"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.
Use AI to make things clearer, not longer. Let it sharpen your thinking, not replace it.
What bullshit is this? AI doesn't sharpen anyone's thinking. It replaces thinking with misinformation
I, umm, I sometimes write essays in slack/matrix
One idea I read is that showing others LLM output should be considered rude. If folks find the tools useful, by all means use them, but do others the service of translating the output back into verified, truthful, human signal first.
If folks find the tools useful, by all means stop.
No. It's still rude. If I want to read an AI response, I can generate crap myself. Who gave you the right to throw crap at me?
theres people that actually do this?
Plenty
Lol go open any AI related community on reddit and half the replies are chatgpt copy paste that's full of fluff and filler.
Which is even funnier because I've seen actual AI bots give better responses which makes me wonder why those users bother to generate an answer at all.
Oh absolutely, and on all sorts of communication.
I sat befuddled at what the hell I was reading when I went to read a reply to an email I had fired off at the end of the day and didn't check until next morning. Turned out he wanted to be 'cute' and so he had AI slop up his out of office reply. So instead of "Out of office, will be back on MM/DD" I had read two paragraphs of bizarre fan fiction looking junk apparently intended to make his three day weekend sound like some fantasy epic.
Someone asked me to help with something and I said sure, and then within they minute, they pasted a wall of text saying "this is what ChatGPT had to say, so it should help you get going". It didn't work for them, so why they thought I would both find it useful and couldn't have asked an LLM myself, no idea. It was totally wrong and that's why they screwed it up harder, but still thought the AI must have been right.
I've received multiple bug reports and pull requests with confusing long winded essays that bothered to bring up details that were both not relevant and yet also incorrect.
Had this coworker who wrote 1-liner "bug reports". When I told him I need more info where it happens, what happens and what the expected result should be he started to use AI to "improve" his reports. The ai slop just added more filler words and obvious statements, but nothing useful that would help me to reproduce the issue, because it did not know the software, application or code behind it.
It's probably more common in some communities than others. I hang out in a lot of AI spaces and many newcomers think it's acceptable to slop-post there just because of the topic of the community.
LinkedIn is the definition of this.
I have had conversation with people who when can't explain something difficult start sending ai messages to protect their ego. It is so easily visible, lol
Especially obnoxious when they ask for your help in the first place and keep pestering you with GenAI bullshit saying they are 'helping' you figure out how to do whatever it is they couldn't figure out with GenAI help in the first place...
We gotta start using ai;dr more.
It's kinda bananas that people think that's even ok... why even ask you? We all have access to LLMs, if I wanted an AI response I'd just talk to an AI. It's not like boomers using voice to text or something, it's more akin to setting up an out-of-office reply.
You're absolutely right! And honestly? It's not just infuriating — it's detrimental to our society's education.
Would you like me to write a succinct response instead of a wall of text?
—
I use them all thr time. Alt + 0150 or 0151 for en-dash or em-dash respectively on Windows. On mobile, it's long-press on the -, on Linux (German keyboard layout) Alt Gr (+ Shift) + -
Lol I assure you, I wrote it all myself. I intentionally used the em dash for this purpose.
Sure you did....
ChatGPT!!!
Sorry, let me restart...repeats the entire response again for no fucking reason
😡
Indubitably
A lot of people at my job started having bots auto post summaries to human channels. It’s a waft of AI garbage that’s hard to parse and people get pissy when you don’t read the whole thing.
Like if you fucking care hand type it yourself instead of vomiting garbage at me.
I get pissy when the AI tries to summerize my emails or messages when I'm driving. I just want it to parrot what my friends in the group chat have said but it keeps trying to paraphrase and I honestly don't get what is so fucking hard about letting the phone just read the god damn words that my friends wrote!
To be fair, before GenAI a lot of them managed to manually vomit a lot of useless garbage and I ignored them then too... Of course, now they are more empowered to waste my time than before...
God, I hate this so much. I used to frequently ask the senior engineer on my team for input on whatever I was working on, but he started just replying with copy-pasted Gemini responses.
Now I don't really ask him for input anymore.
Sounds like it's working as intended for him then.
Maybe, but I don't think his intent is to shut me up. I think he's just a true believer. I think he's honestly trying to be helpful with it.
At first I ignored that behavior, but now I respond asking them to refrain, it's ok to tell me they don't know or to tell me a hint they might know even if they don't know for sure, but I have access to the same LLMs they have and I don't need them to help with that. I tell them to just send me what they would have sent to the prompt, and if I need the LLM, I'll do it.
It's especially bad when they have zero idea and try to fake it with LLM.
If he's actually trying to be helpful, I think it's worth attempting a calm, earnest conversation about it with him explaining how this choice of his is a lost resource to you and how much you value his input over that of an algorithm.
It still may not work, but someone thinking they're doing the right thing is typically easier to persuade over someone who is maliciously trying to push you away.