## Round 1: Foundation Questions **1. [[Personal Discovery]] Moment** When did you first realize that using LLMs could be a form of self-expression rather than just a productivity tool? Was there a specific moment or project that shifted your perspective? **2. Definition of Self-Expression** How do you personally define "self-expression" in the context of work? And [[how does]] using LLMs fit into or challenge that definition? **3. [[The Transformation]] of Work** In what ways has using LLMs changed [[not just]] _what_ you produce, but _how you see yourself_ as a creator/worker/professional? **4. Unique Voice Paradox** Some argue that using AI dilutes personal voice or authenticity. How do you reconcile maintaining your unique perspective while collaborating with an LLM? **5. Creative Partnership** Do you see the LLM as a tool, a collaborator, or something else entirely? How would you describe [[the nature]] of this working relationship? **6. Unexpected Discoveries** What aspects of yourself—skills, interests, or creative directions—have you discovered through working with LLMs that you might not have found otherwise? **7. The New Literacy** Do you think there's an art to prompting and working with LLMs that itself becomes a form of self-expression? Like how [[writing style]] is personal, is "prompting style" personal too? **8. Democratization vs. Distinction** How do you think about the tension between LLMs democratizing certain capabilities while also potentially making it harder to distinguish yourself professionally? Take your time with these—I'm [[looking forward]] to your transcribed thoughts! Feel free to go as deep and meandering as you'd like; the more context and nuance, the better I can understand your perspective. --- [[Nick Young]]: These are all excellent questions. Starting from the first: it was not a moment. What there was was a gradual bit like this. This is for your first question, by the way. Over the last years, I’ve used LLMs increasingly more and more. And this has coincided with an increase in the enjoyment of my work. I travel to work by train, and it’s quite a long train ride, but it’s peaceful. It’s a good time to get work done because. My mobile internet works for two-thirds of the journey, and on that two-thirds I use LLMs extensively to work and think and write. There will be another [[Substack article]] about this individually. [[Over time]] I realised that I was getting more and more eager to sit down and start—an eagerness I’ve never felt before. That’s when I thought there’s something deeply pleasurable, at least to me, about using. Moving on to your question two. What I’m talking about here is not doing testing parts of philosophy. I use LLMs for the sorts of things a lot of people use them for, which is rephrasing, rewriting—all of those slightly less creative tasks which are still essential for writing. So let’s just focus on this for a second. I think I’m enjoying it more, even these more basic tasks, because it is ordinary language that I’m using to talk to these systems. Not all these systems—what I want from them, and this is the important bit, in exactly the way I want elements. I’m not only able to specify exactly what I want them to do, but I also choose how to structure my text. Pleasure to messagecrafting think about . Pleasure to text crafting, getting it exactly right. There’s also a pleasure to prompt crafting, because a prompt can be in your, in your—fill away. What I’m getting at here is quite basic: the mental idea here—sort of art in, yeah, mundane tasks in LLMs as well as more interesting ones. And I want to say that this—or I’m very tempted to say that this—allows me to be expressive in the mundane parts of my work, in a way that I wasn’t before I could natural language As a matter of fact, looking at the rest of your questions—good as they are—I think I won’t answer any more of them right now. But can you give me some new ones based on what I’ve just told you? The above is a transcript btw, so if anything looks wrong it might be due to a mistransription. please ask if you are unsure. --- Raycast AI: # Round 2: Diving Deeper into the Mundane as Expressive Thank you for that rich response! [[The idea]] that LLMs make even mundane tasks a form of self-expression is genuinely intriguing—it's not the typical "AI helps me be more creative" narrative. Let me probe deeper into this specific angle: ## The Texture of Mundane Expression **1. The Language of Instructions** You mention using "ordinary language" to communicate with LLMs in exactly the way you want. Can you give me a specific example of a mundane task (like rephrasing or formatting) where the WAY you instructed the LLM felt uniquely yours? What made that interaction feel expressive rather than mechanical? **2. The Train Ride Transformation** You've noticed increased eagerness during your train commutes. What's the difference between, say, editing a document the traditional way versus crafting prompts to achieve the same result? Is it the dialogue aspect? The immediacy? Something else? **3. Personal Style in the Shadows** When you're using LLMs for these "less creative" tasks, do you find that your personal style still shines through in HOW you structure the requests? Like, would someone who knows you well recognize your prompting style even in mundane tasks? **4. [[The Pleasure]] of Precision** You mention [[the pleasure]] of "getting it exactly right" with prompts. Can you describe what "exactly right" feels like? Is it about efficiency, elegance, or something more nuanced? **5. Before and After** Think back to how you handled mundane work tasks before LLMs. What specific frustrations or limitations did you feel that now feel resolved? What couldn't you express then that you can now? **6. The Ordinary as Canvas** Do you think this ability to be expressive in mundane tasks is changing what we consider "creative work"? Are the boundaries blurring? I'm particularly fascinated by your insight that natural language interaction itself becomes a medium for expression, even in routine tasks. Please continue with your transcription—and yes, I'll ask for clarification if anything seems unclear! --- ## Round 3: The Freedom of Authentic Voice OR ## The Liberation of the Non-Human Audience **1. Unguarded Language** When you say you can "write exactly how I want to write" to an LLM, [[what does]] that [[look like]]? Are you more casual, more technical, more stream-of-consciousness? Can you share an example of something you'd write to an LLM that you'd naturally edit or filter when writing to a colleague? **2. The Performance We Don't Perform** With humans, we often perform professionalism, competence, or certain social roles. What aspects of professional communication feel like performance to you, and how does that performance disappear when working with LLMs? **3. Mistakes and Iterations** When you're working with