How do you compete when Claude already has everything?

I went from not knowing how to code, to learning how to code. Naturally, the next progression is to have big dreams of creating a startup, so I start thinking about niches with software needs, markets with opportunities.

But then Claude came along. Claude did two things:

1. AI coding tools like Claude Code made it easier than ever for anyone to build and copy features
2. Claude itself is such an all-encompassing platform. It has connectors, automations, document creation. What can't it do?

            Claude arrives
                  |
         +--------+--------+
         |                 |
         v                 v
+----------------+ +----------------+
| 1. Claude Code | | 2. Claude is   |
| Anyone can     | | a super app:   |
| build and copy | | Connectors,    |
| features fast  | | automations,   |
|                | | documents      |
+----------------+ +----------------+
         |                 |
         +--------+--------+
                  |
                  v
       Where's the niche now?

If features aren't a moat, what are the moats? Get me a moat

In Rob Walling's writings and podcast, he emphasizes that features can't be a moat for your product. Maybe they were before AI-assisted coding, but now people can literally reverse engineer any app. Much to my chagrin as an engineer, the code isn't even slop. Good products and codebases can be built in days, in weeks; no longer in months, in years.

In Walling's latest book, he talks about 8-10 moats that still work in the age of AI. Out of respect for his work and content, I won't give away the moats specifically, but my takeaway is that product moats boil down to trust.

1. Trust as a company and voice
2. Trust as a product

Trust as a company and voice: do people trust you as an expert and listen to the things you put out in the world. In other words this is part marketing/distribution and part credibility. Mo is a content creator I trust as thoughtful, genuine, and knowledgeable about shipping software, so I'm more likely to buy his courses and products.

Trust as a product: is your product trusted to be a part of a user/company's workflow? Products like TurboTax or Gusto serve a much-needed use case and then are trusted to solve for the use case with little oversight. When a product has real trust, it's embedded in how a user works, where it's painful to rip out. It's the dream for all SaaS products.

Back to Claude

So why is Claude a big deal? It's about how to build trust and those moats.

Building trust as a voice is difficult, because it's marketing, it's distribution. It takes time and building an audience is not as easy as whipping out your Claude Code and building a feature.

Building a trusted product has changed with Claude. Whereas in the B.C. era (Before Claude), niche software seemed more clear to build:

• "This general SaaS leader doesn't have connectors with niche software X and Y, I'll build a product with these niche connectors"

• "This general SaaS leader doesn't have these automation and document storage, I'll build a product with these features"

The post B.C. era, has me question how to compete with Claude, a super app that has:

1. The best models they can subsidize
2. Primitives for connecting any software and automating tasks
3. Great products like Code, Cowork, Design, Artifacts, and who knows what else is coming
4. Distribution and a strong brand presence

This isn't even mentioning the capital they can deploy. How does anyone building SaaS compete?

I think we can still compete by building trust as a voice. That will always be important for marketing your product effectively and sustainably. And the second way is to build trust as a product by finding wedges.

Features aren't a moat, but UX is a wedge

Wedges aren't permanent like moats (I swear I'm going to tie this analogy together). Moats are long-term defenses for your product that a competitor will have a hard time attacking. These are brand trust and product trust.

But wedges are short term advantages that get users to try your product. They prop open the door and give your product an opportunity to learn from your users and gain their trust.

With fl0w, that's all I can hope for. Fl0w is a CLI and a personal remote computer for you to run agents locally and remotely. Everything you can do on fl0w, you can also do in Claude. But what fl0w should offer is a familiar UX, where switching between CLI agents on your laptop and remote agents should feel as smooth as checking your email on different devices. The experience should feel the same and everything just works.

The fl0w remote computer should feel like your local computer, where you can customize your remote environment exactly how you want it.

The blueprint in the post B.C. era

1. Build brand trust. Programmers don't like to do marketing, but in an age where anyone can program with AI, the marketing muscle must be exercised.

2. Wedge your product. In order to build trust for you product, you first need users. UX and differentiated features are a wedge, but unfortunately anyone can copy your wedges. They'll never be moats again.

3. Keep learning and trying. That's all I can do. You can help me though, and try fl0w at https://fl0w.work. Thanks.