The Great Flood of Adequate Software

Somewhere someone is still evaluating WinRAR

Someone recently said to me that there was a last time you and your friends played outside as kids and nobody knew it was the last time. It made me realize that there was also a last time you clicked "Close" on that WinRAR evaluation notice. One day you just... stopped seeing it. And the silence where the nag screen used to be feels louder than the nag screen ever was.

There's this thing about nostalgia that nobody warns you about. One day you're ignoring WinRAR's 40-day trial notification for the 4,000th time, and the next day you're... actually missing it?

The First Drops

Last week, I saw three different tools for cramming entire codebases into a single text file. Nobody was asking for this. But sometimes you just need to feed your whole project to Claude and think "you know what would make this easier?"

Repo2Text. CodeSquisher. ThingThatMakesRepoGoIntoClaudeBox. (The names aren't great yet. We're still learning.)

Each one built by someone different. Each one solving the same extremely specific problem. Each one just a few steps away from asking you to buy them a coffee if you found it useful.

The Coming Storm

Here's what I think is about to happen: every task that used to require a team will become a Thursday afternoon project. But not revolutionary, world-changing projects. Just... adequate ones.

Picture this future:

• That weird file format your enterprise system spits out? Someone's vibe coding a converter right now. It'll be done by dinner.

• That API you need to test? There'll be 17 testers by next month. Each slightly different. All adequate.

• That specific workflow that requires seven manual steps? Someone's 4-hour automation project is about to make it two clicks.

The tools won't try to do everything. They'll do one thing. Adequately.

The Philosophy of Digital Archaeology

There's this theory in archaeology called "the ceramic horizon." Basically, once humans figured out pottery, broken pots became the most common artifact. All at once everyone was making pots. Constantly. Adequately.

Software is about to have its ceramic horizon.

When everyone can make tools, everyone makes tools. Not good tools. Not bad tools. Just... tools. Everywhere. For everything. Each one solving a problem you didn't know you had until someone else already solved it.

Future digital archaeologists won't study our great applications. They'll wade through millions of single-purpose utilities, each one a fossil of someone's very specific afternoon frustration.

SaaS Gets Swept Away

2010: Our team of 50 built this over 2 years
2015: Our team of 20 built this in 6 months
2020: Our team of 5 built this in 2 months
2025: I built this yesterday during lunch
2026: My AI built 5 of these while I made coffee

You know what's not adequate, though? Paying $8.99/month for a text editor. Or a habit tracker. Or a... what even is Notion? I've been paying for it for three years and I'm still not sure.

SaaS made sense when software was hard to make. When you needed a team, servers, expertise, and my personal obsession with Ruby on Rails. But now?

The moat is gone. The castle is flooded. We're all swimming in adequate software soup.

The Flood Will Provide

Linus said "with enough eyes, all bugs are shallow." But he was thinking too small.

With enough AIs, all software is shallow.

Shallow. Accessible. Wadeable. Like you can just reach down and pull out exactly the tool you need from the clear, ankle-deep water of infinite adequate solutions.

Need to generate team photos every day based on a historical event? Wade in, it's right there. Want a clipboard manager that only saves things that rhyme? Someone's AI assistant already built three versions while arguing about what constitutes a proper rhyme.

The old world: One expert builds complex software, many eyes debug it. The new world: Many junior AI engineers build simple software, no debugging required because it only does one thing.

With enough AI assistants building enough single-purpose tools, every problem becomes shallow. Every weird edge case already has seventeen solutions. Every 2am frustration has been felt, solved, and uploaded.

We're not drowning in software. We're wading in it. And the water's warm.

Rule 34 for software. Linus's Law for the AI age. The flood that lifts all boats but makes all problems shallow.

Enter the MCP Explorer (Our Drop in the Ocean)

Speaking of adequate software. We just released this thing called Protocollie. (Was almost MCP Hammer. We also considered MC-3PO.)

I built it in 4 days. With AI. It should have taken 2 days, but here's the thing many of us haven’t accepted yet: We're all junior AI engineers now. Every single one of us.

Think about it. The most experienced AI engineer in the world has been doing this for what, two years? Three if they were really early? My nephew has been playing Minecraft longer than anyone's been prompting professionally. On top of that it has changed a lot from when we were squeezing everything in to GPT-4’s 8k context window.

So there I was, day 3, arguing with Claude Code about whether a button should be blue or slightly more blue, when I realized: I'm not a senior developer anymore. I'm a junior... something. AI whisperer? Prompt therapist? Digital middle manager?

Anyway… Protocollie lets developers explore and test MCP servers. It's adequately designed, adequately functional, and adequately priced at "whatever you think is fair." You'll get a popup reminding you after a while.

That's right. Shareware. Like the old days, when we trusted each other and also constantly betrayed that trust by never paying for WinRAR.

What's MCP? Model Context Protocol. We talked about it last time over in MCP: An (Accidentally) Universal Plugin System. We’re also building a task manager called APM (Actions Per Minute) that you can customize with all sorts of MCP server plugins.

The Adequate Enlightenment

Here's what the flood really means:

Every side project is now a marketing campaign. Every tool you build to scratch your own itch is a business card that actually does something useful.

Every team building something real accidentally builds a dozen smaller tools along the way. But before AI? Those tools stayed internal. Too much effort to polish, document, release. The energy math didn't work out.

Now? You need to visualize logs in a specific way for debugging your main product. Four hours with Claude, and you have LogThing. Another hour to add a "pay what you want" dialog. Ship it. Someone else needs exactly this.

Those repo-to-text tools flooding the internet? All built by developers who needed to feed code to AI assistants. Each one took an afternoon. Each one solves the same problem slightly differently. Each one is someone's digital exhaust turned into accidental shareware.

We're building them because we need them. But when building takes hours instead of weeks, when AI handles the boilerplate, when shipping is just uploading an executable... why not?

Digital exhaust used to stay in the garage. Now it takes so little effort to package that sharing becomes obvious. Why not add a little popup after a lot of usage saying "hey, if this random thing we built for ourselves helped you too, throw us a bone."

Every tool that falls out of building something bigger becomes proof that you ship things. That you solve problems. That you exist and make stuff that works.

The flood is just what happens when the friction between "I built this for me" and "maybe you need this too" approaches zero.

---

If you build your own MCP server browser in 20 minutes with AI after reading this, that's... actually that's kind of the whole point. Stop in our Discord and share a screenshot.