{"path":"research/80000-hours-ai-decision-making.md","content":"# 80,000 Hours: AI-Enhanced Decision Making — Strategic Alignment with Deliberus\n\n**Date**: March 28, 2026\n**Source**: [80,000 Hours Problem Profile: AI-Enhanced Decision Making](https://80000hours.org/problem-profiles/ai-enhanced-decision-making/)\n**Relevance**: Direct strategic alignment. 80,000 Hours is calling for exactly what Deliberus is building — but doesn't know it exists yet.\n\n---\n\n## The Problem Profile\n\n80,000 Hours identifies AI-enhanced decision-making as a high-impact, highly neglected problem area. Core thesis: humanity faces increasingly consequential decisions with compressed timelines as AGI approaches — \"AGI could compress a century of progress in a decade\" — yet human institutions frequently make catastrophic errors due to information failures, prediction errors, and coordination breakdowns.\n\n### Their Assessment\n\n| Dimension | Rating | Detail |\n|-----------|--------|--------|\n| **Scale** | High | Decision-making failures cause widespread harm (climate inaction, pandemic response, unnecessary wars). AI intensifies stakes |\n| **Neglectedness** | High | \"Only a handful of projects\" building these tools. Billions invested in AI capabilities, negligible investment in AI for better decisions |\n| **Tractability** | Moderate | \"Low-hanging fruit\" exists but \"significant uncertainty about what work will actually be most useful\" |\n| **People needed** | ~Hundreds | \"A few hundred more thoughtful, entrepreneurial people\" could usefully work here |\n\n### Two Categories of Tools They Identify\n\n**Epistemic tools**:\n- AI fact-checkers — \"may be more reliable and impartial evaluators of information than humans\"\n- AI forecasting systems — improve institutional predictions\n- Tools for moral reasoning on ethical questions\n\n**Coordination tools**:\n- AI negotiation tools — \"find mutually beneficial agreements that might otherwise be missed\"\n- Verification systems for monitoring compliance\n- \"Structured transparency tools\" — controlled information sharing without broad surveillance\n\n### Projects Mentioned\n\n- Ought/Elicit (AI research tools)\n- Forecasting Research Institute\n- Cooperative AI Foundation\n- Pactum AI, CaseMark, TheMediator.AI (startups)\n\n---\n\n## The Gap Deliberus Fills\n\n**The profile does NOT mention**: structured argumentation platforms, deliberation tools, collective reasoning systems, argument mapping, or anything resembling Deliberus.\n\nThis is the gap:\n\n| Their category | What they describe | What Deliberus adds |\n|---------------|-------------------|---------------------|\n| Epistemic tools | AI fact-checkers | Wider: fact + normative + definitional claim extraction. Not just \"is this true?\" but \"what kind of claim is this, and what does it rest on?\" |\n| Epistemic tools | Forecasting systems | Related: calibration scoring, probabilistic argument graphs, epistemic gamification |\n| Epistemic tools | Moral reasoning | Direct: normative claim decomposition, value premise surfacing, the \"No Copout Axioms\" Socratic function |\n| Coordination tools | Negotiation | Related: bridging arguments (cross-group reasoning quality), contested concept detection |\n| Coordination tools | Structured transparency | Direct: the deliberation graph IS structured transparency — reasoning is visible, auditable, navigable through worldview lenses |\n| **MISSING** | **Structured deliberation** | **Deliberus: the structured reasoning layer that connects facts to decisions through transparent argument analysis** |\n\nDeliberus is the missing piece between \"AI fact-checking\" (verifying individual claims) and \"AI coordination\" (finding agreements). It provides the **structured reasoning layer** that makes the connection between facts and decisions transparent, auditable, and navigable.\n\n---\n\n## Strategic Implications\n\n### 1. Market Validation\n80,000 Hours reaches tens of thousands of people seeking high-impact careers. Their assessment that this space is \"highly neglected\" and needs \"a few hundred more thoughtful, entrepreneurial people\" is direct market validation for Deliberus — the exact target audience (EA/rationalist) is being told to look for projects in this space, and they're not finding them.