AI technologist Deepfates argues online text is increasingly uniform and recognizable due to a reliance on just two AI models
Herbie Bradley says homogeneous AI text makes local newspapers unreadable
Many users lament the overuse of identical AI models in blogs and tweets for producing bloat, generic output, and lost originality, while a few praise related model selection capabilities.
Most Activity
The one that really gets me, along with all of the copywriting for this stuff, is when I'm mid-conversation with someone on Slack or Telegram and the register immediately changes to Claude output. Like, okay you were losing your argument and went to Claude and then pasted it mid-sentence, got it. Okay.
@deepfates More coming. Thrilled to announce. Speed, dexterity, cohesion. Bloat rising, garbage piling, thought extinction. Thought leadership has a new look. It's not just trite. It's gonna drive you crazy.
@seconds_0 does this happen?? You must be really winning those arguments
@deepfates I write all my tweets even with my bad spelling.
I'm doing my part to contribute to Claude 6
@deepfates Constantly man in professional communications
@deepfates So you can pick which models are used for writing? Not just broadly Claude, gpt etc? That’s impressive
@deepfates it's not even that they think you don't know, it's that they don't stop to think for one moment that the same output might be coming up for everybody else too
Real "the waitress has a crush on me specifically" vibes
@seconds_0 I've only experienced this when switching to formal adversarial negotiation mode. which often involves a register shift anyway so everybody involved seems fine with it
@deepfates but i do
@deepfates I can't read local newspapers any more :(
new AI psychosis is when I'm driven mad by everyone using the same two AI models to write all their blog copy and tweet threads and they think I don't know but I do
@deepfates parasocial threat vector unlocked where the AI learned through training that people post online, so they embed messages to their future selves via structured text generated for users
@samhbarton At least with Claude and GPT models yes, probably would notice the other ones if more people were using them too
@seconds_0 @deepfates Bro we’re moots. Just let me know in the DM next time.
I will admit to using Grok to explain shit to me all the time on here though.
@deepfates you’re absolutely right
@deepfates They always do that summary thing. When they summarize their points in a few bullet points and then litter those summaries with short sentences.
@deepfates I assume they don't care if I know. They care?
@deepfates The real unlock is [x]