# When everything changes, what stays the same?

By Brian Casel, Nov 14, 2026

I'm kinda obsessed with change. Observing change. Living through change. Experiencing the world before and after.

I feel lucky to have seen Michael Jordan and Tiger Woods play the game. I feel lucky to have known the world pre-internet and post-internet, pre-smartphone and post-smartphone, pre-AI and post-AI. I find that perspective interesting, and useful in what I do as a product creator and entrepreneur.

> _"100% of what I do has changed from a year ago."_

I say that a lot when I talk to friends and family about what I do at work these days. It's my shorthand for the impact of AI in my world.

For **my craft**, no question. The way I do what I do today looks, feels and _is_ 100% different from the ~20 years I spent making things on the internet before AI.

**My business**, and the products and markets I operate in, are changing fast too. As an entrepreneur, it's my job to keep a finely tuned radar for new patterns, behavior changes and shifts I can learn from. That's where ideas and opportunities present themselves.

Noticing what's changed helps us acknowledge, adapt and look forward. What sometimes gets lost is staying firm on what's _not_ changing, and why.

In this post, I want to collect both.

## Design

**Design fundamentals haven't changed.** To me, designing a product or a website is understanding the problem deeply enough to design the simplest, most elegant solution.

If I'm designing a marketing site, I need to know who my customer is and what they want, so I can design the best way to show them this is their solution. If I'm designing a user interface, I need to know the job my user needs done, so I can design the most intuitive path to doing it.

**I don't design by hand anymore.** And I love it.

I always felt my strength as a designer was my eye, my taste and my obsession with _solving the problem_. My weakness was the hand-execution. I was never a sketch-on-paper guy. I could operate Figma and Photoshop, but it always felt like a laborious chore that made my wrist hurt.

Now I can lean into my role as "design director" and leave the execution to AI. My current favorite workflow is having Claude operate [Paper](https://paper.design) via MCP.

I develop specs that lay out the problem and the solution. Claude designs from those. I look them over, spot 50 things that need tweaking, fire off a few prompts and watch those iterations happen in minutes.

The real unlock is mocking up 4 completely different directions to get to the version that's just right. My projects routinely accumulate 50+ fully detailed artboards in a couple of hours.

Before, that much exploration and execution would have been unthinkable, impractical and impossible for me.

## Code

**I don't write code by hand anymore.** Obviously.

Now I don't even review the code, and I don't care about code style. Before AI, so much of my style and approach as a programmer was driven by "maintainability." Now that everything is a prompt or spec away, that's a non-concern.

I've even stopped needing to learn how the frameworks under my codebases work. I've built Mac apps in Electron and React without learning to write either by hand.

**I still obsess over architecture.** Like design, the architecture of my products (not the code) was always the part that mattered most to me. It still is.

While the industry races to "vibe code" new apps in a day, I prefer to take it slow and consider the architecture deeply before I let my agents loose on the code. I spend weeks, round after round, writing specs, rethinking and deciding every architectural detail before I feel ready to build.

Architecture and design are 90+% of the work now. Building the result is the final step, and it takes a day or two.

## Marketing

**Marketing basics have changed.** SEO, AI visibility and other recurring, repeatable marketing activities can largely be delegated to agents with real jobs now.

That doesn't mean we humans are out of the loop on these things. It does mean we no longer need to hire freelancers, agencies or employees to make sure these jobs get done.

By the way, that whole "agents with jobs" thing is something I'm obsessed with at the moment. I started building [agentreports.io](https://agentreports.io) to explore it even further.

**The need to stand out hasn't changed.** To me, standing out isn't about being the loudest, the spammiest, the spendiest or the most persistent.

Standing out is about being different. And being authentic. If your authentic self (or the human impulse underneath your product) is naturally different and attractive to people who see the world the way you do, then _that's_ what makes all the other marketing work… work.

And that's still hard. And rare.

## Products & opportunities

**"Is there a product here?"** That's an increasingly difficult question to answer.

There are plenty of tools worth building. I build new ones to power parts of my business all the time now.

But the tools (or problems worth solving) that also have potential as products to bring to market feel like a slimmer slice of the pie, with murkier risks and unknowns.

I'm constantly asking:

> _"Why would my product be better than customers building their own?"_

And

> _"Will this problem even exist a year from now?"_

Before, it felt fairly straightforward: identify a known product category, find a subset of customers (a niche) and design a more specialized or differentiated take on that category. Now it's an open question whether any given category will exist in the near future.

Finding those product opportunities requires retuning our radar. Entire product categories are going away, new ones are coming online, and others are reshuffling in ways that are hard to predict.

The shapes of products are changing too. We went from "SaaS that can replace a spreadsheet" to "software that wraps an AI model," and now to products that let customers ["bring their own agent"](https://x.com/CasJam/status/2119557120093757817).

One way I'm staying tuned in, and giving my product radar more surface area, is intentionally working with more of my people. So I've officially opened my [coaching](https://buildermethods.com/coaching) and [project advisory](https://buildermethods.com/project-advisory) services at Builder Methods.

They're my way of getting exposure to more people across more industries and problem spaces. By working with them directly on their businesses and projects, and helping them solve immediate challenges in their business and their relationship with AI, I'm more likely to spot patterns and openings I wouldn't have seen otherwise.

So it comes back to the thing that still hasn't changed:

When you're faced with unknowns, talk to more people.

---

This is the Markdown copy of https://briancasel.com/what-stays-the-same.
