Who Keeps the Savings?
Yes, everyone says AI is here to stay. I think the harder question is who it stays for, and the answer, written in data centers and trillion-dollar IPOs, isn't you.
Henrick F. — Writing
Long-form essays and case studies on AI, engineering, and where things are going.
Latest posts
When LLM Cost-Saving Automation Backfires: 6 Patterns I've Watched Burn Production Pipelines
Most LLM cost-saving automations make sense on a whiteboard and break in production. Here are the six failure modes I've seen most often, why they backfire, and the pre-flight checklist I use before automating any LLM cost lever.
AI Should Never Choose For You: A Buyer-Builder Case Against Agentic Hype
Most B2B buyers in 2026 don't want AI to choose for them. They want help to choose better. Here's why I bought AI as augmentation at Bedu, why I build it as augmentation at MatchWise, and what that means for any AI-native startup's GTM.
The 6-Lever LLM Cost Stack: A Production Playbook (One Backfired, One I'd Reverse Today)
Most LLM cost advice is written by people who don't pay the bill. Here's the 6-lever framework I used to cut MatchWise's per-candidate AI cost roughly 10–20×, the lever that backfired (RAG), and the architectural choice I'd reverse today.
The Rebuild Tax: A 4-Signal Decision Rule for When Internal Tools Are Worth Building
Most teams build internal software for the wrong reason. Rebuild only when at least three of four signals fire: workflow drift, vendor lock-on-data, integration fan-out, and cost-per-seat slope.