Hello!
Today I'm thinking about a line from Elon Musk that people are buying Full Self-Driving with a car attached, not the other way around. People buying the outcome (it drives me places) instead of the tool (a car) is going to impact a lot more products, especially software. At a high level, this is how AI upends existing companies. Why does this matter so much for software? Because software teams spent decades making tools people use to get the outcomes they want. Photoshop is a good example. It has thirty-plus years of interface built around manipulating pixels, but when you want to edit a photo you're thinking "make her hair flow to the left," not "select these pixels and apply a transform." The tool sits between you and what you actually want. I've been chasing this idea for a while. Back in 2024, I wrote about how the continuous march of "more features, more UI" leaves people with more functions they don't know about and don't use, and how AI lets you control software with hundreds of features by just telling it what you want to do. Last month I pushed it further: start from the outcome ("here's the news you asked for") and let people engage with the tools later if they want to. Outcome first, tool second. The car example just makes it really concrete. Nobody wants a steering wheel, they want to get somewhere. As we've said for years, we don't call carpenters "hammer users." So if you're building software today, ask yourself: what's the "it drives me places" version of your product? Because if you don't ship that, someone else will, with your tool attached as an afterthought.
What's the biggest difference between how you thought about AI two years ago and now?
My views have shifted mostly around how you build products with this technology, and where the durable value actually sits.
Two years ago, I was watching AI go through distinct stages. First it was working behind the scenes. Back in 2016, Google Translate was "completely reinvented," but the interface itself changed not at all. They replaced separate translation systems with a single neural network, and YouTube did the same with deep learning for video recommendations. The UIs stayed the same; everything transformative was happening under the hood. I spent years at Google where the conversation was always about making machine learning more of a core part of the experience, but it never got to the point where people were explicitly interacting with an AI model.
Then the past three years accelerated everything. From behind the scenes models to chat to agents, that's how things evolved for the AI-first companies we built. Augment is the clearest example: they started with AI-powered code completions, then launched a chat interface, then an agent that pairs models with tools to get more done. Their transition isn't an isolated example.
So the first big change is that these very capable models are now accessible to anyone, anywhere. The dramatic shift over the past two-ish years is that these models are now accessible to anyone, everywhere. You can download them or run them from the cloud, and all these capabilities are showing up in so many different places. That reframed how I think about it. Instead of a scary monolithic AI brain, I now think about all the different things AI models can do as new ingredients we can use to rethink experiences and do things we couldn't do before.
But here's the part that's really evolved: where durable value lives. Two years ago you could get excited about building a slick software layer. Now? Anybody can go make software right now. So just focusing on the software piece is fragile. If you're going all the way down the stack, where each piece is expanding and complementing and depending on the others, then you have something a lot more durable and useful going forward.
And it changed how I think about roles. It used to be designers illustrate a feature in mockups, then developers sort out what code is needed. Now developers build features using AI coding agents, and designers sort out what UX is needed to ship them.
The honest truth is we're still early. When web came around, everybody copied their desktop app and put it online. I think we're still in that stage of AI right now, where we haven't really found the final form of the AI application. Everybody rewrote their mainframe app for PC, their PC app for web, their web app for mobile, and every single person is going to rewrite whatever app they have now for AI. It's kind of inevitable.
So the biggest difference? Two years ago AI was a feature you dropped into an experience. Now it's AI from the bottom to the top, and AI begets more AI. The whole stack is getting reshaped, and the best advice I've got is the same thing I keep saying: you've got to jump in.