an LLM and something doesn't come out right, how do you feel about iterating or correcting? Is there less embarrassment about "getting it wrong" compared to asking a human colleague for multiple revisions? **4. The Intimacy of Tool Use** Do you find yourself developing a more intimate working relationship with LLMs than with other tools precisely because you can be unguarded? Is this different from, say, talking to yourself while working? **5. Power Dynamics** There's no power dynamic, no judgment, no office politics with an LLM. How does this absence of social hierarchy affect not just what you produce, but how you feel while producing it? **6. The Real You** This might be philosophical, but: do you think the "you" that emerges when writing to LLMs is more authentic than the "you" in human professional interactions? Or is it just different? Your point about letting your guard down really resonates—it suggests that LLMs might be revealing how much energy we typically spend on social filtering in professional contexts. Please continue sharing your thoughts! --- It answers your first question. Sometimes I’m casual, sometimes I’m technical, sometimes I’m more stream of consciousness. Sometimes I correct my spelling, sometimes I don’t. Again, this allows me to be expressive as to how I am feeling or how I am thinking in the moment. that is a nice quote from a philosopher which I’ll find in a moment, describing speech as a "unique sensory imprint of a person". And I think we can use a related idea to get at what I’m talking about here. Let’s just assume we’re typing here rather than using voice. I’ll talk about voice in a differnet article. We can say that what makes a spoken sentence or a piece of typed text, a _unique_ sensory imprint, is that is in unique to that moment. that to me is expression. I’m not going to answer questions two and three of yours because they weren’t very good questions. So, moving on to four: four is an interesting one because we can think of people using certain types of what a philosopher called Lowe—spelt L-O-W-E—called utensils. You can wield utensils like knives or swords or whatever with skill, with a sort of present-to-hand— I always forget. But you know what I mean, right? And there’s a lot of interesting fine-grainedness there. I would argue that LLMs allow for a similar sort of fine-grainedness. Question five is not good either. I’m not interested in these sorts of things. Six—Is there anything interesting there? Let me see. I mean, you say this might be philosophical. This is a Substack about philosophy, logic. So good. So, yes. I mean, one way of thinking now, in answer to question number six as well: typing to an LLM is like “dance like nobody’s watching” or “dancing like nobody’s watching.” In fact, maybe the title of this piece is some sort of play on “dance like nobody’s watching” or “type like nobody’s watching.” maybe, there is a good title somewhere in this vicinity ## The Phenomenology of LLM Interaction **1. The Moment of Expression** You say what makes text a unique sensory imprint is that it's unique to that moment. When you're typing to an LLM in that particular moment—say, on your morning train versus evening train, frustrated versus flowing—how does the LLM's responsiveness to these momentary states enhance or reveal that uniqueness? **2. Skillful Wielding** Building on Lowe's utensils concept—what does it feel like when you're "wielding" an LLM skillfully? Is there a flow state? Can you describe a moment when you felt particularly adept at this fine-grained interaction? **3. The Texture of Thought** When you don't correct spelling, or when you let your stream of consciousness flow—does the LLM's ability to parse through this "raw" expression feel like it's accepting your thought in its native texture? Is there something philosophically interesting about a tool that can work with unpolished thought? **4. Temporal Expression** You mention this is about being expressive "in the moment." How does the immediacy of LLM interaction change your relationship to time in work? Are you capturing thoughts that would otherwise be lost to the editing process? **5. The Dance Metaphor** If typing to an LLM is like "dancing like nobody's watching," what's the equivalent of rhythm, movement, or choreography in this context? What makes some prompt-writing feel graceful versus clunky? **6. Beyond Tool and Collaborator** Given your philosophical framing—is the LLM something like a medium (like paint or clay) through which expression flows? Or is it something else entirely in phenomenological terms? I'm trying to get at the lived experience of this expressivity you're describing. What does it FEEL like in the body, in the moment, in the flow of work? --- "If typing to an LLM is like "dancing like nobody's watching," what's the equivalent of rhythm, movement, or choreography in this context? What makes some prompt-writing feel graceful versus clunky?" the first thing that springs to mind is that in dance there is a thing you are aiming to do, like a particular movement, and a way you can do it –think of an expert dancer making a particular move just so, and taking pleasure in how well he does it, but also the precise way in which he does it. The analog with writing. i guess, would be something like the following: I want to achieve something by writing, that is, I want to get the llm to react in a particular way, but there is also a pride in doing things correctly, and, there is a paritcualr way I could try to achieve the same result. a synonymouse phrase or word, or way of expressing the same idea, or the tone in which I put things, or how well I structure things. I guess you can probably guess what I would say for qs 1, 3, 4. based on the answer I have given here. 2 is triky for me to answer but one thing I am sure of is that I am not interested in flow states, the term has been used to mean so many things as to be useless, imo. Finally. I want your opinion on something. I want to create an over view of all of the answers I have given to every question you have asked me in the course of our conversation. I want to look at them all together so I can then draw on them when I am writing the article. My problem is the following: this sort of task is typically one that you, claude 4.1 opus, are not great at. you will tend to leave out details, or summarise, or make things more succinct –any and all of these make the output worthless. the turns of phrase in my answers are what I want to record and draw on, so every tiny little bit of detail helps. At the same time, I would like all of this information arranged in a slightly more organised form, so I can take all of this information in at a glance. Am I asking for the impossible here? why do you think previous attempts have failed, what further information would you need before I ask you to perform this task? (DON'T perform the task yet)