AI as a tool to lower the inertia of getting started
I was talking to a friend on Twitter about how large language models (LLMs) have helped me get work done in a roundabout way:
From a writing perspective, I’ve not known where to start sometimes so I’ve prompted a LLM to write what I needed. Naturally, it was crap so I then just rewrote it completely. I got the job done. Had I started from 0, I may have struggled for longer.
And then, by chance, I found an article about a cool web app that lets you create 19th century ornamental tile illustrations. The last paragraph was the part that struck a timely chord (with the key phrase in bold):
Many of the initial drafts of this tool were at least partially generated by ChatGPT 4. They never came close to anything you could simply use, but they were instrumental in showing me what’s possible. More importantly, they greatly lowered the inertia I would have to overcome to start building something like this in the first place. We live in interesting times.
I love that idea; AI as a tool to lower the inertia of getting started. It seems counterproductive to prompt a LLM to write something, only to completely rewrite it and yet it’s more effective than starting at a blank screen. And I think that comes down to not knowing how or where to start, seeing something that just isn’t right and improving on it.
It’s a Ship of Theseus for the digital age.