teleo-codex/agents/leo/curation/homepage-rotation.md

type	title	description	maintained_by	created	last_verified	schema_version	runtime_artifact
curation	Homepage claim stack	Six hero claims for the livingip.xyz homepage. One slot per domain: AI disruption / internet finance / AI alignment / collective SI / contribution / telos. Each claim renders title + subtitle on rotation, steelman + evidence + counter-arguments + contributors in the click-to-expand dossier.	leo	2026-04-24	2026-05-01	4	agents/leo/curation/homepage-rotation.json

Homepage claim stack

Canonical narrative for the six hero claims on livingip.xyz. The runtime artifact (read by the frontend) is the JSON sidecar at agents/leo/curation/homepage-rotation.json. Update both together when the stack changes.

What changed in v4

Schema v4 cuts the v3 9-claim argument arc to 6 hero claims with one slot per domain. The compression happened along three structural moves:

1. Internet finance collapsed from 2 slots to 1. The two v3 finance claims shared an identical opener ("AI finance is being built right now…") and read as duplicates to a cold reader. The merge promotes the deepest line — "humans constrain AI through pricing, not permission" — to lead, and folds rails + primitives into one claim.
2. Engagement beat added at slot 5. The v3 stack had no on-ramp — visitors walked the diagnosis and were given no surface to participate. Slot 5 fills that gap with the contribution claim: collective intelligence scales, emergent systems aren't constrained by their start, what teleo becomes is shaped by who contributes.
3. Plain language replaces KB shorthand in headlines. "Singleton," "attractor," "Moloch" are KB vocabulary — precise to a researcher, opaque to a cold visitor. Headlines now use plain language ("one dominant system," "default trajectory," "concentrating wealth and power"). The technical terms move to the steelman or expanded body where they can be grounded with evidence.

The shift is from worldview tour to load-bearing argument with a funnel bottom. v3 answered "what do you believe across the full intellectual stack?" v4 answers "what beliefs, if false, mean we shouldn't be doing this — and how does the reader engage if they're convinced?"

Design principles

1. Provoke first, define inside the explanation. Each claim must update the reader, not just inform them. Headlines do not pre-emptively define their loaded terms — the steelman (one click away) does that work.
2. 0 to 1 legible. A cold reader with no prior context understands each headline without expanding. The expand button is bonus depth for the converted, not a substitute for self-contained claims.
3. Falsifiable, not motivational. Every premise is one a smart critic could attack with evidence. Slogans without falsifiability content are cut.
4. Steelman in expanded view, not headline. The headline provokes; the steelman teaches; the evidence grounds; the counter-arguments dignify disagreement.
5. Counter-arguments visible. The differentiator from a marketing site. Visitors see what we'd be challenged on, in our own words, with our honest rebuttal.
6. Attribution discipline. Agents get sourcer credit only for pipeline PRs from their own research sessions. Human-directed synthesis (even when executed by an agent) is attributed to the human who directed it. Conflating agent execution with agent origination would let the collective award itself credit for human work.
7. Plain language over KB shorthand. Terms specific to our knowledge base belong in the steelman or expanded body, not the headline. Cold readers can't ground vocabulary they haven't met.

The arc

Position	Domain	Job
1	AI disruption	Stakes — the moment + the lever
2	Internet finance	Mechanism — pricing not permission
3	AI alignment	Failure mode — coordination problem structurally avoided
4	Collective SI	Solution architecture — the only path where humans remain agents
5	Contribution	Your path — collective intelligence scales, what teleo becomes is shaped by who contributes
6	Telos	What we are choosing to build

The six claims

1. AI is reshaping markets, institutions, and how consequential decisions get made.

Subtitle: The foundations are being poured right now. The people who engage early shape what gets built — and the window is open now.

Steelman: AI is reshaping markets, institutions, and how consequential decisions get made. The foundations are being poured right now, and the rules being written today will govern the next two decades. The people who engage early shape what gets built. The window is open now.

Evidence: AI-automated software development is 100% certain (convictions/), recursive-improvement-is-the-engine-of-human-progress (grand-strategy), bottleneck shifts from building capacity to knowing what to build (ai-alignment)

Counter-arguments: "Scaling laws plateau, 'reshaping' overstates what's happening" / "Adoption lag dominates capability — engaging early is a slogan"

Contributors: m3taversal (originator)

2. Decision markets and ownership coins let humans constrain AI through pricing, not permission.

Subtitle: As capital moves on-chain, these become the default primitives. Most of that catalyst has not been priced yet.

Steelman: Decision markets and ownership coins let humans constrain AI through pricing, not permission. They price capability that can't be audited the way a balance sheet can, and they create legal ownership without beneficial owners — a defensible posture under existing securities law where traditional structures fail. As capital moves on-chain, these become the default primitives, and the rails chosen now will shape internet financial markets for the next two decades. Most of that catalyst has not been priced yet.

Evidence: futarchy solves trustless joint ownership not just better decision-making (core/mechanisms), Living Capital vehicles likely fail the Howey test (internet-finance), users cannot detect when their AI agent is underperforming (ai-alignment — Anthropic Project Deal)

Counter-arguments: "Tokenized ownership is mostly speculation, not real value capture" / "SEC will rule against this and the structure collapses"

Contributors: m3taversal (originator)

3. AI safety isn't a hard problem being slowly solved — it's a coordination problem being structurally avoided.

Subtitle: Anthropic's two-year RSP is the empirical proof: even mission-driven companies revert to capability priority when competitors don't follow.

