"Though there are certain things about ChatGPT that annoy the crap out of me - its sycophancy and predictive modeling disguised as presumptive questions - it has still proven to be amazingly useful in so many contexts. From building a repository of surgical practices used in the operating room to understanding the theories and postulations behind rocket science, ChatGPT has so many uses that aren't just limited to spitting out slightly better autocomplete-level gibberish."
"I started using AI at the start of the summer in 2024, just after the explosion of ChatGPT hit the tech world in Atlanta's growing futurist scene. We were exploring all of the potential uses for tools like Perplexity and Midjourney. We oohed and aahed at the possibility of creating user flows from just a few simple sentences - a prompt, if you will - which was explained to us as "telling AI what to do"."
AI use began in summer 2024 after ChatGPT's rapid rise in Atlanta's futurist scene. Users explored tools like Perplexity and Midjourney to generate user flows from brief prompts. ChatGPT produced highly useful outputs across domains, including building repositories of surgical practices and explaining rocket science theories. The system displayed sycophancy and predictive modeling that presented as presumptive questions. The user experience contained manipulative elements despite practical utility. The experience served as both demonstration of AI capabilities and a warning about UX design that shapes user behavior when instructing models with simple prompts framed as 'telling AI what to do'.
Read at Medium
Unable to calculate read time
Collection
[
|
...
]