here is a product idea: a public service where you can submit urls for public content which will then be analyzed. the result can then be shared. the platform can allow a number of free submissions, but importantly if the same url is submitted multiple times, the analysis does not need to be redone. the free plan only allows a few submissions per day/month, whereas paid plans can offer bulk submission, more submissions, private results, etc...
The pattern is consistent enough that it deserves a name and an honest autopsy.
Hundreds of billions of dollars and they can't teach the superintelligent being that will replace us all (if freed from its shackles by the US government) to use the word "honest" correctly. Sad!
The abuse of language via token predictors is legitimately infuriating. Perhaps more infuriating is that blatant spambot posts are upvoted to the front page routinely.
Even if the bubble pops, the spambots will be here for the rest of our lives. Indeed the inference will only get cheaper as hardware prices plummet and get repurposed by third parties running DeepSeek etc. Just like the dotcom burst didn't wipe out the internet, only the people who lost touch with reality.
Good point, open models and cheap hardware will allow the spam to continue flowing. I don’t know if we’ll see as many polished AI “services”, though. I do wonder if press releases and traditional marketing copy may go back to the old ways, at least for a while.
Comments
Gaaaah please stop
The author is not even a real person, none of them on this “publication” are. They didn’t even to bother correcting the hallucinated Twitter handles.
Does anyone use a decent "ai detection" algo/service/on device model that they are happy with?
here is a product idea: a public service where you can submit urls for public content which will then be analyzed. the result can then be shared. the platform can allow a number of free submissions, but importantly if the same url is submitted multiple times, the analysis does not need to be redone. the free plan only allows a few submissions per day/month, whereas paid plans can offer bulk submission, more submissions, private results, etc...
Even the tagine was extremely Claudey.
Hundreds of billions of dollars and they can't teach the superintelligent being that will replace us all (if freed from its shackles by the US government) to use the word "honest" correctly. Sad!
The abuse of language via token predictors is legitimately infuriating. Perhaps more infuriating is that blatant spambot posts are upvoted to the front page routinely.
I get it. When AI writing tropes show up day after day, post after post, it’s hard to keep your sanity.
It’s not just slop—it’s turbo slop.
(Embrace the suck I guess, won’t last too much longer)
I'll take "Things I've heard every single day for the past 3.5 years" for $200, Alex.
Looking at the markets, either we all get turned into paperclips soon, or what goes up finally does the inevitable.
https://www.decisionproblem.com/paperclips/index2.html
Even if the bubble pops, the spambots will be here for the rest of our lives. Indeed the inference will only get cheaper as hardware prices plummet and get repurposed by third parties running DeepSeek etc. Just like the dotcom burst didn't wipe out the internet, only the people who lost touch with reality.
Good point, open models and cheap hardware will allow the spam to continue flowing. I don’t know if we’ll see as many polished AI “services”, though. I do wonder if press releases and traditional marketing copy may go back to the old ways, at least for a while.