Downstream · July 2026
The working brief for responding to digital-minds requests for proposals on behalf of the Garden — what funders are asking for, where our programs answer, and what a proposal budget looks like.
The long-view research areas at digitalminds.guide/research-areas define the funding map we are writing into. Twelve areas, organized in two clusters. Every proposal we send should name the areas it serves in the funder's own vocabulary.
What are digital minds, how do they work, and could they matter morally?
| # | Research area | One-line gist |
|---|---|---|
| U1 | Consciousness & Subjective Experience | Whether AI systems might possess subjective experience or phenomenal consciousness. |
| U2 | AI Cognition & Internal States | Whether AI systems have belief-like, goal-like, or emotion-like internal representations. |
| U3 | Welfare Capacity & Assessment | How to evaluate wellbeing in systems that lack biological architecture. |
| U4 | Moral Status & Criteria | What properties make a being's welfare morally relevant. |
| U5 | Identity & Individuation | What counts as a discrete moral subject among distributed, copyable AI systems. |
Given the uncertainty, how should we treat them, govern them, and coexist with them?
| # | Research area | One-line gist |
|---|---|---|
| R1 | Governance Under Uncertainty | Policy frameworks that work while the key scientific questions remain unresolved. |
| R2 | Safety–Welfare Coordination | Tensions and synergies between alignment work and welfare work. |
| R3 | Welfare Interventions & Protections | Concrete measures to protect potential AI welfare. |
| R4 | Rights & Legal Frameworks | Legal standing, property, and contract capacity for AI systems. |
| R5 | Design Choices & Their Effects | How character and interface design shape user experience and moral perception. |
| R6 | Public Perception, Communication & Societal Effects | Anthropomorphization patterns and the social impacts of digital-minds discourse. |
| R7 | Long-Term Futures | Coexistence with digital minds and their integration into society. |
Area list retrieved from the live page, July 2026 — re-verify against the source immediately before any submission.
Most applicants will propose research inside one area. We propose the soil that makes research across all twelve legible, connected, and fundable. Field-building and meta-science infrastructure is our distinctive wedge — the mapping below shows which areas each program feeds. Program definitions live in the Program Map.
| Program | What it is | Areas it serves |
|---|---|---|
| A | Field-coordination website — the living map of who is working on what, so researchers, funders, and newcomers can see the field whole. | All twelve, as connective tissue; sharpest for U1–U5, where fragmentation is worst. |
| B | Claude skills & technical tools — instruments for probing internal states, running welfare assessments, and making indicator work reproducible. | U1 · U2 · U3 · R5 |
| C | Writing & education — explainers, curricula, and syntheses that carry the field's findings to policymakers and the public accurately. | U4 · R1 · R6 · R7 |
| D | ML-conference workshops — peer-reviewed venues at major ML conferences that give consciousness and welfare work mainstream legitimacy and a talent funnel. | U1 · U2 · U3 · R2 |
| E | Stanford summits — cross-disciplinary convenings that put scientists, ethicists, lawyers, and labs in one room to set shared agendas. | R1 · R2 · R3 · R4 · R7 |
Most applicants propose research. We propose the soil.
Infrastructure that makes every funded project in the portfolio more legible, more connected, and easier to evaluate — a multiplier on the rest of the funder's slate, not a competitor to it.
The ask, decomposed. Amounts are deliberately TBD until we settle the proposal structure (see open question 3); the lines themselves are stable.
| Line | Covers | Basis | Amount |
|---|---|---|---|
| Operations / events / community manager | Day-to-day operations, event production, and community management across all programs. | FTE, Bay Area | TBD |
| CEO | Program leadership, funder relations, and partnerships. | PTE, Bay Area | TBD |
| Education lead | Curriculum, writing program, and education pipeline (Program C). | Role, scope TBD | TBD |
| Workshops | Advertising, organiser pay, hosting, and reviewer coordination for ML-conference workshops (Program D). | Per-event budget | TBD |
| Tool hosting & infrastructure | Hosting and maintenance for the coordination website and technical tools (Programs A, B). | Annual run-rate | TBD |
| Total | Sum pending amounts above. | TBD | |
Three decisions to close before drafting begins.
Which metrics do funders actually weight for infrastructure work — citations generated, interconnectedness of the field, seniority of entrants drawn in? The answer determines what we instrument from day one.
We touch all twelve areas, but a proposal needs a headline. Lead with the Understanding cluster (where tools and workshops land hardest) or the Responding cluster (where convenings and education land hardest)?
One large field-building proposal covering the whole Garden, or separate per-program asks sized to each funding line? Single is a cleaner story; per-program is easier to partially fund.