Philosophy
Philosophy & Approach
Everyone promises AI revolution. I deliver reality checks. Here's what I believe — and more importantly, what I've learned works.
Philosophy
Everyone promises AI revolution. I deliver reality checks. Here's what I believe — and more importantly, what I've learned works.
Foundation
I've spent 30 years building things. Starting companies, scaling teams, shipping products. Learning what works through doing — and learning what doesn't through expensive failures.
These convictions aren't theory. They come from carrying problems all the way to production, from training developers who had to own every line, from telling clients what wouldn't work before they spent money on it.
Personal accountability creates everything else. Accept ownership and the rest follows — the end-to-end thinking, the deliberate friction, the sharper judgment. One foundation, four principles, 30 years of shipping production systems.
The foundation
Personligt ansvar
“You own every line of code. AI is just the tool.”
Accountability can't be delegated. Consequences create judgment — AI has no skin in the game. The moment you accept 'I own every line,' everything else follows. It's not one principle among many. It's the foundation that makes the others exist.
◆ Read the insight →
Skarpare tänkande
Because you must validate what you own
Everyone sells AI as 'think less, do more.' Reality is the opposite. You stop writing and start defining, validating, deciding. The shift goes from builder mentality to reviewer mentality — and that's cognitively more demanding, not less.
◆ Read the insight →
Begränsningen
Because the model solves what you ask — not what you forgot to ask
An LLM processes what is in the context window — nothing more. The rounding error that charges customers three cents too much on every invoice, every month, for three years: no model finds it unless you know it exists. Understanding this limitation is what makes context architecture necessary and domain knowledge decisive.
◆ Read the insight →
Kontextfönstret ÄR verktyget
Because you're responsible for output quality
The real skill isn't prompting — it's what you fill the context window with. Your code, your ideas, your transcripts, your domain data. The more unique content you feed, the more tailored the result. Input curation, not prompt engineering.
◆ Read the insight →
Kunskap förstärks, den ersätts inte
Because all of the above compounds your expertise
Senior + AI = 4-5x productivity. Junior + AI = faster mistakes. AI is a catalyst, not a replacement. It doesn't make unknowing people knowing — it makes knowing people faster. The professional identity transforms: from code writer to system conductor.
◆ Read the insight →Value-based engagement focused on outcomes, not hours. Real conversations about real challenges. Hands-on implementation alongside your team. Knowledge transfer that builds lasting capability.
“Ansvar är inte en del av processen — det ÄR processen.”
I write about these ideas in Swedish — personal reflections on accountability, validation, voice, and what one person can deliver.
Read the insights →