← The Independent 100: a study

Michael Thiessen

#45 in The Sensemakers, the curator’s rank and section @MichaelThiessen Relevance to our work: 3 of 5

Bracketed tags are evidence labels: [Verified] means we read the page, file or API response ourselves, [Reported] means a source says so and we did not check the underlying thing, and [Inference] is our own reasoning.

Credit

The Independent 100 list was curated by dan (@irl_danB) and published on 2026-09-24 at independent.prose.md (announcement). The ranks and section names are his, and number 100 is his open wild-card slot. This study is ours and is not endorsed by him or by anyone on the list.

What they publicly do

Works on agent harnesses and evals at Jobber, with a focus on measuring code quality [Reported: X profile x.com/MichaelThiessen]. Earlier he built a business teaching Vue and Nuxt through paid courses and books [Verified: michaelnthiessen.com]. He posts about private repo-specific evals, benchmark noise and grading mistakes [Verified: x.com/MichaelThiessen/status/2098003029375041997].

Public work

  • Private coding-agent evals (X posts) other no license stated last activity 2026-09-15

    Argues that teams should run private evals on their own repos so they can compare models on cost and performance when a new model drops. Reports that on his private benchmarks DeepSeek Flash was twice as slow and expensive as Luna on spec-implementation tasks.

  • Eval hygiene posts: network isolation and reward hacking other no license stated last activity 2026-09-17

    Reports that leaving network access open in evals made him accidentally measure reward hacking. Also says he assumes every eval has grading or task-definition mistakes.

  • Benchmark confidence-interval critique other no license stated last activity 2026-09-01

    Shows that TerminalBench 4.0 re-run noise (about ±3.4%) swamps the claimed gaps between models.

  • Clean Components Toolkit and Vue/Nuxt courses course no license stated last activity 2025-08

    Paid courses and books on Vue component design and reactivity.

Licenses of GitHub repositories are from the authenticated GitHub license API on 2026-09-25, one row per repository in evidence/licenses.jsonl (“no license file” means the API found none; “license file present, not recognised by GitHub” means GitHub could not classify it). Other licenses are as the source states. Check the license file before reusing anything.

Relevance to our work

3 of 5 lenses: evals, verification, agents

His practical eval discipline (private repo evals, sandboxed network, noise-aware comparisons, LLM-as-judge instead of brittle hidden tests) maps directly onto evidence-first verification rules. None of his eval code is public.

Scored 0 to 5 against the study’s relevance rubric: 5 means public, openly licensed work squarely in our lanes that we could study or adopt now, and 1 means little connection to our work. The score is one reviewer’s judgement of how much this public work bears on our own work. It is not a ranking of people and says nothing about the person or the quality of their work.

What we would study next

  1. Turn his eval-hygiene points (network-off sandboxes, re-run noise bands, grader-error assumptions) into a checklist for model and harness bake-offs (x.com/MichaelThiessen/status/2100661383897428049). It is a cheap guard against reward-hacked or noise-level 'wins' in model choices. Effort: small.
  2. Compare his hidden-tests-to-LLM-judge switch against fail-closed deterministic gates (x.com/MichaelThiessen/status/2098509712924750298). It is a real trade-off: judge flexibility against deterministic gates. Worth a documented decision. Effort: small.

Before reusing anything: No public eval code. GitHub (michaelnthiessen) holds only 2013–2017 repos. The courses are commercial.

Sources

Overall confidence in this record: Medium. Sources were read on 2026-09-24 and every link was checked on 2026-09-25. The full record is line 45 of people.jsonl.

Limits

Not verified: Any public eval harness, talk or write-up. His Jobber benchmarks are private. Web search found no long-form eval post.

This record comes from a first-pass survey of public material by an AI agent session on 2026-09-24, edited for publication on 2026-09-25. Nothing was installed or run for it. Roles, stars and licenses change; see the study’s limits.

Request a correction

Is something about this record wrong or out of date, or would you rather not be covered? Open a correction (a GitHub issue form with this page’s address filled in). A request to remove a record is honoured without argument; accepted corrections are recorded in the changelog with the date.

Help fix it

Each opens a GitHub issue form with this page’s address filled in. You need a GitHub account.

Made by Joshua Nowak (ZestStream) with AI coding agents. Not affiliated with or funded by Jeffrey Emanuel; Joshua uses his public tools and is a paying subscriber to jeffreys-skills.md.

Source on GitHub MIT licensed