Under active development · AI tools for political and policy analysis.
Development status →Context collection / staged product direction
Start by preserving the method behind your own repeated briefs. The proposed offering is curated, maintained evidence-linked context with declared scope, source rights and review state. Pricing and purchase are unavailable.
Local illustrative listings and buyer-review artifacts work here. Seller publishing, payments, identities, hosted sharing and independently verified credentials are not implemented.
Source-backed learning collection / editorial drafts
Inspect a policy question, its sources and its unresolved judgments. Try a cited context export, then create a draft of your own.
Browser-created editorial downloads use tacitus-editorial-draft/1, a readable JSON teaching format. PRAXIS attachment, seller publishing and payments are separate capabilities. No purchase or hosted upload takes place here.
Preparation time is not the date a statement became true or a guarantee of source freshness. Inspect each source publication date and each claim’s scope.
The earlier v1 downloads are withdrawn: they required a world-validity start these sources do not establish. Preparation and publication dates cannot substitute for that missing fact. Corrected native-v2 research packages preserve unknown validity and are not compatible with the native-v1 catalog or extractive brief compiler. They remain unreviewed educational material, with no verified PRAXIS integration or general reuse license.
Read the corrected package verification receipts. These receipts concern packaging and preserved metadata, not expert conclusions. An embedded earlier pending-review checkpoint is historical; the external receipt describes the final package checks.
3 of 3 editorial capsules
AI governance and energy infrastructure
A decision capsule for evaluating data-center proposals without confusing annual energy forecasts, peak capacity and local infrastructure obligations.
Prepared 2026-10-18T04:56:00Z · 2 sources
Humanitarian diplomacy and policy analysis
A decision capsule for structuring an access discussion while preserving uncertainty, civilian protection and the distinction between public law explanations and actual permissions.
Prepared 2026-10-18T04:56:00Z · 2 sources
Urban resilience and public health policy
A decision capsule for comparing affordable home cooling, accessible cooling spaces and neighborhood outreach before the next warm season.
Prepared 2026-10-18T04:56:00Z · 2 sources
compute-infrastructure-governance · Prepared 2026-10-18T04:56:00Z · Human review required · Revision recorded 2026-10-18T05:35:00Z
fact · us-historical-demand
The LBNL report estimates U.S. data centers used 192 TWh in 2024, or 4.7% of U.S. electricity.
Provenance: attributed source summary
Qualification: A modeled historical estimate for all data centers within the report’s scope, not AI workloads alone. Cryptocurrency mining is excluded.
World-validity is not established by preparation or publication dates.
{
"publisherDates": {
"lbnl": "2026-06-18"
},
"recordedAt": "2026-10-18T05:35:00Z",
"observationPeriod": {
"year": 2024
},
"forecastHorizon": null,
"worldValidity": {
"status": "unknown",
"from": null,
"until": null
}
}Sources: lbnl
fact · us-forecast-demand
The LBNL 2030 reference projection is 649 TWh; a compounded sensitivity stress-test range spans 521–843 TWh.
Provenance: attributed source summary
Qualification: Projection assumptions use late-2025 information. The stress-test range is not a probabilistic confidence interval and does not capture every structural discontinuity; it is neither observed demand nor committed project load.
World-validity is not established by preparation or publication dates.
{
"publisherDates": {
"lbnl": "2026-06-18"
},
"recordedAt": "2026-10-18T05:35:00Z",
"observationPeriod": null,
"forecastHorizon": {
"year": 2030
},
"worldValidity": {
"status": "unknown",
"from": null,
"until": null
}
}Sources: lbnl
fact · global-demand
The IEA’s 2026 executive summary projects global data-center electricity use rising from 485 TWh in 2025 to about 950 TWh in 2030.
Provenance: attributed source summary
Qualification: A global projection; it cannot be substituted for a U.S. forecast or a local grid impact study. It concerns all data centers, not exclusively AI workloads. IEA and LBNL models are not one harmonized dataset.
World-validity is not established by preparation or publication dates.
{
"publisherDates": {
"iea": "2026"
},
"recordedAt": "2026-10-18T05:35:00Z",
"observationPeriod": {
"baselineYear": 2025,
"status": "publisher estimate"
},
"forecastHorizon": {
"year": 2030
},
"worldValidity": {
"status": "unknown",
"from": null,
"until": null
}
}Sources: iea
analysis · capacity-distinction
Annual TWh cannot by itself determine a proposed facility’s peak MW demand, interconnection cost or reliability impact. Local load shape and engineering evidence are necessary.
Provenance: analyst proposal
Assumptions: General evidence may not transfer to a specific decision without local inputs
Analytical dependencies: us-historical-demand, us-forecast-demand, global-demand. These do not establish native invalidation propagation.
