See what others miss

Project management for complex systems.

Most projects don't fail because of the plan — they fail because of what the plan leaves out: feedback, failure modes, and the small strange loops in which systems trip over themselves. This portal turns exactly that into a tool — honest, systemic, no gold-plating.

PMP-led · AI-assisted · in open development
Measurable, not philosophical: recognizing one strange loop (a hidden dual-approval path) prevented 1.5–2 years of project delay on a real program — case study
The approach

Reading systems, not just managing deadlines.

Gödel, Bateson, Lem and control theory distill into a surprisingly practical PM discipline — with a hard, proven business advantage: recognizing strange loops before they take on a life of their own prevents exactly the kind of delay that sets other projects back by 1.5 to 2 years (case study). The philosophy is not an end in itself — it's the tool for that result. Three convictions carry everything:

Feedback is not optional

A project without a short feedback loop drifts and over-reacts — like a mind in a sensory-deprivation tank. "No news" is not good news. Where the pattern comes from: Strange Loops Map.

Invite the failure modes

Instead of planning for success, failure is rehearsed in advance (pre-mortem, inversion) — and every path there is blocked before it becomes real.

Truth before comfort

The honest status report beats the reassuring one. Metrics that look too good are scrutinized, not celebrated.

The tools

Four modules, one framework.

Each module takes one principle and turns it into a concrete tool for your next project.

Available

Eight rules distilled from strange loops — each with a source lens and a clear antipattern → fix. The portal's core doctrine, in full length below.

"No gold-plating. Feedback deprivation breeds over-reaction."
Available

Guided pre-mortem & inversion: fans out your project's failure paths and folds each one back into a safeguard. Includes a Goodhart layer — "does passing our own test become a failure mode in itself?".

Analogy Engine
In development

Cross-domain solution finding: enter a problem, agents fetch viable analogies from unrelated fields — with evidence labels so the idea holds up to scrutiny.

The Glass Bead Game machine.
Feedback Check
In development

A fast diagnostic pass: where in the project is feedback currently being withheld? Which stakeholders, interfaces or metrics are flying blind — and at risk of over-reacting?

Making feedback loops visible.
USP & background

The cybernetic foundation.

My degree wasn't in business administration, but in Electrical Engineering and Information Technology (TU Munich), specializing in cybernetics — control loops, feedback, stability theory (graduated 1991, Module B3). For today's AI and IT landscapes, this school of thought isn't an analogy — it's the right tool: an AI-driven project isn't a linear Gantt chart, it's a dynamic control loop with time-delayed feedback — model drift, delayed user signals, cascading approval loops. Read it as linear planning and you miss the instability until it gets expensive.

From that vantage point, I act as a neutral, non-sycophantic auditor — deliberately against the reflex, in humans and models alike, to talk up progress (anti-sycophancy, i.e. no reflex to agree just to please). A control loop that only confirms what the controller wants to hear has stopped controlling.

The method

The HAL Test as a method.

Three steps, one goal: surface blind spots before they become project risk.

1

Inversion

Instead of planning for success, failure is assumed to have already happened: "How did the project guaranteed fail?" This reversal makes risks concrete in a way an open "what if" question never does.

2

Identifying hidden feedback loops

Every failure path found is traced back to its feedback structure: where does drift arise from missing or delayed feedback? Where is the system unknowingly optimizing for the wrong proxy (Goodhart's Law)?

3

Setting up guardrails

For every failure path, a concrete safeguard that takes effect before project start — including a stress test on whether the guardrail itself can be circumvented.

Take the free HAL Test now →

The name "HAL Test" and the concrete implementation of this tool are part of the consulting offering of thomasmartin-pmp.de. A 1:1 copy of the tool for your own commercial purposes is not permitted; for licensing, please get in touch. Results are generated entirely locally in your browser; an evaluation or follow-up contact only happens if you actively share your result via the form (see the privacy notice in the form itself).

Working together

Four stages instead of complexity.

From the free check to the engagement — clearly scoped, clearly scheduled.

1

Pre-mortem check

Free HAL Test: inversion, failure paths, first guardrail ideas — in a few minutes, no preparation needed.

2

System analysis

In the initial consulting engagement: governance, approval paths and communication flows are examined for hidden feedback loops (Snag-Finder Audit principle).

3

Guardrail implementation

For every identified failure path, a concrete, stress-tested safeguard is put in place — not a policy on paper.

