How we work
Luminesca News is a small independent analysis desk. This page explains what we cover, where our information comes from, how we use AI tools, and how we handle corrections.
Who writes here
We publish under a single site byline rather than individual names, because there are no individual staff: this is a solo project. Research, drafting and publication are carried out by the operator with AI assistance and programmatic checks. We do not present ourselves as subject-matter authorities, and we do not claim professional credentials we cannot document.
What we cover
Technology and AI (models, developer tools, infrastructure and cost), energy and climate (grid, storage, policy and markets), and digital policy (privacy, platform regulation and AI governance). We focus on explainers and analysis rather than breaking news.
Sourcing rules
- Primary sources first: official documents, company announcements, regulator publications, standards bodies and peer-reviewed material take precedence over secondary reporting.
- Every article cites its sources in a Sources section at the end, so you can check our claims yourself.
- Key figures are cross-checked against at least the cited source, and we avoid reporting numbers we cannot trace to a public document.
- Unverified reporting is labelled as such — where a deal or event is reported but not officially confirmed, we say so in the text.
AI-assisted drafting
We use AI tools to assist research and drafting. This is disclosed on every article. Claims and figures are checked against the sources cited at the end of each piece, and we correct errors we find or that readers report. We do not present AI-assisted drafts as first-hand reporting, field testing, or professional analysis.
In practice that division of labour looks like this: AI tools accelerate the literature sweep, the outline and the first draft; humans decide what the piece should say, verify every load-bearing claim against its cited source, and take responsibility for what is published. The byline is the desk because the responsibility is the desk's — and the corrections policy above is how that responsibility is enforced in public.
Corrections
If you spot a factual error, email [email protected]. We check the claim against the cited source and publish a correction note on the article for material errors.
How stories are chosen
We pick topics the way a reference librarian would: questions that keep coming back, where the public record already contains a real answer. A comparison between AI chatbots, an explanation of how long-duration storage works, a walkthrough of what the EU AI Act requires and from whom — these are stable questions that deserve stable, correct answers, not a fresh take for every news cycle.
Ideas come from three places: our own reading of official publications, reader emails, and patterns we notice across the beats we already cover. We keep a running list, and topics that survive on it for months — because people keep searching for them and the public record keeps answering — get written first.
Review checklist before publication
Every article goes through the same checks before it goes live:
- Claim-to-source traceability: every number and factual assertion can be traced to the public sources listed at the end of the piece.
- Disclosure: the AI-assisted drafting disclosure appears on the article, in plain language.
- Balance: where the record is genuinely uncertain, the text says so instead of picking a side for readability.
- Plain English: jargon is defined on first use, and a reader with no domain background should finish the piece understanding it.
What we are not
To be equally clear about the limits: we are not first-hand reporters, we do not attend press events, and we do not conduct interviews behind closed doors. Everything we publish is built from the public record. We are also not professional advisers — nothing here is legal, financial or investment advice. Where a decision is high-stakes, read the primary sources we cite, and if you find a gap between our summary and the source, tell us: that is exactly the kind of error the corrections process exists to catch.
Last updated: September 9, 2026
← Back to Luminesca News