A model handed me back a line I couldn't decide was mine. Here's what AI actually does across my newsletter process — research, brain dumps, editing, checks — and what it doesn't.
The distinction between disclosing an AI percentage and disclosing a workflow feels crucial. A surface score cannot show where judgment entered, which sources constrained the draft, or which claims were rejected. I would love to see "How I make this" evolve into a small provenance record: what the writer originated, what the model transformed, what evidence was checked, and where the human made the final call. That would reward examined practice rather than clever evasion. It also gives readers something more useful than a verdict: a way to understand how trust was actually produced.
I've appreciated your thoughtfulness and transparency and it's given me a lot to think about. Now that I'm actually writing, I'm really appreciative of it. Yesterday, I wrote a post, did the pangram analysis, and it came back "100% human." Is that saying that it sounds dumb like a human, not smart like a machine? Is my voice really that different, authentic? What does that really mean, what's the signal it's giving? I don't know, maybe good to have the conversation, but I don't think we know the answer, and like everything else in AI it's moving so fast.
"And I’ve put real work into training these tools on my own writing, creating skills, instructions, and markdown files so what comes back sounds like me instead of sounding like everyone."
Just imagine if you put the same effort into, you know, writing and thinking for yourself.
The distinction between disclosing an AI percentage and disclosing a workflow feels crucial. A surface score cannot show where judgment entered, which sources constrained the draft, or which claims were rejected. I would love to see "How I make this" evolve into a small provenance record: what the writer originated, what the model transformed, what evidence was checked, and where the human made the final call. That would reward examined practice rather than clever evasion. It also gives readers something more useful than a verdict: a way to understand how trust was actually produced.
I've appreciated your thoughtfulness and transparency and it's given me a lot to think about. Now that I'm actually writing, I'm really appreciative of it. Yesterday, I wrote a post, did the pangram analysis, and it came back "100% human." Is that saying that it sounds dumb like a human, not smart like a machine? Is my voice really that different, authentic? What does that really mean, what's the signal it's giving? I don't know, maybe good to have the conversation, but I don't think we know the answer, and like everything else in AI it's moving so fast.
"And I’ve put real work into training these tools on my own writing, creating skills, instructions, and markdown files so what comes back sounds like me instead of sounding like everyone."
Just imagine if you put the same effort into, you know, writing and thinking for yourself.