About this note

The Ten-Year Note is a timeline for the next decade of software development and consulting (2026–2036), written by Jörn Dinkla — a software developer and consultant writing about his own professions from inside the change.

The note is priced on one thesis: the cost of producing the artifact is falling fast, while the cost of specifying, verifying, and taking accountability for outcomes stays high or rises. Every phase carries explicit confidence levels and “re-price when” triggers; the claims audit shows which earlier claims held and which needed revision, and the dashboard lists the numbers that would change the conclusions.

Method

The evidence base is a structured research pass over primary sources — METR, DORA, HBR, NBER, the Stanford Digital Economy Lab, HFS Research, Bitkom, SEC filings, and named press reporting. Vendor and marketing figures are either omitted or explicitly hedged as vendor-sourced. Contested questions (the productivity paradox, the benchmark-to-reality gap, fragmentation versus consolidation) are flagged open rather than resolved artificially. Anything past 2030 is presented as ranges and badged speculative.

Signed AI-authorship note

This site was researched, written, designed, and built with AI — openly, because that is the point. The research report behind it was compiled with AI-assisted deep-research runs against primary sources; the prose was drafted by Claude in Claude Code sessions directed, verified, and edited by the author; the site itself was generated by his website-factory pipeline (brief → design → scaffold → judged build loop, with independent AI reviewers for pixels and prose). The chapter infographics are generated with his open-source polyptych-studio pipeline and the pixbridge image library.

The division of labor is the one the note itself predicts: the machine produced the artifacts cheaply; the human specified what to build, verified the numbers against their sources, and takes accountability for the conclusions.

Jörn Dinkla, with Claude (Fable 5) · Claude Code · July 2026

Working with me

This note doubles as a work sample. If your organization is repricing its own delivery model — engineering or consulting — and wants help that is transparent about what AI can and cannot yet carry, that is exactly the practice I run.