Skip to content
ParallaxNexus
Subscribe
Signals
◆Open-source AI models are becoming strategic infrastructure.◆AI agents are moving from chat interfaces into enterprise workflows.◆Power availability may become a constraint on AI infrastructure.◆Physical AI is moving from laboratories into commercial environments.◆Every deployed agent is a new identity with credentials.◆Inference, not training, is becoming the dominant compute story.◆Seat-based software pricing is under structural pressure.◆Software engineering is the proving ground for every other kind of agentic work.◆Open-source AI models are becoming strategic infrastructure.◆AI agents are moving from chat interfaces into enterprise workflows.◆Power availability may become a constraint on AI infrastructure.◆Physical AI is moving from laboratories into commercial environments.◆Every deployed agent is a new identity with credentials.◆Inference, not training, is becoming the dominant compute story.◆Seat-based software pricing is under structural pressure.◆Software engineering is the proving ground for every other kind of agentic work.
Trust

How We Use AI

AI assists our research and production. Humans decide what is true and what gets published.

Last updated 13 September 2026

What AI does

  • Monitors permitted sources for emerging topics (the Trend Radar).
  • Collects and organises sources into research dossiers.
  • Extracts claims and links each to its source.
  • Produces first drafts, outlines and metadata suggestions.
  • Cross-checks draft claims against the dossier and flags anything unsupported.
  • Suggests charts, internal links and social distribution copy.

What AI does not do

  • Publish. Every piece passes editor review before it goes live.
  • Invent facts, figures, quotes or experts. Drafts that contain unsupported claims are rejected by the fact-check stage.
  • Rewrite published articles automatically. Refresh alerts go to editors, who decide.
  • Scrape sources that prohibit it. We use official APIs, RSS feeds, licensed data and public datasets.

Verification

Important factual claims undergo verification by a human editor against primary sources before publication. Charts derived from editorial estimates are labelled Illustrative. Launch content that has not completed this process is labelled Demo.

Disclosure

Pieces produced with AI assistance carry a byline for the desk or the responsible editor. We do not present AI-generated text as the work of a named human who did not write it.

Your data

We do not use subscriber data to train models. See our Privacy Policy.