Claim
When experts say "it's impossible," ask who said it and what is actually breaking. Most "we can't do this" claims are inherited beliefs, not constraints. Push until you reach the real boundary — a law of physics, a specific contract clause, an identified harm. Many "impossible" things turn out to be inconvenient.
Mechanism
Experts hold their domain's received wisdom efficiently — but the received wisdom is often a freeze of a prior generation's tradeoffs. When you ask "who said this is impossible?" the answer often goes "everyone knows" or "vendor X told me" — a chain that decays into hearsay. Pushing until the chain ends in an actual citation either produces a real constraint (in which case you redesign) or surfaces an opportunity nobody else has tested.
Conditions
Holds when:
- The team can absorb the time cost of investigating each claim of impossibility.
- The hire who runs the investigation has both technical depth and willingness to question authority.
Fails when:
- Inherited beliefs are correct often enough that re-investigating each one is wasteful. (In genuinely well-understood domains, this rule produces churn.)
- The "law of physics" is a regulator with discretionary enforcement. Some constraints are real because someone with power says so — even if the contract is fuzzy.
Evidence
Cash App's Cash Card story: Ayo's team tried 1,000+ combinations of plastic, overlay, paper, envelope, finishes before shipping. When experts said "it's impossible to laser-engrave at this density," he asked who said and dug in. Laser-engraving turned out to have thousands of settings nobody had explored. The result: a physical card no competitor could produce.
"Most people ask an expert and stop. But experts often have outdated or inherited beliefs. You have to keep pushing until you get to the actual law of physics or contract."
— Ayo Omojola on Lenny's Podcast, 2026-04-28
In regulated industries: "There are many things used as excuses because someone's understanding from their last job is applied here, creating a worse experience for patients."
Signals
- "Impossible" claims get tracked and investigated, not deferred.
- The team produces unexpected outputs in well-understood categories (Cash Card, no-API screenshot detection at Snap).
- Founders do this themselves at first, then hire people who do it natively.
Counter-evidence
This pattern eats time. Many startups die from over-investigating known-impossible problems instead of shipping the known-possible ones. Apply selectively: at the moments where the impossible thing is a 10x competitive advantage if true, not at every minor friction.
Cross-references
- Differentiation requires three checks: different, better, and matters viscerally to users — what the bedrock-depth investigation produces at the surface