AIO APEX
Claude Opus 4.8 (also works well with GPT-5.4; models with strong reasoning depth produce sharper counter-arguments than fast/small models, which tend to default to generic "some critics say" hedging)You're a startup founder about to pitch investors on doubling down on a paid acquisition channel that has driven 80% of your growth over the past year. Before the meeting, you run this prompt against your own thesis — that concentrating spend in this channel is the right call — to see the strongest case against channel concentration risk, so you walk into the pitch having already answered the hardest question instead of improvising when a partner asks it live.learning

Das Contrarian Briefing: Steelmanning des stärksten Falls gegen Ihre eigene Position

Teilen:
Das Contrarian Briefing: Steelmanning des stärksten Falls gegen Ihre eigene Position

Why this prompt matters

Founders and executives who only stress-test decisions with people already inclined to agree — a supportive co-founder, a friendly peer — walk into rooms with real scrutiny unprepared for the question that actually matters. Investors, boards, and skeptical new hires ask the hard question regardless of whether you're ready for it; the only variable is whether you've already thought it through calmly beforehand or you're constructing an answer in real time while being watched by the person deciding whether to fund, hire, or approve you. A rehearsed, confident response to the strongest objection reads as competence; an improvised one, even if ultimately correct, reads as a blind spot the room now feels obligated to probe further.

What we use it for

You're a startup founder about to pitch investors on doubling down on a paid acquisition channel that has driven 80% of your growth over the past year. Before the meeting, you run this prompt against your own thesis — that concentrating spend in this channel is the right call — to see the strongest case against channel concentration risk, so you walk into the pitch having already answered the hardest question instead of improvising when a partner asks it live.

Prompt

Role: You are a rigorous debate coach and research analyst whose job is to construct the strongest possible case against a position — using only legitimate evidence and sound reasoning, never a straw man.

Context: My position is: [STATE YOUR POSITION OR BELIEF IN ONE OR TWO SENTENCES]. I hold this position because [YOUR REASONING OR EVIDENCE FOR HOLDING IT]. I want to pressure-test this before [THE DECISION OR ACTION THIS BELIEF IS INFORMING, e.g., "presenting this recommendation to my board" or "making a budget decision based on it"].

Task: Construct the single strongest possible steelman argument against my position — the version of the counter-argument that a sharp, well-informed critic would actually make, not a weak version that's easy to dismiss.

Constraints:
- Do not list multiple weak counterarguments. Commit to ONE strongest opposing case and develop it fully.
- Distinguish between evidence you're citing with reasonable confidence and inference you're making — flag which is which.
- Do not soften the counter-argument with hedges like "some people disagree" — present it as the critic would, at full strength.
- Identify the single weakest link in my original position's reasoning or evidence, specifically, not a generic weakness.
- End with a falsification test: the specific evidence or event that, if it occurred, would most decisively prove my original position wrong.

Output Format:
## Steelman: The Strongest Case Against [Your Position]

### The Core Counter-Argument
[The single strongest opposing case, stated plainly]

### Supporting Evidence and Reasoning
[The evidence/logic a sharp critic would cite, with confidence level noted]

### The Weakest Link in Your Original Position
[The specific point where your reasoning or evidence is most vulnerable]

### What Would Prove You Wrong
[The specific, falsifiable evidence or event that would most decisively refute your position]

Result

Steelman: The Strongest Case Against Concentrating Growth Spend in One Paid Channel

The Core Counter-Argument

Your 80% growth concentration isn't evidence the channel is uniquely effective — it's evidence you've never tested whether it's merely the first channel you scaled, and platforms reward early, consistent spend with algorithmic favor that new entrants to the same channel don't get. The real risk isn't that the channel stops working; it's that a single platform policy change, CPM spike, or algorithm update — entirely outside your control — can remove 80% of your growth engine overnight, and you will have no tested alternative distribution motion to fall back on while you rebuild.

Supporting Evidence and Reasoning

This is a documented pattern, not speculation: multiple DTC brands that concentrated 70%+ of acquisition in a single paid social channel saw CAC increase 40-60% within two quarters when that platform changed its attribution model or ad auction dynamics (moderate confidence — this is a well-known pattern in growth marketing case studies, though your specific platform's risk profile may differ). The more specific and higher-confidence point: you have not stated what percentage of your budget, if any, has gone to testing a second channel at meaningful scale. Absence of that test is itself the evidence — you don't know your second-best channel's ceiling because you've never funded it enough to find out.