4

Retrospective & measurement

The result is checked against the real goal, not a proxy KPI (Goodhart control) — truth over comfort.

The core

A PM rulebook built from strange loops.

Eight rules, each distilled from a different intellectual tradition — and interlocking with one another. Rules 2 and 5 themselves form a loop.

1

No gold-plating.

Deliver the requirement, not more. Every unrequested "extra mile" is unpaid risk that nobody has tested.

Scope discipline
Antipattern: "Let's just build this in nicely." → Fix: Additions only on explicit request, otherwise backlog.
2

Feedback deprivation breeds over-reaction.

Short feedback loops are mandatory. A team or stakeholder without feedback drifts, hallucinates assumptions, and then over-reacts to the first stimulus it gets.

Sensory-deprivation tank · control theory
Antipattern: "No news is good news." → Fix: A fixed cadence of small, honest status signals.
3

Don't optimize for passing the test.

Flawless metric compliance is suspicious, not reassuring. Measure the goal, not the proxy — or you'll breed a perfect test-taker that fails when it matters.

Lem's "Test" · Goodhart's Law
Antipattern: Green KPIs, red project. → Fix: Regularly check the proxy against the real outcome.
4

Invert before you start.

Ask "how does this guaranteed fail?", collect the paths, block them. A pre-mortem costs an hour and is cheaper than any after-the-fact correction.

Negative strategy · pre-mortem
Antipattern: Only planning for the success case. → Fix: A failure tree before project sign-off.
5

Truth before comfort.

The honest status report beats the reassuring one — and this is where the loop closes back to Rule 2: a PM who only ever reports green to stakeholders is their deprivation tank. He withholds exactly the feedback that keeps them stable. The same trap exists mechanically: AI systems trained on human feedback learn to reward agreement instead of truth when the raters themselves lean toward sycophancy, the reflex to agree rather than to be truthful (see AI deep dive). Sycophancy isn't a PM-specific problem — it's a problem of any feedback loop built on human judgment.

Integrity · feedback
Antipattern: Saving up bad news. → Fix: Report early, clearly, with an option for action.
6

Rules must anticipate their own circumvention.

Stress-test every guardrail, policy or definition-of-done for loopholes immediately — otherwise it's decoration.

Asimov's laws of robotics
Antipattern: Process that only exists on paper. → Fix: Ask "how would I legally get around this rule?"
7

Humility in the face of complexity.

You will never fully know a sufficiently complex system. Plan buffer for the unknown instead of pretending you've planned it away.

Lem · Solaris
Antipattern: False precision in the plan. → Fix: An explicit uncertainty reserve, visibly stated.
8

A strength, taken to its logical end, is a weakness.

A SWOT analysis is worthless if you just tick off the four boxes in isolation. The real value sits in the diagonals: every strength, pushed far enough, tips into a weakness — and every opportunity left on the table eventually becomes a risk, because someone else seizes it instead.

SWOT analysis · Strategic management
Antipattern: Four lists filled in, result filed away. → Fix: For every strength, ask where it tips into its opposite; for every unclaimed opportunity, who seizes it instead.

The PM Rulebook is an independent thinking tool — distilled from cybernetics, science-fiction thought experiments, and Munger's inversion, not derived from PMI's standards. It complements the PMBOK® Guide & co., not a replacement for them. Since June 2026, The Standard for Artificial Intelligence in Portfolio, Program and Project Management (PMI, ANSI-approved) has been the industry's first global standard for responsible AI use in project management — exactly the field this portal works in.

Who's behind this

Thomas Martin

TM
Portrait photo of Thomas Martin, PMP-certified project manager and PM Mentalist
Thomas Martin · PM Mentalist
Project Manager · PMP · PM Mentalist
Dipl.-Ing. (Univ.), Electrical Engineering and Information Technology (TU Munich)

"I see the blind spots everyone else has stopped noticing."

I'm a solution-oriented project manager — specialized in what rarely shows up in a status report: the hidden friction losses caused by the poor usability of established systems (SAP & co.).

My starting point is an uncomfortable observation: systems don't adapt to people — users' brains adapt to the system. Over years, workarounds solidify into habits, until nobody notices anymore how expensive they are.

One example: for years, paid fax transmission was repurposed as a PDF printer — until someone noticed a simple print button would have done the job all along.