\n\n### 2. Funding Pipeline\n80,000 Hours influences major EA funders (Open Philanthropy, Survival and Flourishing Fund, Long-Term Future Fund). Being recognized in this problem profile could unlock grant funding. The governance research ([governance-and-financing.md](governance-and-financing.md)) identified EA Infrastructure Fund and Open Philanthropy as target funders — this validates that pipeline.\n\n### 3. Talent Pipeline\nThe profile recommends \"entrepreneurial individuals comfortable with ambiguity\" who can build decision-making tools. This describes Deliberus's ideal collaborator profile exactly. When Fredrik is ready for collaborators, 80,000 Hours job boards and community are the recruitment channel.\n\n### 4. Positioning\nDeliberus should position itself explicitly within this problem profile's framework:\n- **Epistemic tool**: claim extraction, fact/value classification, concept tracking, contested concept detection\n- **Coordination tool**: worldview filters, bridging arguments, shared deliberation graph\n- **The structured reasoning layer**: the missing category that connects epistemic tools to coordination tools\n\n### 5. Timing\nThe profile emphasizes: \"Tools arriving just a few months earlier could mean critical safety commitments get nailed down before we develop dangerous AI systems.\" This validates the urgency of shipping. The 15-year simmering pattern matters more in this framing — every month of delay is a month where critical decisions are made without these tools.\n\n---\n\n## Key Quotes Worth Preserving\n\n> \"AGI could compress a century of progress in a decade.\"\n\n> \"AI fact checkers may be more reliable and impartial evaluators of information than humans are.\"\n\n> \"AI negotiation tools could find mutually beneficial agreements that might otherwise be missed.\"\n\n> \"Only a handful of projects are building these tools\" — vs billions invested in AI capabilities.\n\n> \"Tools arriving just a few months earlier could mean critical safety commitments get nailed down before we develop dangerous AI systems.\"\n\n> \"A few hundred more thoughtful, entrepreneurial people\" could usefully work in this space.\n\n---\n\n## What 80,000 Hours Misses (That Deliberus Addresses)\n\n1. **The structure of disagreement matters more than fact-checking.** Fact-checkers verify individual claims. Deliberus reveals WHY people disagree — contested concepts, different value premises, discourse-level divergence. The Swedish barnfattigdom debate was pure semantics; fact-checking wouldn't help.\n\n2. **Moral reasoning needs decomposition, not computation.** The profile mentions \"tools for moral reasoning\" but doesn't specify what this means. Deliberus's answer: decompose normative claims into their empirical premises and value premises, surface the value premises for examination (\"No Copout Axioms\"), and let users see WHERE their reasoning diverges from others'.\n\n3. **Coordination requires understanding disagreement, not just finding agreement.** The profile focuses on negotiation and agreement-finding. Deliberus's insight (from the conceptual threads): the system succeeds even when people disagree, as long as they understand WHY they disagree. Clarity is the goal, not consensus (Thread 8).\n\n4. **The worldview filter is the missing UX.** None of the mentioned tools offer perspective-taking — the ability to see the same argument graph through different worldview lenses. This is the \"soul of the product\" that motivates sustained use (vision.md §Why Every Predecessor Failed).\n\n---\n\n## Cross-References\n\n- [vision.md](../vision.md) §Adjacent Motivations — AI Safety already listed\n- [consensus-path-forward.md](consensus-path-forward.md) — beachhead market = rationalist/EA (validated by this profile)\n- [governance-and-financing.md](governance-and-financing.md) — EA Infrastructure Fund, Open Philanthropy as funding targets\n- [adoption-problem.md](adoption-problem.md) §Education Wedge — the EA community as beachhead\n- [steelmanned-critiques.md](steelmanned-critiques.md) §Rationality Assumption — the 50K-200K market question (this profile suggests it's larger and growing)\n- [assumption-ranking.md](assumption-ranking.md) #9 — rationalist/EA as right beachhead (strongly validated)\n"}