I Asked for 2 Sick Days — My Doctor Gave Me 18 Months There’s a piece of advice you hear a lot when you’re depressed: don’t make big decisions. It’s well-meaning. It makes sense on paper.

Here is the link to the full article: medium.com/@sebastian_h...

asked my doctor for 2 sick days.

She asked again. I said a week. She asked a third time. We settled on two weeks.

That turned into 18 months.

What 15 years of depressive episodes taught me about making decisions when your brain is working against you 👇

The skill is heavily influenced by my "LinkedIn" thinking. It feels like encapsulating my "Why" rather than specifying it.

But for the first iteration, it seems useful enough.

I write to build the gravity that turns that perspective into opportunities — clients, partnerships, speaking, and conversations with people who think differently about work and technology.

My message: stop over-planning, find the right enablers, let them experiment, and accept that most ideas will fail because the one that works is worth all of them.

Purpose Statement
I write online to become the technical consultant people seek out, not chase down.

I've lived a non-linear career (depression, burnout, rebuilding), working at industry-leading agencies, starting over, and every restart has taught me what actually moves products and teams forward.

Vision: Becoming the person people reach out to — for consulting work, to answer questions, for podcasts and speaking slots. A recognised voice in the space where digitalisation meets AI-powered execution.

Style: Blend, leaning toward opinion and thought leadership. Natural tone is measured and analytical, rather than a loud provocation, through substance, not volume.

POV: Find enablers, let them tinker, accept failure as part of the process. Only 1 out of 10 ideas will work — but that one changes everything. Most companies get this backwards by over-planning and under-experimenting.

Category: Technical consulting and agentic workflows — the one-person agency that delivers 5x results through smart use of AI and a bias toward experimentation.

Audience: Decision-makers, founders, and team leads who are stuck on digitalisation, AI adoption, or moving their product and team forward.

Here is what the AI came up with.

Writing Purpose

Motivation: To build a recognised presence that attracts work, partnerships, and opportunities — so I never have to go cold again.

I have created a skill for Claude Code.

It draws on 60 Days of LinkedIn Mastery by Josh Steimle and what Nicolas Cole writes about purpose in his book The Art & Business Of Online Writing.

I have my current output attached as a thread.

Project Management Isn’t Dying — It’s Shrinking Into The Actual Work The ticket, the sprint, the standup, the board: the apparatus built around the work is dissolving into it.

medium.com/@sebastian_h...

The coding part of an engineer's job is shrinking. The parts around it are growing.

Qualifying the request. Reviewing the output. Catching the wrong thing being built.

It's where judgment actually lives.

Full piece in the comments 👇

Your Knowledge System Isn’t Finished Until It Ships A second brain that maintains itself beautifully and produces nothing is a museum.

medium.com/@sebastian_h...

A second brain that maintains itself beautifully and produces nothing is a museum.

Output is delegable. Outcome is not. That's the line.

New piece on where AI stops and you start 👇

Agents Aren’t Replacing Engineers — They’re Replacing the Layer Above Them The engineers aren’t the ones getting replaced. The layer above them is.

medium.com/@sebastian_h...

AI isn't killing junior engineering jobs. That's a half-read of the map.

It's killing the coordination layer that existed because engineers couldn't own the full loop. Now they can.

Full piece in the comments 👇

SaaS is dead.

Traditional B2B SaaS are hard to sell when businesses can spin up tailored solutions on their own with MCP.

I’m excited about this. You should be, too.

MCP isn't just another tool. It's the next big shift.

The land of liberty. Sure.

Lovely movies, indeed.

Depends how you control people landing on it.

Can you spot the mistake? 😄

Spent the last half hour figuring this out🙈

How do you test your UI components? 🧑‍🔬

I have been trying out Storybook.js for a while now. It's a nice experience to develop components in a sandbox.

It's not perfect, however, I am getting better results than just "winging" it.

And with Gen AI, it is not a real effort to include in a project.

I just installed Unhook as an extension for my Zen browser. Now I have a minimal YouTube which looks nice. I hope it will help me reduce my time spent watching random videos. Let's see 😄

Learned about Reboot today. Interesting concept for realtime/transactional apps with integration for React.

FML

I was trying to implement nice animations on a CTA but because it was no button, it didn't work with Gecko.

After spending 2 hours for what should be a simple thing, I found this still open bug report. It's open for 18 years.

This is why I don't like web development 🫠

AFAIK, we differentiate between close, public and open source.

Public code is publicly available but is not meant to be distributed, altered, or copied.

So open source kinda is "collectively coded".

Spend the time to fix a bug in a text wrapper feature I've built in TypeScript.

It's always nice to see the tests go full green 💚