
AI should be better
Three years on from melting-eyed AI images, the technology has leapt but the taste gap has not closed. What that means if you are building a portfolio career.
The Path5 min read
“Ok fragile ego, I’m ready to be hurt again.” Me, 2026.
I’m starting a marketing challenge called everywhere30. October 1st seemed like a good place to start.
The question I want to answer is simple: can I hit a 30-day content cadence with no breaks? You’re welcome to join me, follow along and cheer, or at the very least not heckle too much from the sidelines.
This is for the build-in-public founders like me who just can’t lock in when it comes to content consistency.
The last time I tried this, I failed laughably quickly. Today I’m doing it again, and this time I’m better equipped and I’m tracking the hours.
Most creator advice says pick one channel. Some says forget that, post everywhere. That second camp made me ask myself two things. Could I ever be the person who posts everywhere? And how much time would “everywhere” actually take?
So I’m going to find out.
Seven platforms, 30 days. A cap of 15 hours a week, with a stretch goal of under 8.
The floor, every day:
Miss any of those and the day’s a fail. On top of the floor, every week:
I’ll log the hours daily and post the numbers every Friday, including the days I get it wrong.
Everyone reading this probably has a healthy or productive habit they’d like to adopt. Eating better, exercising, sleeping or reading more. Mine has always been creating more.
My plan last time was to use Distillr, a tool I’d vibe-built to pull content ideas out of my GitHub PRs and voice notes, so the work I was already doing all day would turn into posts.
In theory, lovely.
On day one it gave me 40 ideas, and they were all meh.
Picture it like this. You have four slices of toast (the content calendar). You’re incredibly low on butter (inspired material). But you don’t let that stop you, and you try to make it work (the content). You’re no fool, though, and you can see the outcome is four dry bits of toast.
Tasteless, bland, crispy bread that used a lot of words to say very little. Every post that came out of Distillr needed so much work. And not one of them passed the simple test: what’s in it for the reader? Why should they care?
So I dipped out and spent a week tinkering aimlessly with the code. If you have ADHD like me, you’ll recognise that as the most productive-feeling way to get absolutely nothing done.
That’s when I realised Distillr was missing two things:
Which made me ask: how could I build the most efficient human-in-the-loop content system out there? One made for a solo founder who wants to move like a content team of six.
If you’re building an AI app, there’s a lesson in this for you.
It put the entire infrastructure of my app into question. So I stopped building a front-facing product with UI, user journeys, auth, API keys and all the rest, and rebuilt Distillr as a folder of plain files and instructions that Claude runs on my laptop. I rebuilt the content model from first principles.
Strategy first, then sprinkle in ideas, then draft.
Getting rid of the frontend, and even the database, was a weight lifted. Now the only test is whether it gets me posting.
The early signs are promising.
I’m stress-testing a content framework I’ve mapped out. (I looked a little like that Always Sunny Charlie meme while doing it.)
The long-term goal is to see whether it can help me build a one-person media company for all my ventures, which for now means LiveDocument and Distillr. Then whether I can repackage it and share it with other solo founders, creators and builders.
Each day I’m also learning one thing about making posts good, and sharing it. I’ve worked through 19 pieces of advice so far, from people like Kallaway, Brock Johnson and Jenny Hoyos, and turned the useful parts into rules Distillr follows.
Tomorrow’s lesson is the 20 signs your post was written by AI. I found plenty of them in my own drafts.
Like most habits, you start by measuring the floor, not the idealistic end result. That means inputs and outputs only. Results like followers and views have too many moving parts to judge in a month. Those come later.
So instead of dreaming about 10,000 followers, my bar for success is this. Did I hit the floor every day? How much of the weekly plan shipped? How many hours did it take? And how many posts did I make without Distillr? The target there is zero, which will be tough by week two.
The result lands on 1 November, the day Distillr launches. Either Distillr holds up, or you get an honest write-up of where it failed.
Day one starts today. I’ll report back.
If you’re new here, I’m Cam, cofounder of LiveDocument (documents that tell you who read them), building Distillr in public. The daily log and the Friday hours will land in The Path.

Three years on from melting-eyed AI images, the technology has leapt but the taste gap has not closed. What that means if you are building a portfolio career.
The business models behind a portfolio career: the comfy number, 1000 true fans, proximity offers, flywheels, recurring versus one-off, and distribution.
On the days you are tired, sick or hungover, productivity advice fails. Pick the two or three things that move the needle, do those, and bank the win.
Join 800+ solopreneurs learning how to create their life's work.