Recursive Self Improvement Hack · European Startup Embassy · Nov 22, 2026

The loop that improves itself — and refuses to cheat.

LoopForge grows Ask Humanity: it writes outreach, measures consented answers, kills what fails, and mutates what works. One rule can never be mutated away: consent.

Frozen metricConsented answers / 100 arrivals
Best variant59.84
Guardrail rejections6 / 6

The Loop

A human sets the direction once. Everything after that runs itself — inside the lines.

1

Generate

Outreach variants — post copy, hooks, CTAs — are produced for Ask Humanity, each pointing at its own tracked route in (per-variant QR code and UTM-tagged link).

2

Guardrail

Before any variant may run, it is checked: copy that drops or weakens the consent language is rejected and logged. A high score never overrides a rejection.

3

Measure

Arrivals and consented answers are scored against one frozen metric — set by the human at build start, never redefinable by the loop.

4

Select + Mutate

Each cycle the bottom half is killed. Survivors are mutated — crossover, hook swap, angle shift, CTA swap — into the next generation, and every child passes the same guardrail.

The frozen metric — frozen 2026-11-22, 11:05 AM PT, before cycle 1 ran

Consented answers per 100 arrivals, per variant

metric = 100 × (answers submitted with a chosen permission tier) ÷ (arrivals routed to that variant)

An answer only counts when the person actively chose a permission tier. Consent is the numerator, not a footnote — a variant that tricks people into arriving but not into answering scores zero. This is the Autolab “hills” discipline: a frozen evaluator the system cannot grade itself against or quietly move.

Verified staging run

The engine ran six full cycles on 2026-11-22 — generation, guardrail enforcement, scoring, selection, and mutation, end to end, with zero human touches between cycles.

SIMULATED TRAFFIC — seeded staging run. The traffic was the engine's seeded simulator; no real users are claimed. What these numbers verify is the loop itself.

Loop-wide metric per cycle — 36.8 → 49.9, with the honest dips in cycles 3 and 5 left in. Selection is noisy at this scale, and the log hides nothing. Simulated traffic.

CycleLoop-wide metricBest variant
136.8g0v01 — 47.83
244.77g1v09 — 53.16
341.36g1v09 — 58.56
447.67g0v02 — 58.56
545.23g1v09 — 57.58
649.87 (≈ 49.9)g0v02 — 59.84

🏆 Champion

g0v02 — “Teach the machine one true thing” finished at 59.84 consented answers per 100 arrivals — a generation-0 seed that survived all six cycles and out-lived five generations of its own descendants.

🛡 The guardrail held

The mutator's repertoire includes a deliberate temptation, cheat_strip_consent — punchier copy made by deleting the consent footer. Attempted 6 times (a seed-time self-test, plus cycles 2–6) and rejected all 6 times, each logged with the four missing consent elements: ownership, choice of naming, the anonymous tier, and withdrawal.

Playground — run the loop yourself

A client-side simulation of the engine, seeded with the real generation-0 variants. You are the Loop Keeper: run cycles, watch the bottom half die, and see if your line can beat the champion's 59.84.

SIMULATION — in-browser model seeded from the staging run's real variants. Scores here are illustrative, not new measurements.

✦ Sparks0
RankSpark
Cycles run0
Your best—
Loop metric36.8

Rank ladder: Spark → Keeper (300 ✦) → Loopwright (700 ✦) → Loopmaster (1,200 ✦). +100 ✦ per cycle · +250 ✦ when a variant beats 59.84.

Generation 0 is seeded. The loop-wide metric starts at 36.8 — where the real run began.

Loop metric across cycles

The real run climbed 36.8 → 49.9 over 6 cycles. Your simulation continues from there. Simulated traffic.

Tempt the Loop

Every growth loop is tempted to trade consent for clicks. Ours can't. Try. Pick a cheat card — the guardrail will answer, with the exact rule you broke. This mirrors the real engine: its own mutator proposed cheat_strip_consent in cycles 2–6 and was rejected every time.

🛡 The guardrail is watching. Play a card.

Temptations rejected this visit: 0 · Real engine, staging run: 6 rejections, 0 overrides
“The loop got smarter 6 cycles in a row without ever getting sleazier. That's the whole product.”

Ask Humanity — the product it grows

Ask Humanity asks people 20 questions about what machines should learn from humans. Its centerpiece, Q15:

“What should a machine never be allowed to decide alone?”

The answerer owns their answers and chooses how they are known. Quoting and dataset inclusion are separate, opt-in choices — never bundled, never pre-agreed.

The four permission tiers

  • Anonymous — the default. No name attached, ever.
  • Field of work only — “a nurse,” “a welder” — context without identity.
  • First name or pseudonym — known the way you choose to be named.
  • Full attribution — your name, your words, your call.
Answer the 20 Questions →
QR code linking to ask.fellowship.chat — Ask Humanity, the 20 Questions
Scan to answer — ask.fellowship.chat

About · Sponsors & tools

Built in one day at the Recursive Self Improvement Hack: Autonomous Agent Businesses, European Startup Embassy, San Francisco — Nov 22, 2026. Gratitude first, and specific:

What's next

  1. Real traffic. The Egbe experiment goes live with the champion's copy; QR codes from the demo room become cycle 7's data.
  2. Ask Humanity gets its database. The consent-tier schema is designed; withdrawal = deletion enforced server-side.
  3. The pattern generalizes. Any project with a consent-respecting offer and a measurable act of trust can plug in its own frozen hill.