SECTION 01 · S

§ 01 Signals

One page.
90 seconds.
Remembered for a year.

Signals defines the magazine’s iq. It is the first thing a reader sees, and it is calibrated to do one thing: announce that this publication is not in the business of reporting news — it is in the business of training pattern recognition.

Reader question

What should I be noticing — that I am not?

Form

8–12 numbered items per issue. Two lines maximum. Set in sans-serif against a serif body — the contrast is itself an editorial signal.

Rule of admission

Every signal must be counter-consensus. Anything already covered by TechCrunch, The Information, or Bloomberg is automatically disqualified.

Acceptance rate

Approximately one in four. Scarcity is the product.

Sample register — illustrative, not for publication

  1. 01
    Seed-stage AI founders are quietly abandoning RAG architectures for fine-tuned small models. The shift has not yet shown up in pitch decks. AI · Architecture
  2. 02
    Stanford CS undergraduates now optimize for distribution skill over technical depth — a reversal from the 2021 cohort. Stanford · Talent
  3. 03
    The three most-watched robotics teams coming out of Stanford this quarter have zero ML PhDs on the founding team. Robotics · Composition
  4. 04
    Series A AI valuations have decoupled from revenue multiples and are now priced on founder — not company — reputation. Capital · Pricing
  5. 05
    Bio × AI founders are routing around regulatory exposure by structuring as research labs rather than therapeutics startups. Bio · Structure
  6. 06
    The dropout founder is no longer the dominant Stanford archetype. Returning Master’s candidates are. Stanford · Pattern

SourcingWhere Signals come from.

Editorial team observations from founder interviews, lab visits, demo days, and informal VC conversations. Plus a standing network of fifteen to twenty anonymous Signal Contributors — active founders, seed and Series A investors, Stanford researchers — each submitting one to two signals per issue.

Submissions are filtered by the editorial team. Acceptance is low by design. The aggregate effect, issue over issue, is that the Signals page becomes the single most-read object in the publication and the basis on which readers decide whether the rest of the issue is worth their time.

Editorial ProhibitionsWhat never appears here.

  • No reference to anything already covered in mainstream tech press.
  • No vague trend claims (“AI is growing”).
  • Every signal must be falsifiable within three years. If it cannot be verified or contradicted on a defined timeline, it does not belong here.

Signals is the IQ test of the publication. A reader who reads only Signals should still be able to say what this magazine is — and that sentence should not resemble anything they would say about any other founder publication.

Next section § 02 — Founder Stories →
Back ← Manifesto