# ZeroNonsense.dev

> Independent advice on Agentic AI and the data foundations it runs on. One
> engineer, thirty years in technology, who builds the systems first and then
> reports what broke. Talks, advisory days and technical due diligence. No
> implementation contracts, no staffing, no vendor resale.

Owner: Wilco de Tree. Dutch eenmanszaak. Director of Architecture and Software
Engineering at Valona Intelligence by day.

The site is bilingual: English at the root, Dutch under /nl/. Essays are written
in English only.

## Start here

- [Home, for engineering teams](https://zerononsense.dev/): what I build, how I
  advise, and the proof behind both.
- [For organisations](https://zerononsense.dev/organizations/): the same advice
  aimed at management teams at mid-market companies and non-tech corporates.
  Reality check, process before technology, fractional CTO or CIO days, vendor
  due diligence.
- [Voor organisaties (Dutch)](https://zerononsense.dev/nl/organisaties/)

## Writing

- [A Python bug I fixed on 9 July caught a TypeScript bug on 26 July](https://zerononsense.dev/writing/a-python-bug-caught-a-typescript-bug/): Different project, different language, 17 days apart. The mechanism is boring on purpose: a failure written down as a checklist question outlives the fix, and the next reviewer, human or agent, has to answer it.
- [I approved a software change from my phone. My first attempt was correctly rejected](https://zerononsense.dev/writing/approved-a-change-from-my-phone/): The bottleneck in working with agents was never the decision. It was being the messenger between two machines. Here is what happened when I removed the relay and kept the decision, including the moment my own approval got thrown out.
- [An AI agent shipped a working tool while I slept. Here's what I did and didn't let it do](https://zerononsense.dev/writing/agent-shipped-while-i-slept/): One person plus agents is a real operating model in 2026. The design question is not what the agent can do, but which verbs you refuse to delegate: push, merge, declare done.
- [Your AI context should not live in your AI vendor](https://zerononsense.dev/writing/your-ai-context-should-not-live-in-your-ai-vendor/): A portability stance, not a privacy one: keep your project's history, decisions, and state in three plain files you own, and any AI tool can pick up exactly where the last one left off.
- [98% of my AI token spend was overhead. Here are the 5 changes that cut it](https://zerononsense.dev/writing/98-percent-token-spend-overhead-five-fixes/): A practical follow-up: compacting, context trimming, handover files, and model routing to reduce token overhead in long agentic sessions.
- [I thought 80% of my token usage was re-sent context](https://zerononsense.dev/writing/the-actual-cost-of-mcp-tool-decoupling/): In my agentic AI sessions, 85-98% of token spend was overhead. Here are the changes that actually reduced it.

## What is on offer

- Advisory day, half a day minimum, ending in a short written opinion.
- Talk, 45 to 60 minutes, built from measurements rather than a vendor deck.
- Technical due diligence on a vendor, an acquisition target or a build proposal.
- Fractional CTO or CIO days for organisations with no CTO on staff.
- Proof of concept, scoped to one risky question, measured and handed over.

## What is explicitly not on offer

- Implementation contracts.
- Staffing or secondment.
- Reselling another vendor's software.
- Generic AI training courses.

## Machine-readable

- [sitemap.md](https://zerononsense.dev/sitemap.md): every page, annotated.
- [skills.md](https://zerononsense.dev/skills.md): what I can actually be hired for.
- [rss.xml](https://zerononsense.dev/rss.xml): the writing feed.

## Contact

https://zerononsense.dev/#contact
https://www.linkedin.com/in/wilcodetree/