Writing is a particular way of thinking. Talking is another particular way of thinking. Writing (and talking) are not separate, distinct acts that we do once the thought is done being thought. It's not like we sit quietly and think.., think.., think.. Then we suddenly go "Done!", and then proceed to verbalise or write the completed thought, after which we return to quietly thinking the next one. ------- Dear Jesus, It looks to me like what you're doing is an emerging format for which we may not have a name yet. For lack of a name, I'll call it *"transcribing+editing+publishing".* 1.transcribing a recording of a piece of verbal communication, 2. AI-led editing of that written transcription and then 3. publishing that edited transcription as this new thing we don't have a name for yet.  What *"transcribing+editing+publishing"* is NOT is writing. There doesn't seem to be *any* writing *anywhere* in this 3-step process. There's *thinking* in this process, for sure: - there's the type of thinking that you do when you think words in your head by yourself - there's also the type of thinking that you do when you speak words out loud to other people, but there's no _writing_ - which is the type of thinking that you do when you write words down. Now, there may be some writing done in the steps before step 1 (transcribing). But you don't mention that in your blog post. Consider mentioning it? So.. how do you prepare your talks? How have you prepared the concrete 15 or so talks you mention in your post? Is there any writing involved during that prep phase? Do you jot anything down? Do you use speaker notes in the deck? Crucially, do you use AI for _that_ prep process? If so, how? Do you use AI for the _writing_ bits (if you have any) of that prep process? If so, is there _any_ writing at all that you do yourself, without AI? If there is, then _that_ is the "fire" of "what's running inside you" that the Plotnik quote mentions. That is the "communication" that you could have AI edit (with or without AI). Editing a transcription is not editing a piece of writing, it's editing a piece of speaking. The thinking you do in your head and the thinking you do in your verbal communication and the thinking you (would) do in your writing return very different outputs. Your writing self is NOT your speaking self. What would come out of you if you _would_ write, will never come out in any other way. If you were to try to _write_ your thoughts instead of speaking them, that will result in, I promise you, a very different story. You don't have to give us that story. You don't _have_ to write, obvs. But know that you also can not _replace_ writing or _outsource_ writing with editing transcriptions of live talks or with anything else for that matter. You mention when you describe the Voice Profile in your blog post that "You feed it a few of your transcripts or writing samples, and it figures out how you actually talk." How you talk or how you write? Feeding it transcripts will maybe teach it how you talk, yes. I would not mix it with the "writing samples" (whatever these are. what are they btw? this is important. your masters thesis? your emails? your.. prompts?) You also mention first trying to "Just tell AI to write it in my style." when you describe the bad idea. What style? The style of your speaking? If you meant writing style, then again, the question is, from what would it infer that style? The example of slop you then give doesn't You've told me when we spoke in the kitchen a while back that you struggle with writing and that you start writing and you cringe cause that's not you; that's not how you sound in your head and that's not how you sound when you speak to people or when you give a talk, you told me. I think the reason for that is because you don't have a "writing voice" (yet). You only get your writing voice by writing. And I know you like your speaking voice, you mention it in the blog as well, you love talking. And people love talking to you and listening to your talks (myself included). So why aren't you happy with the raw transcription then? That IS your real speaking voice. Your literal speaking voice. Why wouldn't it work written down? I'll tell you why. There clearly must be something that needs to happen *to write in a way that reads as if the writing was spoken*. If that's the job you wanted to achieve with Verbatim, I don't think you achieved it (at least not to me, Jesus is not in there and I prefer listening to the orginal than reading the transcript). But if the job is to make sure that your spoken thoughts are recorded in written form to not be lost, Verbatim is doing it! (and has a very cool name). You explore and build stuff and then you talk about the stuff you explore and build and then you transcribe-write the stuff you talk about. This is enough. You don't have to *write*-write.