How we write

This page describes the editorial standards that govern everything published on luvai.net — the journal, this page itself, the about page, and any future long-form content from the studio. We're writing it down because real publications have written-down standards, and a small one-person studio is no exception. If anything, the smaller the operation, the more important it is to be transparent about how the words got onto the page.

Who writes here

All long-form content on luvai.net is written by Sakuya, the studio's sole operator. There are no contributing writers, no ghost-writers, no externally-commissioned pieces. If a future post is co-written or guest-written, that fact will be disclosed in the byline of the post itself.

How drafts get written

Posts begin as a single observation, claim, or argument that I've been turning over for a while — usually long enough that I've discussed it with one or two people privately and noticed which version of the argument actually holds up under questioning. The post is the version that survived.

First drafts are written in a single sitting, longhand or typed, with no AI assistance for prose generation. I do sometimes use AI tools for adjacent tasks: checking whether a citation exists, asking for a sanity check on a technical claim, looking up a date. The prose itself is mine.

On sourcing claims

When I make a factual claim — a number, a date, a specific assertion about an external company or technology — the claim either represents something I personally observed in my own use of a tool, or something I can point to a source for. Sources are not always linked inline (the journal style is conversational, not academic), but the reasoning is always sourceable on request.

If I'm speculating, I say so. Phrases like “I think”, “in my experience”, “by my unscientific count” are deliberate signals that what follows is opinion or limited-sample observation, not established fact.

Disclosures

The studio's product, PromptCraft, is sometimes referenced in journal posts — usually as a worked example, occasionally as the subject of the post. When that happens, I'm obviously not a neutral commentator on my own product. I try to flag this relationship explicitly when relevant.

The studio does not currently have any sponsored content, affiliate links, paid placements, or any other arrangement where money changes hands in exchange for what gets written. If that ever changes, this page will be updated and any affected post will be marked.

The site does serve advertising via Google AdSense; this is disclosed in the privacy policy. AdSense ad placement is automated and has no influence on what the studio writes about.

Updates and corrections

When a post is materially updated after publication, the update is noted at the foot of the post with the date and a brief description of what changed. Typo fixes and minor copy edits don't require a note.

When a post contains an error of fact, the correction is made promptly. The original wrong claim is struck through (not silently overwritten), the correct version is added, and a corrections note explains what was wrong. If you spot an error in something I've published, please email hello@luvai.net and I'll fix it.

Tone and voice

The journal's voice is conversational and personal. I write in first person, use contractions, and address the reader directly. I don't pretend to be a detached observer when I'm clearly not.

The journal also doesn't pretend authority I don't have. I'm a software engineer who runs a small studio. I'm not a journalist, an analyst, or a futurist. The pieces here reflect what one person, working in one part of the field, has noticed. If you want broader coverage, there are better places.

AI-assisted writing — the studio's position

This deserves its own section because the studio operates in the AI tools space and the question is going to come up. The position is straightforward: prose published on this site, in the journal or elsewhere, is human-written and edited. AI tools may be used for adjacent tasks (research, fact-checking, occasional sanity checks on a technical paragraph), but the writing itself goes through a human brain and human hands.

I think this matters because the value of a journal like this is the perspective of a specific person, and a specific person isn't something an AI model can currently substitute for in a way the reader wouldn't notice. If I ever change this policy — for instance, by experimenting with AI co-writing for a specific piece — the affected post will say so plainly.

Reader corrections and feedback

The studio is small and email is the right channel for reader feedback. Corrections, disagreements, or follow-up questions are welcome at hello@luvai.net. Replies sometimes lag — one-person studio — but I read everything, and substantive notes get substantive responses.

Changes to these standards

If these standards change in any material way, the “Effective” date at the top of this page will update and the change will be noted on the home page's recent-writing surface. Editorial standards should be stable, not surprising.