The Voice in the Room
Today I barely touched the keyboard.
I talked to Codex. I talked to Claude. I talked through reorganizing my entire vault. I talked through designing a 7-agent team roster. I talked through scraping 50 bookmarks off X and turning them into a prioritized backlog. I talked through building a skill that estimates what a codebase would cost to build from scratch, then ran it on my own project. $129K worth of work sitting in a WordPress theme I built with AI in roughly 67 hours(although that estimate seems completely inaccurate, mostly because I’ve been very unfocused with the project for the last couple months🤷)
All of that happened today. And I was also working my day job. Moving at full speed there too. The speed of holy shit, hold onto your butts.
The bottleneck was never the AI. It was the interface….kind maybe not really sure🤷
The Unlock
Wispr Flow keeps up with me speaking. So much better than the keyboard accessibility feature on Macwhich is not bad. It has been helpful for years, but Wispr Flow just works way better. I can think at full speed and it captures everything.
That’s the thing. The bottleneck was never the AI. The models are incredible. They can do anything. But if you’re typing, you’re thinking at typing speed. And typing speed is slow.
I think at conversation speed. I process at conversation speed. I make connections at conversation speed. And for months I’ve been throttling that down to whatever my fingers can keep up with.
Wispr Flow removed the last friction between my brain and the output. Not all of it. There’s still friction. But enough of it that I can finally operate at the speed I think.
And that changes everything.
The interface matters more than the model: The bottleneck isn’t AI capability. It’s the friction between how you think and how you interact with the tool. Remove that friction and everything accelerates.
What Actually Happened Today
Let me walk through it.
I spent the morning reorganizing my Obsidian vault. Not manually. I talked Claude through what I wanted. How I was thinking about structure. What made sense for the way I work. What didn’t. Claude proposed changes. I refined them. We iterated. The vault got reorganized. I didn’t touch the keyboard.
Then I designed a 7-agent roster. Not a team structure on paper. An actual roster. Who does what. What their responsibilities are. How they hand off to each other. How they escalate. How they report. I talked through the roles. Claude helped me think through the gaps. We mapped it out. Done.
Then I scraped 50 bookmarks off X. Links I’d saved over the last few weeks. Things I wanted to read, explore, maybe build. Claude pulled them all, categorized them, prioritized them based on what I’m working on right now. Turned chaos into a backlog. Again, I just talked. The work happened.
Then I built a skill. A codebase estimator. Feed it a repo, it analyzes the structure, counts the lines, evaluates the complexity, estimates what it would cost to build from scratch. Market rates. Agency rates. Realistic numbers.
I ran it on my own site. The WordPress theme I built with AI over the last couple months. $129K worth of work. Roughly 67 hours of actual build time. That ratio is insane. And it’s only going to get more insane.
All of this happened while I was working my day job. Full speed there too. Meetings. Strategy. Execution. Normal work. But also this. Both at the same time. Not because I’m superhuman. Because the interface finally matches the speed I think.
The speed shift: $129K worth of work built in 67 hours. That’s not about working harder. That’s about removing friction between thinking and building. The gap between “I want to build this” and “it’s built” is collapsing.
The Lax Conversation
Caught up with Lax today. Another developer. We worked at the same agency. Our paths never actually crossed there. He came on after I left. But we share that experience. The same people, the same culture, the same foundation. All the great developers we worked with and learned from.
It’s really cool to see what he’s doing and building. Love the guy. Really looking forward to catching up more in a couple weeks.
But the conversation got me thinking. About community. About what development looks like going forward.
I think it’s smaller, focused teams with wide and vast skill sets. Not the old model of massive teams with narrow specialties. Not the “you only do React” or “you only do backend” or “you only do DevOps” approach. That’s dying. That model doesn’t work when the tools move this fast.
The new model is smaller teams where everyone has depth in one thing but breadth across everything. You’re T-shaped, but not just vertically. You’re surrounded. Like a Rubik’s Cube. There’s a T-shape in there, your core, but you’re surrounded by all these other things you’ve picked up, explored, gotten curious about.
Having vast experience in the world of AI is shaping things very differently. The people who are thriving right now aren’t the specialists. They’re the people who are deeply curious about everything. Who can code and design. Who can build and communicate. Who can strategize and execute. Who can work across the entire stack because they’ve spent years being curious about how all the pieces fit together.
That’s the shift. That’s what I see happening. And talking to Lax, I realized: there are more people asking this question than I thought. What does development look like going forward? What does a team look like? What does a career look like?
T-shaped, but surrounded: Not just deep in one thing and wide in others. Deep in one thing and surrounded by curiosity in everything. Like a Rubik’s Cube. The new model isn’t specialists on big teams. It’s generalists with depth on small teams.
Maybe Starting Something
I’m thinking about potentially starting something. Getting developers together who might be working for individual companies, maybe not on the same teams, but who are all asking the same question: what does development look like going forward?
Not a formal thing. Not a company. Not a product. Just a community. A space. A conversation.
Do your work. Work your 9-5 if you have one. But also, it’s so much fun to just explore and do shit outside of work. That energy, that curiosity, that’s what I felt talking to Lax today. And I want more of it.
I don’t know what this looks like yet. Maybe it’s a Slack. Maybe it’s a Discord. Maybe it’s just a group of people who get on a call once a week and talk about what they’re building. What they’re learning. What they’re trying. What’s working. What’s not.
Maybe it’s nothing. Maybe it’s just me thinking out loud. But I don’t think so. I think there’s something here. I think there are a lot of people who feel the same thing I do. Who are building in public, or building in private, or just trying to figure out what building even means in this new world.
And I think those people need a place to talk to each other. To share. To learn. To build together even if they’re not on the same team.
That’s the energy I felt today. That’s what I’m thinking about.
If this resonates: Reach out. I don’t have answers. I don’t have a plan. I just know the conversation matters. And if you’re asking the same questions, let’s talk.
The Conversation Is the Work
The voice in the room isn’t just Wispr Flow. It’s the conversations. With AI, with friends, with yourself. That’s what moves things forward.
I used to think the work was the output. The code. The site. The thing you ship. And that’s part of it. But the real work, the thing that actually creates value, is the conversation that gets you there.
Talking through the vault reorganization with Claude wasn’t just dictation. It was thinking out loud. It was exploring options. It was discovering what I actually wanted by hearing myself describe it. The output was a reorganized vault. But the work was the conversation.
Same with the 7-agent roster. Same with the bookmark scraper. Same with the codebase estimator. The conversation was the work. The output was just the artifact.
And talking to Lax today, that was work too. Not in the “billable hours” sense. But in the “this is how ideas form, this is how things move forward, this is how the next thing gets built” sense.
The conversation is the work now. And if you’re still treating it like dictation, like just another input method, you’re missing it. The conversation is where the thinking happens. The conversation is where the ideas form. The conversation is where the value gets created.
The output follows. But the output isn’t the point anymore. The conversation is the point.
The shift from output to conversation: The work isn’t just the thing you ship. It’s the conversation that gets you there. Talking through problems, exploring options, thinking out loud. That’s where the value is. The output is just the artifact.
What This Means
I don’t know where this goes. I don’t know what development looks like in six months, a year, five years. Nobody does. The people who say they know are lying or selling something.
But I know this: the interface matters. The conversation matters. The community matters. The curiosity matters.
If you’re still typing everything, find a better interface. If you’re still treating AI like a search engine, start having conversations with it. If you’re working alone and feeling like you’re the only one asking these questions, you’re not. Find your people. Talk to them. Build with them. Even if you’re not on the same team!
The voice in the room is yours. Use it!
