The fire of fires

Daniel Miessler calls it the fire of fires. I read his essay. Then I pulled up my credit card bill from a year ago. I had already lived it. In one year I cut about fourteen SaaS tools. One private AI stack took their place. Here's what burned.

Warm charcoal sketch of a cluster of small glowing amber lanterns flickering out while one larger central amber fire glows steadily against a deep obsidian night sky, with thin wisps of smoke rising from the fading lanterns

What Daniel is pointing at

If you haven't read it, go read The Fire of Fires by Daniel Miessler. The core argument is simple. Models got cheap enough that a private assistant can now do what we used to rent whole SaaS products to do. Scheduling. Meeting notes. Research. CRM follow-up. Content outlines. Inbox triage. Each of those is a five-to-fifty dollar subscription. Add them up. That's your SaaS stack. That stack is what Daniel thinks is about to burn.

I think he's right. Most folks will read it, nod, and keep paying for the same tools. The gap between "I agree" and "I cleaned up my tools" is wider than it looks. I know. I spent a year on it.

What I actually replaced

A year back I ran a small service shop. My tools looked like everyone else's. A booking link so folks could grab time. A note-taker for meetings. A research sub. A CRM that pinged me to follow up. A tool that built content briefs. A tool that drafted my newsletter. Each one made sense on its own. Stacked up, they were a monthly bleed I had tuned out.

Over the last year I rebuilt my own working stack on top of a private AI infrastructure I now call the Digital Worker. Here's the honest tally of what got replaced.

Scheduling tools like Calendly or Motion. Gone. My agent reads my calendar, knows my rules, and proposes times in natural language.
Meeting-notes tools like Fathom or Fireflies. Mostly gone. I still record, but the summary, the action list, and the follow-up draft all come out of my own stack, and they go straight into my notes.
A Perplexity Pro research subscription. Gone. My agent does research for me across the open web and writes up the brief in my voice.
A content-brief tool I was paying for monthly. Gone. Same workflow, one prompt, better output because it already knows my audience.
CRM follow-up automations I had glued together with paid tools. Gone. Replaced by a small set of scripts that watch my inbox and draft the follow-up for me to review.
Form-filler and doc-generator subscriptions. Gone.
Two or three smaller tools I don't even remember the names of. I found them on the credit card statement when I was cancelling.

Rough count: fourteen subscriptions. Some were small. A couple were real. The total was several hundred dollars a month, none of which I miss.

Warm charcoal overhead sketch of a wooden workbench where twelve small amber subscription cards are being swept off one edge into darkness while a single steady amber ember glows where they used to sit — The stack I had. The stack I kept. Almost everything went into the fire.

What I didn't try to replace

I want to be careful here. The fire is real. It's also not universal.

My accounting is still in a real accounting product. My email still sits on Google's servers. My domain registrar is still a domain registrar. My phone is still my phone. The core systems of record, the places where regulators or auditors or banks expect to look, stayed where they were. The stuff that burned was the thin subscription layer on top of those systems. The glue. The reminders. The summaries. The outlines. The drafts.

I see this over and over. The SaaS world splits into two piles. One pile holds the truth. Call it the system of record. The other pile helps you act on that truth. Call it the assistant. AI is eating the second pile fast. The first pile is moving slow. For most small shops, it may never move at all.

Why this is a product, not a hobby

My background is ethical hacking. I spent years learning how to map a system I don't know, find what it's made of, and see how the parts fit. I pointed that same mindset at my own SaaS stack. What does this tool do. Where does its data live. What would I need to build to stop paying for it. Once I started treating the stack as a system to map, not a list of products to keep paying for, the replacements showed up on their own.

After about six months, the stack stopped feeling like a side project. It started to look like something I could hand to another small operator. I had done the hard part. I picked what to swap out. I picked what to keep. I wired the new pieces so they did not fight each other. That mix is the real thing people want. They do not want "a model." They want the calls already made. Which tools to kill. Which tools to keep.

That is why the Digital Worker tier is here. It is my year of burning down my own stack. I packed it into something you can set up in two weeks, not a year.

The part Daniel doesn't say out loud

The fire of fires is real, but it has a condition.

Nothing burns if you don't light it. The subs don't cancel on their own. The workflows don't move on their own. Your team built habits on the old tools. They won't switch without a push. I'm not the first to see that the SaaS stack can be swapped out. I might be the first on my block to actually do it. The only reason I did? I treated it like a project with a deadline. Not a neat idea to think about.

If you read Daniel's essay and nod, give yourself a calendar quarter. Pick three tools. Replace those three. See how it feels. You will be shocked at what you don't miss.

That's how the fire spreads.

← Back to all posts