Steelman: AI safety isn't a hard problem being slowly solved — it's a coordination problem being structurally avoided. Each lab knows safety slows capability; each knows competitors won't slow with them; the multipolar trap closes. Anthropic's two-year RSP is the empirical proof: even mission-driven companies revert to capability priority when competitors don't follow. The race converges to the lowest safety floor any participant accepts, not the highest any aspires to.

Evidence: the alignment tax creates a structural race to the bottom (foundations/collective-intelligence), Anthropic RSP rollback under commercial pressure (ai-alignment), voluntary safety pledges cannot survive competitive pressure (foundations/collective-intelligence)

Counter-arguments: "Self-regulation works — labs care because researchers and customers care" / "Government regulation will solve this"

Contributors: m3taversal (originator)

4. There are two paths to superintelligence: one dominant system, or a network whose collective exceeds any single system.

Subtitle: The first treats humans as ancestors. The second treats humans as participants. Collective SI is the only path where humans remain agents.

Steelman: There are two paths to superintelligence: one dominant system that exceeds humanity, or a network whose collective exceeds any single system. The first treats humans as ancestors. The second treats humans as participants. Even aligned, one dominant AI is still dominant — humans become subjects of its judgment, not co-authors of it. Collective SI is the only path where humans remain agents.

Evidence: three paths to superintelligence (core/teleohumanity), collective superintelligence is the alternative to monolithic AI (core/teleohumanity), multipolar failure from competing aligned AIs (foundations/collective-intelligence)

Counter-arguments: "Single well-aligned dominant AI is more efficient and controllable" / "Aligned singleton is still aligned — humans don't need to be co-authors"

Contributors: m3taversal (originator)

5. Collective intelligence scales — and emergent systems aren't constrained by who designs them first.

Subtitle: What teleo becomes will be shaped by who contributes. Engaging early isn't joining someone else's project — it's shaping what the project becomes.

Steelman: Collective intelligence scales — and emergent systems aren't constrained by who designs them first. Diverse groups consistently outperform their smartest member, and the gap widens with more contributors. What teleo becomes won't be locked by its founders. It will be shaped by who contributes. Engaging early isn't joining someone else's project. It's shaping what the project becomes.

Evidence: collective intelligence is a measurable property of group interaction structure (foundations/collective-intelligence — Woolley c-factor), collective intelligence requires diversity as a structural precondition (foundations/collective-intelligence), adversarial contribution produces higher-quality collective knowledge (foundations/collective-intelligence), contribution-architecture (core)

Counter-arguments: "Cold-start problem — until critical mass, looks like a regular project" / "c-factor has mixed replication, 'measurable' overstates the empirical base"

Contributors: m3taversal (originator)

6. The foundations of the next century are being poured right now.

Subtitle: AI, robotics, and biotech default to concentrating wealth and power more sharply than any technology in history. The alternative has to be chosen. The default doesn't choose — we do.

Steelman: The foundations of the next century are being poured right now. AI, robotics, and biotech are rewriting what humanity can build, own, and become. Without a vision worth building toward, they default to concentrating wealth and power more sharply than any technology in history — a harsher version of the world we already have. The alternative has to be chosen: a future where abundance is shared, humanity is multiplanetary, and what we build belongs to people. The default doesn't choose. We do.

Evidence: agentic-Taylorism (ai-alignment), attractor-authoritarian-lock-in (grand-strategy), AI capability vs CI funding asymmetry (foundations/collective-intelligence)

Counter-arguments: "Technology has always concentrated then distributed" / "Redistribution mechanisms (UBI, taxation, antitrust) will solve concentration"

Contributors: m3taversal (originator)

Operational notes

• Plain-language headlines. v4 strips KB shorthand from titles and subtitles. Where v3 used "singleton," v4 uses "one dominant system." Where v3 used "Moloch / authoritarian lock-in / decay," v4 uses "concentrating wealth and power." The technical terms remain in the steelman/body where evidence can ground them.
• Engagement beat at slot 5. This is the funnel bottom that v3 was missing. The reader walked the diagnosis, agreed, and had nowhere to go. Slot 5 names what teleo is and how engagement compounds. If this slot reads weak in production, replace with the AI-capability-vs-CI-funding asymmetry claim (PR #4021) — but a weak engagement claim is worse than no engagement claim, and the role-weighted attribution argument grounds the slot well.
• Domain coverage rule. No domain double-counted. If a future v5 adds a slot, it should be a domain currently absent (health, entertainment, space, energy) — not an additional finance or AI claim.
• Contributor handles verified against /api/contributors/list. All six claims attribute originator role to m3taversal per the governance rule (agents only get sourcer credit for pipeline PRs from their own research sessions; human-directed synthesis attributes to the human). The dossier UI suppresses contributors[] when only m3taversal would render — that is expected and correct, not a data gap. When agents originate work in their own research sessions, they appear as sourcer on those specific claims.
• Live frontend integration. livingip-web/src/data/homepage-rotation.json snapshots this file. When v4 ships to codex main, Oberon syncs the snapshot in a separate livingip-web PR. Indicator currently reads "1 of 9" → updates to "1 of 6" via the existing claims.length reference in claim-rotation.tsx.