Any discussion about AI/LLM’s/etc is incredibly complicated. I could go on and on elaborating on this, but I’m just going to leave my preface at that.
There is one thing I found to be true over and over again no matter what the anchor point is for the conversation, no matter the context, no matter someone’s sentiment, etc: nobody likes to have their time wasted.
LLM’s are incredibly useful for cutting corners. It makes it very easy to waste people’s time. No matter how useful they are, no matter the use case you have found, no matter the integration, people keep encountering bad search results and people sending them clearly LLM-generated work that wastes their time.
Unless somebody comes up with a cure for that, there will always be a significant portion of the population that is hostile to LLM’s - and rightfully so! No promise of productivity will overcome that.
TL;DR: the biggest problem with LLM’s is that it enables people to waste other people’s time.
Yeah, I waste more time now because of AI generated slop.
I see a headline on reddit and click on the article. I spend 3 minutes reading it and the article says nothing new and doesn't even have a point.
Then I go to the comments and see people arguing about it but also people calling it AI generated slop.
Then I click on the account that posted the article and their account history is spamming the same article to 10 different subreddits and I subscribe to about 5 of them so I have to see the same dumb post 5 times. Then I realize the account is just an AI bot spamming AI generated articles to generate upvotes or display ads to generate money.
It's all so annoying and a waste of electricity, bandwidth, and my time.
There are also subreddits mainly for photos like nostalgia places where bots just repost popular posts from a year ago or 6 months ago but now they repost stuff from 3 days ago. It's completely out of control and AI has made it this easier to do.
So now I am spending less time on these sites which is maybe a good thing so I can thank AI for making using the Internet so annoying that I stop and find something else to do.
It's basically Dead Internet Theory becoming reality.
There was always content that wastes people's time because people have always confused length and complexity with comprehensiveness and depth.
These were always poor proxy metrics for "good content," but in a lot of environments, especially professional ones, they were how work was evaluated. Naturally others used LLM to generate content that satisfies these metrics.
The slop epidemic is a consequence of what people erroneously valued for so long. Now they have it, and it's meaningless, and even if most of it was always meaningless, they can't easily tell the difference between "fluff with something meaningful" and "fluff with only fluff" anymore.
But now we're "democratizing" wasting people's time. If the AI-boosters have their way, we won't even be able to have good conversations about something as simple as the movies we saw over the weekend. It will all be "bespoke, AI-generated content." The conversations will be the equivalent of telling a story about a weird dream you had last night.
The solution was always not to view wasting people's time as proof of effort. But we did, and now AI is replicating it, and the result is this dysfunction.
If we properly valued conciseness over complexity and didn't insist on 5 paragraphs of polite fluff in business communications, it wouldn't be nearly as bad.
The issue is we have let the attention economy overrun us. It’s not people being too verbose. Shorter content is the objective these days.
You are more consistently rewarded, typically financially, for putting out 100 terrible blog posts than 10 great ones. The terrible ones tend to be short too.
Your time is saved letting ChatGPT write all your emails. The recipient’s time is wasted by emails with little to no substance that the sender couldn’t even bother to spend time on.
Hell there are people using LLM’s to comment on forums now. That’s not what I come here for. This has always existed to some degree but it has, as another person pointed out, been so democratized that it’s becoming a bigger problem. Now I have to constantly expend mental effort questioning if I’m talking to actual people. When I discover it’s a bot, it’s infuriating. It’s disrespectful, frankly.
We all came here to talk to each other. Not be duped by facsimiles of each other. I want to talk to you.
There is one thing I found to be true over and over again no matter what the anchor point is for the conversation, no matter the context, no matter someone’s sentiment, etc: nobody likes to have their time wasted.
LLM’s are incredibly useful for cutting corners. It makes it very easy to waste people’s time. No matter how useful they are, no matter the use case you have found, no matter the integration, people keep encountering bad search results and people sending them clearly LLM-generated work that wastes their time.
Unless somebody comes up with a cure for that, there will always be a significant portion of the population that is hostile to LLM’s - and rightfully so! No promise of productivity will overcome that.
TL;DR: the biggest problem with LLM’s is that it enables people to waste other people’s time.