There are two camps within AI. There’s the folks who sleep with one eye open, convinced that robots are plotting to steal their jobs. And then the ones who see them as their new BFFs.
Our low-key obsession with AI should be a clear indicator of where we stand on machines taking over.
But we’re not blind to its limitations. Even the smartest algorithms can't handle nuance or use common sense.
We thought we had it all figured out, though. We’ve been working with ChatGPT for a couple of months. We had carefully trained conversations about our business plan to mimic embedding—the method you use to customize the LLMs’ knowledge. It was pretty cool, predicting our every move and correctly responding to all our prompts.
We should have waited with the cartwheels, though.
A recent OpenAI update caught us mid-air. Turns out ChatGPT has a 3,000-word memory limit—looks like it forgot what you said earlier.
So much for hacking a shortcut to our digital soulmate.
About Susan Rov
Susan traded cocktails and lattes for words that move. As a copywriter on a mission to end boring, she brightens brands and makes an impact. Number 463 on the list of fiction readers left standing, you'll find Susan at sunrise—running, writing, and wondering how to change the world next. That drive led her to become a founding team member of BRIL.LA.