Joined a new AI-native company this week and it’s kind of wild how different it feels already. The laptop arrived, I logged in, and an agent basically took over from there. It set up my dev env, pulled repos, fixed dependency issues, got permissions approved, pointed me at the backlog, linked the architecture docs, and surfaced the Slack debates I actually needed to read before touching production. When I needed context on something, I asked the agent and it found the exact thread from months ago explaining why a decision was made, who owned it, the related Linear issues, and the PRs connected to it. I’ve only been here 3 days but it honestly feels like I’ve worked here for a year because the usual friction and scavenger hunt for context just isn’t there anymore. We should probably stop calling this “onboarding” and rename it to “mounting” because this feels a lot more like mounting a distributed filesystem called “institutional memory” than slowly getting drip-fed context over 6 months.

Muse is cool, I'm a fan... but what's been on my mind lately is less about tech. We've created two big inventions over the last few thousand years: societal and physical. Democracy, markets, institutions, culture, shared values vs cars, planes, fire, etc... AI is about to massively level up individual capability. massively. This Muse post is exactly it. It's all about the individual. We can likely make anything physical soon. The best brains + dollars on the planet are pilling on. But when individuals become dramatically more powerful, more autonomous, and less dependent on each other, what new societal inventions do we need? What replaces the institutions, rituals and communities that evolved for a world where humans needed each other much more than they may in the future? Or is this "build me a society upgrade with..." prompt running in a backroom somewhere?

Had to give the agents working hours. Turns out 24/7 autonomy was bad for their and my morale. I’m so obviously the bottleneck. But who cares. eNPS is through the roof. Human in the loop was a trap. I finally have some peace.

Hiya @saragordic this is awesome! Thanks for sharing. I've been using for the day and have a few fun improvements. All PRs in github.com/jmwind/rooms/pull… for you to check out. - 3x faster room switching - window previews - custom resizing - kb only access to everything - edit access for rooms (add windows, name...)

Spot on. So what replaces conferences about language and frameworks which has been the default for 30 years? I'm thinking we call it "After Code" what software engineering becomes when writing code is no longer the primary job. There's tons to learn, I see it daily with everyone doing it differently with vastly different outcomes. Would you come? What should we call it?

It felt surreal a few weeks ago to see my friend Doug help rename Lake Ontario “Lake America.” Yeah. That's him behind the President. Doug once taught me one of the best lessons I've ever learned about leadership and communication. It helped me a ton and I've repeated it for more than a decade. Long before Doug Burgum was the U.S. Secretary of the Interior, he was the chairman of our board. He'd built Great Plains Software (outside of silicon valley, which was exactly what we were doing), scaled it, and eventually sold it to Microsoft for $1.1B. It’s an ERP rebranded as Microsoft Dynamics and still used around the world. I have a soft spot for people who build companies and products that outlast themselves. When he came to visit us in Sydney, we peppered him with questions about scaling companies, culture, leadership. He gave a great talk at our company all hands. He was smart about business, but particularly good at the people stuff. I got some time alone with him a few times. We shared a taxi and beers in SF. During one of those chats he asked: “JML, what's your top priority?” I told him it’s definitely communication. We had a few big cultural changes underway and I was spending a lot of time trying to make sure they actually landed across the company. “That's a good top priority.” Then he asked: “How do you know when your comms have landed?” I started trying to answer. He didn't give me very long (he does have the gift of the gab which I don't). Then Doug said, almost like he'd told hundreds of people before: “When your words are repeated back to you and the person telling you doesn't know they came from you.” Whoa. That landed. I've repeated that advice to founders, execs and teams ever since. I've applied it myself over and over. It's so simple, but hard to do in practice. Communication is hard, and that advice always reminded me to do more of it than my logical brain told me I wanted to. It didn’t come naturally to me. And here's the funny part. I never told anyone it came from Doug. Until now. His words were literally being repeated by people who had no idea they came from him. Exactly as advertised. I don't even know if Doug came up with it. He never told me. Maybe someone told him. Fast forward more than a decade. Doug is now the U.S. Secretary of the Interior. And I'm watching this thoughtful, serious guy I knew stand behind the President, smiling, while he renames Lake Ontario “Lake America.” As a Canadian, it's a tad strange thing to watch. I've been thinking about what this means to me for the last few weeks. We're all complicated humans. Someone can do something you completely disagree with and still have taught you something important. A good idea doesn't become a bad one because you disagree with the person who gave it to you. And maybe Doug's advice was never really his anyway. Maybe someone gave it to him. He gave it to me. I gave it to hundreds of other people. Some of them probably passed it along too. Nobody owns it anymore. What I've been asking myself is this: If I sat down with Doug today, would I be as open to his advice as I was back then? I'm not sure. It's easy to listen to people we admire. Much harder when they've done things we strongly disagree with. I'd like to think I would. Because after all these years, maybe that's the part of Doug's advice I hadn't understood yet. Good ideas will come from anywhere. I just hope I'm still open enough to hear them when they do.

How are ppl using Muse for different types of work? In grok bot I have a few different helpers with specialized roles. It feel like a team. With muse I have one assistant who’s a Swiss Army knife? Muse is better in a lot of ways, but I prefer having a team.

This aged well

Touching grass. Inspired by @graceongrid style pics.

A GPU-rendered wind animation at 60 fps with zero JavaScript per frame. Thanks @wcandillon for Skia. I think this is native?

Team offsite was perfect. Nobody got drunk, overshared, no awkward smalltalk, and no one posted to LinkedIn afterward.

The last 3 weeks have felt like the most productive and fun of my entire 30-year career. I’ve always had way more ideas in my head than I could actually build. That part has never been the problem. Instead of just vibing harder with agents and congratulating myself because Claude wrote 40000 lines of code, I took a step back and tried to build a team of agents for a real product. Agents that know what we’re trying to do from a product and technical perspective. Know my taste for good architecture and what it can enable and why, wrote a product taste brief, goals, where we are going, how we monitor production, etc... this took changes to our codebase, a lot of our manual tasks had to be automated much better. A lot of "we just know" had to be written down. Which was a win anyhow. We discussed how to catch dumb ideas before they become production incidents. We had to write a lot down to start. Just like building a team, there was a lot of upfront alignment, and then microalignment. The goal was the factory, not just code. So you do have to know when to take a step back on pure product output to invest in the team. Sounds familiar? And this is real production code, with real users. Not a todo app with six agents arguing about the button colour. Looking back, the difference between that and developers still single-threading prompts now feels archaic. Giving the workbench dev an agent doesn’t really solve it either. Now they just have a pneumatic hammer and are extremely proud of how fast they can hit things. I was surprised at how much the investment in the code base with tools like dependency-cruiser, run a full test loop locally, e2e tests, helped keep the quality of our product increasing. The weirdest part is once you’ve spent a few weeks building factories, it gets really hard to work beside someone at the bench. They’re asking you which hammer to use. You’re trying to redesign the assembly line. I’ve just engineered myself out of ever having a normal software job again... companies that are still organized around the workbench are dying a lot faster than they realize.

Oops forgot to do one of those free sneaker give aways. What was I thinking!

The finale was @offline and I appearing in the NYT in a 3 page spread about our sneaker twitter shenanigan. It’s framed in my office, my grandkids will have a chuckle.

sneaker twitter woke up today... from the queue. missed you guys 😂