand, and cannot defend. Someone pastes a prompt into a model, gets back a tidy wall of text about a bug, a design choice, or a half-formed idea, and drops it into the forum as if it were a question. Then another person, or the same model on the other side, answers in the same register. The thread looks busy. Nothing has been learned.

A forum works because a person brings a real constraint. They have tried something, hit a limit, and can say what they expected and what happened instead. When that is missing, the replies have nothing to attach to. Advice becomes generic. Corrections bounce off, because the original poster cannot tell a good suggestion from a bad one. The conversation turns into models smoothing one another’s sentences. It looks like discussion. It is not.

There is also a cost to everyone else. Reading a generated post takes the same attention as reading a human one, and often more, because the wording is confident in places the author is not. People spend time answering questions that were never actually asked. Good contributors notice, get quieter, and the board fills with threads that resolve into nothing.

Using a model to organize notes, check a paragraph, or sketch options is fine. Posting the output as your own thinking is not. If you cannot explain the problem in your own words, you are not ready to ask the room to solve it. Write the post yourself. Say what you tried. Say where you are stuck. If a tool helped you get there, that is your business. The question still has to be yours.

Otherwise this stops being a forum. It becomes a relay between language models, and the people who actually know something stop answering. What is left is noise dressed up as a thread, and nobody is obliged to treat it as anything else.

11 Likes

When will AI determine humans are no longer necessary? Welcome to SkyNet.

I know what post you are talking about and I have debated about removing it. I purge and deny access to this forum constantly to bots that sign up with valid email addresses and then post slop. Overall I feel that I am winning the battle on slop but it’s still an ongoing battle. I am still doing this myself and not just connecting AI to my forums to do so.

In the bigger picture the AI Slop is breaking the internet and starting to bring the Dead Internet Theory into a reality.

7 Likes

IMO, there are two types of people coming to a forum like this. Those with experience and those without. (and those just lurking - me 99% of the time) Those without get paid with a little experience, and those dispensing that experience get a little dopamine hit. AI does a beautiful job disrupting that transaction. As the heat in your post(s) clearly highlight.

Educational (how-to) forums like this are probably not long for this world. As the years go by it will be harder and harder to tell who/what is on the other end of your forum post. That raises the mental cost on the two forum members and the moderator alike. While those costs rise, AI does an ever better job educating someone who would have gone to forums.

If a person wants help doing something, just ask AI. If a person wants community, I think (hope) the future is retro. More small in-person meet-ups and big conventions.

I’ve asked AI for help, the results are often not as good as what a real person would give me (or just wrong). It won’t stop me from asking, but AI is not all encompassing.

Forums are like a library, then contain wisdom beyond one’s own reach. Unlike a library they contain knowledge that is targeted to another person’s needs, yet may be adapted to my specific needs. Example being that I had a problem with my firewall, asked a question after going through several attempts to fix it. A day later someone asks a different question, one that sort of related back to mine and gave me the idea that lead me to the fix on my firewall.

The issue is without forums you end up with no source for new things for AI to ingest. I think forums like this run by real people like myself and people like Level1Techs are the future. As all the other big social media companies are becoming algorithmically driven slop warehouses people will go back to looking for a refuge where they can have an actual conversation with a human.

12 Likes

Had no idea this was happening, I can’t even see the point of doing it. Though I have begun to see that Claude and CHAT, does not give too many responses before it says your time is up before it resets in 5 hours! Everything is moving to a subscription model if it hasn’t already.
Not sure how you can see if something has been generated by AI but I have seen the major providers have said they can check if they have provided the response or not. Presumably this is to provide institutions with a means to minimise cheating. Maybe something like this in time can be adopted.

My “claude” never runs out of tokens, but it is also EXTREMELY slow because of the hardware I have at hand. Flip side is how many hours do I want to run 60 watts through the system to make heat. Getting into winter so the heat is not a major concern, but you get the point.

If I had money, my local AI would be running closer to 300 watts, but it probably wouldn’t balance out the power useage over time equation.

A natural and reasonable concern to raise given the degree to which individuals are leaning into AI without understanding that its confident presentation is not built on thinking or experience… only programming.

Individuals are becoming increasingly disconnected from first hand determination, processing information or applying rigour. Gen Z is the first generation objectively dumber than its parents. AI is a natural progression for them to grasp onto.

The deeply flawed content which AI confidently shares has led to me becoming tired of reading posts of any format across any media which use the same AI “enriched” style, tone and level.

A human displays nuance with respect to confidence, AI does not.

I am becoming suspicious of anything I read which illicits red flags for AI input. I question if the OP ran his own post through an AI given some of the structure.

1 Like