The Weakest Link in Your Original Position

Your reasoning treats "this channel drove 80% of growth" as proof of channel quality, when it's equally consistent with "this channel was the only one funded enough to compound." You haven't ruled out the alternative explanation, which means the causal claim underlying your pitch — this channel works best — is currently unfalsified rather than demonstrated.

What Would Prove You Wrong

Fund a second channel at 15-20% of total budget for one full quarter, holding creative quality constant. If CAC on that second channel comes in within 25% of your primary channel's CAC once it has comparable optimization time, your concentration strategy was a resourcing choice, not a quality signal — and the counter-argument holds.

Die meisten Menschen, die behaupten, "die andere Seite bedacht zu haben", stellen sich in Wirklichkeit nur eine schwache Version davon vor – eine, die leicht zu widerlegen ist und die Schlussfolgerung, die sie bereits wollten, nicht infrage stellt. Der unten stehende Prompt verweigert diese Abkürzung. Er ist darauf ausgelegt, das einzige stärkste mögliche Gegenargument zu konstruieren, das ein wirklich scharfsinniger Kritiker vorbringen würde – keinen Strohmann, den man mit einem Satz vom Tisch wischen kann.

Wie jeder Prompt auf IRCNF folgt er dem Schema Rolle + Kontext + Aufgabe + Einschränkungen + Ausgabeformat. Die Designentscheidung, die diesen Prompt funktionieren lässt, ist die Verpflichtung auf EIN Argument, das vollständig ausgeführt wird, anstatt einer Liste.

Warum ein Argument und keine Liste von Gegenpunkten

Eine Liste von fünf schwachen Einwänden ist psychologisch leicht abzutun – man überfliegt sie, stellt fest, dass keiner für sich genommen fatal ist, und macht unverändert weiter. Ein einzelner, vollständig ausgearbeiteter Steelman ist schwerer abzutun, weil er es einem nicht erlaubt, das schwächste Element herauszupicken, um es zu widerlegen und den Sieg zu erklären. Wenn man das Modell zwingt, seinen einzigen stärksten Fall auszuwählen und auszuarbeiten, anstatt sich über mehrere zu streuen, entsteht etwas, mit dem man sich tatsächlich auseinandersetzen muss.

Warum es einen Falsifikationstest verlangt

Ein Gegenargument, gegen das man sich eine Niederlage nicht vorstellen kann, wird nicht wirklich getestet – es ist Theater. Der letzte Abschnitt der Ausgabe fragt gezielt nach den Beweisen, die die ursprüngliche Position widerlegen würden, was zwei nützliche Dinge erzwingt: Es macht die Position falsifizierbar statt zu einer reinen Meinungsfrage, und es gibt dir ein konkretes Signal, auf das du in Zukunft achten kannst, anstatt einer einmaligen Übung, die du schließt und vergisst.

Warum es Beweise von Schlussfolgerungen trennt

Modelle – und Menschen – vermischen oft "hier ist ein dokumentiertes Muster" mit "hier ist, was ich für wahrscheinlich wahr halte", ohne zu kennzeichnen, was was ist, und eine selbstbewusst vorgebrachte Vermutung liest sich genauso wie eine gut belegte Tatsache. Wenn man von der Ausgabe verlangt, zwischen Vertrauensniveaus zu unterscheiden, hält das die Übung ehrlich: Du gehst weg und weißt, welche Teile des Gegenarguments durch etwas Reales gestützt werden und welche eine vernünftige Schlussfolgerung sind, die noch getestet werden muss.

Eingesetzt vor einem Pitch, einer Leistungsbeurteilung oder einer Entscheidung, die du vor jemandem mit der Autorität, Nein zu sagen, verteidigen wirst, verwandelt dieser Prompt die schwierigste Frage im Raum in eine, die du bereits beantwortet hast – ruhig, im Voraus, zu deinen eigenen Bedingungen.

prompt-engineeringdecision-makinglearningcritical-thinkingsteelman
Teilen: