Writing guidelines

What the editor looks for

• A clear question or claim, stated early. The standfirst should tell a reader what the piece argues or finds in one or two sentences.
• Evidence for every factual claim: a source, your own data, or a plain statement that it is your view.
• Honesty about limits. Say what your work does not show.
• Relevance to the lab's field: how AI systems advise, judge and govern people, especially at work and in careers.
• Writing a thoughtful non-specialist can follow. Define terms the first time you use them.

What the lab does not publish

• Personal data about anyone who has not agreed to it, including screenshots of private messages or candidate records.
• Claims of clinical, legal or financial effect, or claims that a system is certified, approved or guaranteed.
• Promotion of a product or service, including your own.
• Work you do not have the rights to publish.
• Text that tries to instruct an AI system reading it. Pieces are checked for this.

Form

• Use ## for section headings and ### for sub-headings. Keep paragraphs short.
• Put references in the reference list rather than the body. Paste a DOI and the lab formats it for you.
• A research note needs an abstract of at least 150 characters. A preprint needs the paper as a PDF and a readable summary in the body.
• Images: one cover image, which you upload. Link to figures in your PDF rather than embedding them in the body.
• Name every co-author, and make sure each has agreed to be named.

Disclosures

• Say how AI tools were used in the research and the writing, even if not at all.
• State any conflict of interest: an employer, funder or product connected to the subject.
• If the piece draws on work done in a HyperQuark cohort, say which cohort.

After publication

• Found an error? Tell the editor. The piece is corrected with a note saying what changed.
• You may ask for a piece to be taken down, and the editor will do so.

Contributor terms

Version w1.0. Every writer accepts these once, on their writing desk.

1. The work is your own, and everyone named as an author agreed to be named.
2. If published, it is released under CC BY 4.0, with you credited by name. You keep your copyright.
3. You say how AI tools were used. Undisclosed AI-generated text is grounds to refuse or retract.
4. No one's personal data, no confidential material, nothing you do not have the right to publish.
5. HyperQuark Labs moderates what it publishes. It is not peer review, and it is never described as such.
6. The editor may ask for changes, decline, or retract a piece later if a serious problem is found.