Welcome to The Friday Brief. 🎉

Last week was the reset.

This week, the format starts working for real.

If anything in here lands, hit reply and tell me.

If it doesn't, the unsubscribe link still works.

I'll be sorry to see you go, but your time is important.

Let's get into it. 💪

The Dopamine Trap

There's a recurring theme in the indie hacker world right now that I can't shake.

Let's call it the "Distribution Paradox."

It’s unavoidable and it's everywhere. Reddit, LinkedIn, X.

Builders are spending 40+ hours a week automating their product features but less than 4 hours automating how they actually find customers.

Built the app in a fortnight. Marketing it? No clue.

To me, the reason why seems simple.

Shipping code gives you a dopamine hit and manual outreach feels like a grind.

But in 2026, the market is too crowded for "build it and they will come" to be anything more than a fairy tale.

If you aren't building your distribution engine with the same engineering rigor as your codebase, you aren't building a business.

You're a hobbyist.

Get out there and talk to people.

Five Things Worth Your ⏱️ Time ⏱️

• Kanwas: Hit the top of Product Hunt this month with the highest vote count of any launch in May. Worth a look just to clock what's resonating right now.
• FeatDrop: Public changelog tool for founders who want a single feed for shipping updates and feedback. Solves the "where do I post this?" tax that solo operators pay every week.
• OnTheMap: Geographic founder discovery with revenue signals attached. Useful if you're trying to find operators in your patch without trawling LinkedIn for an hour.
• PostHog: Not new but worth a flag if you've never set it up. Event tracking, funnels, session recordings, and feature flags in one tool with a free tier that actually carries you past launch. The default analytics stack for indie operators in 2026.
• Turso: Distributed SQLite with a free tier that includes 500 databases and 1 billion row reads a month. Read-heavy apps barely cost anything to run on it.

Product Teardown: TypingMind

Tony Dinh built TypingMind on a premise that should have been wrong.

The players were already shipping their own chat interfaces, OpenAI and Anthropic and Google were spending billions on the underlying models, so why would anyone pay for a third-party wrapper on top of an API they could call themselves?

Three years later TypingMind is reportedly the default LLM interface inside a meaningful chunk of enterprise teams, generating millions in B2B revenue with a headcount of one.

I think it's worth looking at briefly.

One frontend to rule every model, sold by one bloke.

The bet is straightforward: models are commodities and interfaces are sticky.

If you're a team of 30 using Claude, ChatGPT and Gemini across different workflows, you don't want three separate vendor logins and three separate billing cycles and a finance team trying to reconcile usage across providers, you want one pane of glass with bring-your-own-key billing that maps to your existing API spend.

And that's the whole product, really.

Plug in your API keys, get a unified interface across every major model, share prompt libraries with your team, pay once for the wrapper. The models get cheaper every quarter, the friction of switching between them does not.

What makes it interesting is the positioning.

Dinh isn't competing with OpenAI, he's selling the layer that sits between the models and the people who use them at work, betting that this layer is more durable than any individual model's lead.

The bet's holding so far because OpenAI and Anthropic and Google keep building for consumers, not always for teams who need governance and audit trails and some way to stop a junior pasting the customer list into the wrong window.

In Case you Missed it

Mark, Moss Piglet

I had a good chat with Mark from Moss Piglet recently, covering his move from stay-at-home dad to building a privacy-first product solo. An inspirational read for any parents out there.

From Stay-at-Home Dad to Privacy-First Founder: How Mark Built Moss Piglet

Mark Thayer taught himself Elixir during his son’s nap times to build alternatives to surveillance capitalism. Six year…

We Are Founders •Chris Kernaghan

One small thing

Sending this one from the Troodos Mountains in Cyprus.

The internet's patchy, there's storm clouds rolling in over the ridge, but I've still got a question for you before I lose signal.

If you've landed on an AI workflow that's genuinely saved you more than 5 hours a week, not just a tool but an actual sequence you run, hit reply and tell me about it.

I'll feature the best one next week. Enjoy the weekend.

See you next Friday.

Chris.