How to Hear Yourself Again

In 2016, Bob Dylan won the Nobel Prize in Literature for creating, according to the Nobel Committee, and I quote, new poetic expressions within the great American song tradition.

Sweet, right? The part I'm interested in focusing on today is people know when they're hearing Bob Dylan lyrics, like whether you love his voice, hate it, respect it, argue with it, imitate it badly during karaoke, you know it's him, right? So here's a real problem with finding your voice.

With AI, a lot of the conversation around writing voice is stuck on this idea of finding your voice, like your voice is like MIA somewhere, you know, a photo of your missing voice on a milk carton.

Tell me that's not an 80s reference.

But for a lot of people, especially people who have spent years writing carefully, professionally, in whatever tone and style they thought was correct, the issue is more basic than that.

They've never really developed the habit of hearing themselves in the first place.

So like you write something rough, maybe it's a paragraph, an email, the beginning of an article, a personal story you've been kind of picking away at and circling for years.

And then you hand it to AI and you ask for help, right? When I say AI, I'm talking about like the LLMs, the chatbots.

And the AI version comes back and it's smoother and, excuse me, cleaner sentences, more organized.

The grammar is fine.

Maybe it even sounds more professional on the surface.

And for a second, you're like, all right, great, that's better.

But when you read it the second time, it's more, eh, I don't know.

You don't feel that little internal recognition of, yes, that's what I meant.

That is super cool that I wrote that.

You're not connected to it.

You don't hate it.

It just doesn't create a spark.

That non-reaction to the thing that you handed to AI without any real supervision or active collaboration, that's data.

So last episode, we talked about the urge to fix the weird parts of your writing.

You know, Emily Dickinson and the problem of rigid editors, whether human or AI, trying to edit the very thing that gives writing its human signal.

The point was about protecting the meaning that you bring to the table when writing.

So for today's episode, we're talking about the ability to hear your own voice before you hand it over, to be revised, cleaned up, or quote unquote, improved.

This weird assumption that generative AI knows your writing voice better than you do.

Pretty fluent, grammatically perfect sentences aren't necessarily an improvement over you.