Solo. Technical.
Zero funding. From India.
A 42-day sprint toward the real Y Combinator Winter 2027 deadline — November 2, 2026, 8:00 PM PT. Not a course. A forcing function: notice a real problem, pre-sell it, ship an ugly version, get paid, and make an honest call on submitting. Every technique below is sourced — Paul Graham's own essays, YC's own application and deal pages, the Mom Test, and 2026 cold-outreach benchmark data.
Is this idea worth a week?
Run every candidate through these five before you get attached. Four of five are directly from Paul Graham's own essay on finding startup ideas — verified, not paraphrased from a summary of a summary.
Six weeks. Sep 22 → Nov 2.
Each week is a boarding pass, not a suggestion. Tap one to open its full briefing — the technique, the framework, tonight's drill, and its sources.
Questions that can't lie to you.
From Rob Fitzpatrick's book. If you ask your mom whether your idea is good, she'll say yes — and so, without realizing it, will everyone else.
Ask this
Not this
Small, sharp lists beat big, vague ones.
2026 aggregate benchmark data across cold-email platforms — real numbers, not a guess.
Zero-edit, nightly, 15 minutes.
A phone, a basic mic, and an accent are not a handicap here — polish reads as fake. Document action, not thinking: the rejections, the bugs, the raw build.
Rotate through these weekly
Nothing about this stays vague.
Fetched directly from ycombinator.com this session — not a secondary blog's guess.
The $500K deal
What actually kills applications
Tonight's entry.
Your real archive. What you built, who you talked to, what broke, what you learned — this is what your application's traction story gets built from in Week 6.
No entries yet. Tonight's the first one.
What's load-bearing, and what isn't.
Established = verified against the primary essay, book, or official YC page directly. Verified stat = checked against named 2026 industry data. Framework = a reasonable working model, not an official position.