A web page that tells you what your browser gave away the moment you arrived. No login, no form, no permission. Most pages do this. None of them tell you.
Vibe coded af, how has nobody spotted this. The website swears the text was written by a human, and either they have contracted chronic GPT-virus or are an LLM
edit: this is made by Rise Up Labs which is an ai psychosis company
AI is quite good at web design now, but it still has a distinct style. Claude in particular LOVES to mix serif and monospace fonts. This isn’t necessarily a guarantee based on just that, but it did trigger my alarm bells.
The second biggest thing is the language. LLMs absolutely SPAM slightly vague, short phrases separated by punctuation.
The language on each data point also is pretty repetitive which implies either sub agents were called or the model was asked individually to write something about it in a specific tone.
The final nail in the coffin was the company that made it, Rise up labs, which advertised all their AI software on their home page
One clue to me is the “how many times you moved” statement. One actual human “move” is worth hundreds of what the site calls a move. A human would notice that but the reality of it means nothing to an AI.
Secondly just the language used being quite dramatic but also generic.
You know it’s just counting the change in acceleration in your phone’s gyroscope chip or whichever it is. If you are typing something the phone “moves” twice with each swipe.
This page is just putting numbers it’s collecting from your phone into a template paragraph.
Vibe coded af, how has nobody spotted this. The website swears the text was written by a human, and either they have contracted chronic GPT-virus or are an LLM
edit: this is made by Rise Up Labs which is an ai psychosis company
How can you tell that it was vibe coded? Genuine question.
AI is quite good at web design now, but it still has a distinct style. Claude in particular LOVES to mix serif and monospace fonts. This isn’t necessarily a guarantee based on just that, but it did trigger my alarm bells.
The second biggest thing is the language. LLMs absolutely SPAM slightly vague, short phrases separated by punctuation.
The language on each data point also is pretty repetitive which implies either sub agents were called or the model was asked individually to write something about it in a specific tone.
The final nail in the coffin was the company that made it, Rise up labs, which advertised all their AI software on their home page
One clue to me is the “how many times you moved” statement. One actual human “move” is worth hundreds of what the site calls a move. A human would notice that but the reality of it means nothing to an AI.
Secondly just the language used being quite dramatic but also generic.
LLMs always write with a very dramatic tone. I really hate that high impact language now.
You know it’s just counting the change in acceleration in your phone’s gyroscope chip or whichever it is. If you are typing something the phone “moves” twice with each swipe.
This page is just putting numbers it’s collecting from your phone into a template paragraph.
Thanks! I’ll have to keep an eye out for those things.
What is a “psychosis company”?