Fortsetzung der Diskussion von Any Stockpickers out there?:
While I agree with @markus654 I wonder about how to best deal with AI content. My personal opinion: use AI to inform yourself. Then post your thought and indicate that they are based on an AI output.
My take:
I’m a big fan of AI and use it for all sorts of things. But just post AI answers makes it impossible to have a meaningful read/discussion. A forum is a discussion among humans and not among bots.
If I wanted to talk to AI, I could talk to an AI myself.
I like the new rule that Hacker News had to implement because of this:
Don’t post generated comments (..). HN is for conversation between humans.
https://news.ycombinator.com/newsguidelines.html
I’d love to see something like that included in the rules here.
And before anyone brings up the language/translation argument:
- I think it’s fine to give an AI text as input and say “please translate from German to English 1:1.” Because the response is still the same length with the same meaning.
- But I don’t think it’s okay to use AI to generate entire new texts. Two sentences then end up becoming several paragraphs or tables, and the actual message gets lost and just repeats itself over and over. The arguments are watered down, and the quality suffers.
…if people didn’t use it to “inform themselves” - then take that information at face value without verifying its veracity. ![]()
For what it’s worth, I’ve been accused (wrongly) of posting AI content online more than just once or twice, merely for structuring a post with a couple of bullet points (and some diligent spellchecking, probably).
I’m not 100% against AI to summarise topics, “sanitise” writing or as a translation tool (most of us probably aren’t nativer speakers of English), but find it aggravating when posts lack personal input, effort or thought.
I think this says it all. The value of a forum is getting new data, personal opinions, and real-life experience from people. AI answers can be obtained directly.
It all started from my post, which I mentioned it was AI.
The reason I did that is that I’ve read somewhere (Hacker news maybe) about the situation with SpaceX. Instead of copypasting what I’ve read there, I just asked again google.
The result is a summary that helps starting a conversation. Without that, I’d have to either not post at all or try to find that comment I’ve read. If that wasn’t enough I would have had to google most of the topic, inform myself of the meaning of everything and at some point in the future reach a conclusion. At that point I might have given up to posting here because it was surely too late.
Consider my post (or any accepted AI post) kind of a newbie question with a bit more information included. If you check back you’ll see my first comment about SpaceX and sp500 which was lacking of enough information to keep the discussion opened. So my second post had way more info that could have been debuinked. It did not contain fake news but it might have contained fake theories and for that you need experts to see through them.
Am I wrong?
You have a few AI posts that, in my opinion, are good examples for this discussion (in the sense that I don’t think AI generated posts are good). Examples:
- wall of text
- let me AI that for you
- back-and-forth between others, you, and your AI (and the AI was wrong)
- not exactly sure, but ChatGPT calculated something for me
And sorry, this isn’t meant as an attack
. To me, these are just examples that have caught my attention. I can’t imagine what the forum would look like if everyone did that.
Maybe I’m exaggerating a bit, that’s possible. I’d just find it a shame if this forum turned into humans discussing with AI through other humans. I prefer discussions direct with humans.
Not that long, to be honest.
That’s a great example.
Took me literally 5min on the loo
to …well, not „google“ it, to pull up the law (AHV-Gesetz on admin.ch) find the answer myself (proving wrong what AI stated with conviction).
PS: not meaning to slight @ma0 here. Also, I quickly gave up to find it on AHV/AVS own website. That said, makes me hopeful not to be replaced myself as an employee too soon.
Thanks for your feedback. No offense taken.
Well you caught my other AI post. That was kind of casual-friday post. I should be a bit ashamed of that, mostly for the way I posted (too in a hurry). but for the content it basically agrees with your points.
My post of today is IMHO something else. I define them as “conversation starter”. I think we might start to discuss also about posts where I myself answer with something along the line of “let me google that for you” or “let me search on the forum for you”., in the sense that I ask the smartest persons on this forum to help me figure out if the AI just bulshitted me or not.
So all in all I agree with you with the difference that I support using AI to start conversations or generate more questions. Not as a final answer though.
Edit: I totally forgot about the python post. That one is totally misunderstood. The AI was just to write the skeleton of the script. The algorithms comes from a forum poster.