Qualification: Analytical distinction; no site-specific engineering determination is made.
World-validity is not established by preparation or publication dates.
{
"publisherDates": {
"lbnl": "2026-06-18",
"iea": "2026"
},
"recordedAt": "2026-10-18T05:35:00Z",
"observationPeriod": null,
"forecastHorizon": null,
"worldValidity": {
"status": "unknown",
"from": null,
"until": null
}
}Sources: lbnl, iea
scenario · compute-scenario
A public planning body is considering a phased data-center project and must decide what evidence to require before infrastructure commitments.
Provenance: illustrative scenario · illustrative scenario, not an actual event
Assumptions: Illustrative context only; no actual event asserted
Qualification: Illustrative governance scenario, not an identified project or an existing permitting decision.
World-validity is not established by preparation or publication dates.
{
"publisherDates": {},
"recordedAt": "2026-10-18T05:35:00Z",
"observationPeriod": null,
"forecastHorizon": null,
"worldValidity": {
"status": "unknown",
"from": null,
"until": null
}
}Sources: No source citation; analyst-generated material
Published 2026-06-18; checked 2026-10-18T04:56:00Z · Printed pages 5, 8–10: historical use, projections and compounded sensitivity analysis
Published 2026; checked 2026-10-18T04:56:00Z · Executive summary, electricity-demand outlook
Original assistant-authored TACITUS paraphrases and analysis; no publisher text, images or datasets bundled. Display and delivery are user-authorized; general reuse terms for original TACITUS material are not specified. Linked source licenses apply only to their respective materials. Hashes identify summary bytes, not publisher identity, endorsement, truth or ongoing rights.
Start with a question and one source. Your form stays in this page until you download it or leave. Source URLs are recorded, not fetched or verified. Avoid sensitive material on a shared device.
Small demonstration collection / manually authored
Inspect applicability, provenance, rights and maintenance before adopting a method. These are illustrative packs, not a live marketplace or independently verified expert credentials.
32 declared members; exact consultation passages, qualified claims, authorship and review history. Synthetic original dossier / Day 8 and Day 15 / fictional jurisdiction / English. Adoption and funding unresolved. Includes source-lineage dependencies; excludes seven neighboring dossiers and live legal assurance.
4 of 4 illustrative packs.
Policy analysis / en
Review a policy brief. A sentence turns a proposed start or a request into an adopted obligation.
legislative-stage@r1 · Illustrative / local review required
Conflict analysis / en
Edit an armed-group description. An alias, a component and an organization are treated as the same identity, or one designation becomes a global label.
contextual-terminology@r1 · Illustrative / local review required
Narrative analysis / en
Update a narrative assessment. Several publications repeat a claim and the brief calls them independent corroboration.
source-independence@r1 · Illustrative / local review required
Statecraft / en
Prepare a negotiation handover. A participant says “we are agreed” without evidence of the other party’s acceptance.
commitment-scope@r1 · Illustrative / local review required
Selected pack / legislative-stage@r1
Expected output: The revised consultation considers July 2027; adoption has not been confirmed.
legislative-stage@r1 · reviewed 2026-10-03 · Illustrative / local review required
What changed, and what can the steering group safely say?
Selected method: legislative-stage@r1. Inputs: S01 §1: proposal and non-adoption · S02 §2: industry request · S03 §1 when available: revised option · S04 §1–3: mandate and reviewer correction. Omitted: Legal advice, automatic deadline calculation, other jurisdictions without a specialist, and live compliance assurance.
Only the reporting-rule method has a complete saved briefing replay; terminology has its own local exercise. Other packs are inspectable guidance.
An author cannot relicense a publisher’s corpus or a former client’s evidence through a capsule. Listing metadata does not enforce access or promise recall of downloaded copies.
Worked incident → targeted questions → draft method → counterexample → human review → scoped artifact → another case → maintenance. This prefilled interview and analyst are fictional; public references inform the practice without endorsing TACITUS.
Edited interview answers are an inspectable artifact. They do not replace the authored replay method or create a compatible automated rule. Maintain definitions, reviewer reservations, revision history and refresh triggers before wider use.
Read the research on articulated practice and tacit knowing →
Preserve a team’s method and test a second brief. Include elicitation, setup, reviewer and maintenance effort.
Confirm recipients, semantic fidelity and source permissions. Imported author approval is provenance; each receiving task needs local review.
A small manually reviewed collection should test author effort, buyer fit, maintenance and willingness to pay before marketplace machinery.
A capsule seal or hash is an identifier, not proof of authorship, factual accuracy, permissions or signature validity. Signing remains a roadmap item.