ExpertsFMHow To Pick A Startup Idea
Y Combinator Startup Podcast · June 17, 2026
5 Selected Lessons:
heuristic · 4 framework · 1
Burn the boats: commit to one idea and go deep
Source
Y Combinator Startup Podcast — How To Pick A Startup Idea (June 17, 2026) · at 2:06 ▸
When this applies
when you're hedging across multiple options instead of fully committing to test one
The takeaway
Concentrated effort on a single option produces clearer, faster signal than diffused effort across many.
Why it works
Splitting attention across multiple ideas produces shallow, noisy signal on each, making it impossible to tell if any idea truly works or fails.
One level up
Information quality depends on depth of commitment, not breadth of options explored.
Where it stops holding
This stops working when an idea is cheap enough to test in parallel with near-zero opportunity cost, or when early signal is genuinely ambiguous even with full commitment (requiring a true pivot test).
Detail
procedure:
- Pick one idea among the equally attractive options.
- Explicitly foreclose ('burn') the other options — stop working on them.
- Tell existing customers you've pivoted.
- Change identifying markers (company name, emails, website, internal narrative) to fully embody the new idea.
- Work with single-minded focus until you get clear signal.
benchmarks: {"identity_shift":"should feel like wearing a new skin, becoming almost unrecognizable"}
rationale: Going deep eliminates the noise of split attention and produces real signal — either validating or invalidating the idea quickly.
In their words
The solution to this is to go in depth first. If you're trying to decide between several ideas, all of which look equally attractive, pick one idea and go deep on it.
Don't seek the perfect idea or perfect founder-fit before starting
Source
Y Combinator Startup Podcast — How To Pick A Startup Idea (June 17, 2026) · at 0:00 ▸
When this applies
when facing any big decision where you keep delaying commitment because you haven't found the 'optimal' option
The takeaway
Decisions that depend on real-world feedback can't be optimized in your head — they must be tested through action.
Why it works
You can't evaluate an idea's viability in the abstract; only contact with real customers generates the feedback needed to know if it's good.
One level up
Analysis paralysis is itself a decision — usually the worst one.
Where it stops holding
This stops working when the idea requires specialized technical skill you genuinely lack and no amount of customer talk can substitute for it (e.g., a non-technical founder attempting deep DevTools).
Detail
rule: Stop trying to find the perfect idea or prove you're the perfect founder; pick one and validate with customers instead of overthinking.
threshold: If you're curious about an idea and willing to go deep and talk to customers, that's sufficient to start — years of domain experience aren't required.
rationale: Overthinking in the form of seeking a perfect idea or doubting founder-market-fit prevents the real feedback loop (talking to customers) that actually determines whether an idea is good.
In their words
The problem with this approach is that it's impossible to figure out the perfect idea in the abstract. You can only figure out what you should be working on by making contact with reality and getting feedback from customers.
The 'could you run your customer's business' test
Source
Y Combinator Startup Podcast — How To Pick A Startup Idea (June 17, 2026) · at 4:37 ▸
When this applies
when deciding if you truly understand a customer's world deeply enough to build something they need
The takeaway
Depth of customer understanding should be measured by whether you could operate in their shoes, not by a quota of interviews conducted.
Why it works
Surface-level conversations (e.g., counting interviews) don't guarantee real understanding; the test of being able to operate the customer's business reveals whether you've internalized their priorities, bottlenecks, and willingness to pay.
One level up
Measure understanding by functional capability, not by activity volume.
Where it stops holding
This stops working when the target customer's business is so idiosyncratic or regulated that outsiders structurally cannot gain operating-level knowledge without insider credentials.
Detail
rule: Ask whether you could actually run your customer's business — know their daily crises, top problems, and what they'd pay to solve them — rather than just counting how many customers you've talked to.
threshold: Being able to 'teach a class' on the customer's problem and being one of the most informed people on the subject.
rationale: True depth of customer knowledge, not interview counts, is what lets you build something customers actually need and will pay for.
In their words
The high watermark I use to help founders answer this question is: Could you actually run your customer's business?
Build at the edge of what models can do (live in the future)
Source
Y Combinator Startup Podcast — How To Pick A Startup Idea (June 17, 2026) · at 4:37 ▸
When this applies
when deciding whether to build on capabilities that barely work today but are improving fast, versus safe, fully-proven technology
The takeaway
Positioning your product at the edge of a rapidly improving underlying capability lets future improvements do work for you, as long as you deeply understand and track the bottleneck.
Why it works
Products built right at the capability frontier will improve automatically as the underlying technology improves, compounding the founder's effort with external progress.
One level up
Ride exponential underlying trends rather than building only on what's already fully mature.
Where it stops holding
This stops working when the underlying capability (e.g., model performance) plateaus or the bottleneck turns out to be something other than what you anticipated, making your product permanently broken rather than temporarily limited.
Detail
rule: Pick ideas that sit at the edge of what today's models can do, understand the bottlenecks intimately, and bet that solving the bottleneck could become the company.
rationale: This is a version of Paul Graham's quote: you should live in the future and then build what's missing.
In their words
This is a version of Paul Graham's well-known quote: You should live in the future and then build what's missing.
Pick the most ambitious version of an idea since difficulty is similar
Source
Y Combinator Startup Podcast — How To Pick A Startup Idea (June 17, 2026) · at 6:52 ▸
When this applies
when choosing between a modest version and an ambitious version of any venture that both require comparably hard effort
The takeaway
When the effort cost of two options is roughly equal, choose the option with the higher payoff ceiling.
Why it works
Because both modest and ambitious pursuits place similarly extreme demands on time and effort, the ambitious version yields disproportionately higher reward (moat, talent, market size) for similar cost.
One level up
Scope ambition up to match sunk effort, since effort is often the real constraint, not opportunity size.
Where it stops holding
This stops working when ambitious scope genuinely requires orders-of-magnitude more capital, regulatory clearance, or time than the modest version, breaking the 'roughly equal cost' assumption.
Detail
rule: Aim at the version of your idea that, if it works, rewrites a sector of the economy — e.g., regulated industries, large incumbents, or hard tech — because it costs about the same effort as a modest idea but protects you via moat, talent attraction, and market size.
rationale: The cost of pursuing a wildly ambitious startup idea and a modest one are roughly the same, since both place extreme demands on your time.
In their words
It may seem unintuitive, but the cost of pursuing a wildly ambitious startup idea and the cost of pursuing a modest one are roughly the same. They're both extremely hard.