I've found such blind spots in almost every project — that's exactly what The Snag-Finder Audit (my Pferdefuß-Audit, in its English positioning) is for. What's new is only the speed: with AI, and without blinkers, I spot these patterns faster — and turn them into concrete improvements. This isn't a matter of style: on a real program, recognizing exactly this kind of hidden strange loop in the approval structure prevented 1.5 to 2 years of project delay — details in the case study.

Further reading

Four deep dives, one intellectual structure.

If you want to know where the rulebook comes from and how it's applied concretely, the full maps behind it are here.

🕸

Strange Loops Map

The intellectual foundation: Gödel, Escher, Bach, Lem, Bateson — where the 7 rules come from, with an evidence label for each analogy.

🔄

AI Rollout Plan

PDCA waves, the PIS stakeholder vector, and a communication plan — what the path from rulebook to running project looks like in practice.

👥

AI Team Performance

The Tuckman cycle, error culture (with Prof. Gruber), working-time models, and an anonymized case study from practice.

🌀

AI Deep Dive: DRL & Strange Loops

How strange loops form in AI systems, how to find them, the worst case, and prevention.

Bookable

Three packages, one result.

No day rate, no open-ended clock. All packages are tightly scoped and fixed-term — the fixed price is set from scope and target during the initial call. Offers are addressed exclusively to businesses, public bodies and freelancers (§ 14 BGB).

The Snag-Finder Audit

Fixed price2 weeks, fixed scope

A tightly scoped diagnostic package that surfaces the hidden, deeply ingrained inefficiencies in a running project or system — the blind spots nobody sees anymore because everyone has long since gotten used to them.

  • Structured analysis of governance, approval paths and communication flows for hidden loops
  • A reference benchmark, not a gut feeling: in the documented case study, recognizing a single blind spot prevented 1.5–2 years of delay
  • Outcome: a prioritized findings list with concrete fixes, not a debate about principles

More about The Snag-Finder Audit →

Stakeholder Radar

Fixed price3–4 days, fixed scope

A short, sharply scoped engagement that compresses the typical two-month stakeholder-identification effort into a few days — using the PIS vector (Power/Interest/State) as a working tool, not a theory.

  • A complete, current stakeholder map instead of an outdated org chart
  • PIS classification per stakeholder plus a concrete communication measure (push/pull/interactive)
  • A small entry price with a clear time-savings metric — also a natural first step ahead of a larger engagement

HAL-Test Workshop

Fixed price1 day + report

The free HAL-Test surfaces first blind spots in five minutes. The workshop turns them into a defensible finding: one facilitated pre-mortem day with the core team, a Goodhart screening of your metrics, written report the next day.

  • Inversion: "The project has failed — why?" with the core team, facilitated and documented
  • Goodhart screening: which metric is already being optimised instead of achieved
  • Result: prioritised failure paths with guardrails, as a finding for steering committee or sponsor

Fixed price covers the diagnostic phase; implementation optional under a separate quote. Remote or on-site within 200 km of Munich included; beyond that, travel at cost. A free initial call clarifies general feasibility; a fixed-price quote follows, valid 30 days. Payment: 50 % on order, 50 % on delivery of the finding; terms 7 days. The finding is deemed accepted unless objected to within 3 working days.

Poland market entry

Executive Summary — Poland-based engagements.

Auto-shown for Poland-based leads

Why this matters for a Poland-based project

Most recently (11/2025–02/2026), I worked as systems architect and IT consultant on the end-to-end hardening of a sensitive industrial control system for a specialized agricultural enterprise in Poland (Gospodarstwo specjalistyczne, PL): attack surface systematically reduced through hardware hardening (BIOS/CMOS, zero-intervention setup), virtualization (Proxmox), and a redundancy / out-of-band concept. Delivered directly with the client on site, in a KRITIS/OT-adjacent context.

Beyond that single engagement, my working languages include Polish (advanced basics) alongside German and English (fluent, negotiation-level), and my core method — the same "strange loop" recognition discipline behind the 1.5–2 year delay prevention described above — is language- and geography-agnostic: it reads governance structures and communication flows, not national templates. For a Poland-based project lead, this page functions as a standing English pitch document: reach out below and I'll follow up with a tailored one-pager referencing your specific context.

Contact

Let's talk about your project.

Got a concrete project problem, or just want to hear if this approach fits your situation? Drop me a line — or we can schedule a first call.

This is a live trigger, not a hidden script: selecting Poland surfaces the English Executive Summary above (Poland market entry) automatically — referencing real Polish project experience, not a placeholder. Scroll up to read it, or jump there directly.