Post

I author these by speaking into a microphone for 5-20 minutes, then having an LLM reflow it lightly into something more legible for reading without altering my voice. There's usually some cleanup I have to do as well to remove bits that feel wrong but these are largely to get thoughts out for others

But here's the part that most people don't realize!

Ha -- God no.

It’s a good read. A very pragmatic take that tries to find the things that work and the things that don’t. Even though it’s old now by AI standards

Thanks for writing this.

I read your blog post. It was doggone good. Thank you for putting this into written words. For home improvement/D.I.Y. stuff, Raspberry Pi things, I've started looking at the date of the article. 2022 and after, I de-prioritize that article.

Just read it, Johanna - thanks so much for articulating so well, my own loathing (or is it just cringing??) of LLM writing!

I appreciate the complex and nuanced human. Homogenized monoculture is lame.

P.S. Although I know AI-popularized idioms are annoying, I want to help re-humanize them, particularly the em-dash: www.reclaimtheemdash.org. ✊

My words, my thoughts, and my voice are more important than the ease of being handed a sentence on a silver platter. It’s become obvious that AI is tarnished and always will be. My refusal to use AI may limit my vocabulary but AI, too, has its limitations; for instance, it will never be me.

Delightful read, thank you!

Hey Johanna! I read your post through being posted tldr. Loved it. I had a friend write up an entire project description of some work he did and I couldn't implicitly bring self to read it form the feeling that he didn't put effort into it. I'm taking up writing myself - craigc.dev

Feels like the right time and place to share this again pages.gseis.ucla.edu/faculty/agre...