Editorial standards

Who's accountable

Luis Chavez-Mattos, Director of Product at Remy, is the editor of this blog. The standards on this page, the corrections process, and the decision to publish or hold an article are his responsibility, not a team's.

Find him on X and LinkedIn. These are the accounts listed in this page's structured data, so you can check that the name belongs to a real person.

Our articles are drafted with AI, so we don't put a writer's byline on them. Naming a person as the author of something they didn't write would mislead you about how it was made. Each article credits the editor who reviewed and approved it instead.

How we use AI

Articles are drafted with AI (Anthropic's Claude), then reviewed and approved by human editors before they publish. We say this plainly because it's true and you'd want to know.

In practice:

• Every draft starts from a source, not from memory. An article begins with the transcript of a technical video or the model card of a new open-weight release on Hugging Face. The draft is written against that material.
• Hardware and benchmark details get a second lookup. For articles about running a model locally or about its measured results, we fetch the model's own published documentation for the specific figures. If that lookup fails, the article goes ahead without those figures. We don't guess them.
• The repair pass is AI too. If a draft fails our automated checks, a second AI pass fixes it before any person sees it. We say so because "AI-drafted, human-edited" often suggests more human writing than actually happens.
• A person approves every batch. Nothing publishes on its own. New articles wait in a review queue, and an editor has to approve them, or pull any of them, before they reach the site.
• We don't publish to hit a number. Our daily limit is a ceiling, not a target. When a day's sources are thin, we publish less. We also cap how many articles one video or one subject can produce, so a single story isn't stretched across many posts.
• Images are AI-generated. The illustration at the top of each article is generated for it. It is not a photograph of real events.

Automated checks

Software enforces some rules on every article, because a rule that depends on someone remembering it eventually slips. Before an article reaches the review queue, checks confirm that:

• It doesn't repeat an article we've already published, including the same story told again within the last three days.
• It has the required structure: a direct answer at the top, a TL;DR, question-shaped section headings, and an FAQ.
• Its title matches what people search for instead of making a claim.
• Every link to another article on this blog points to a post that is live.
• The figures in its TL;DR and FAQ match the figures in the article.
• Our news coverage doesn't mention Remy or MindStudio.
• Its specific claims are checked against the sources it was written from.

When a source contradicts a claim, or supports only part of it, the claim is corrected to what the source says before an editor sees the article. A claim is never removed just because our sources don't mention it. We search for a source that confirms it, and if none turns up, the claim stays as written and the editor is told we couldn't confirm it. Every automatic correction is listed for the editor, who reads the article before it publishes and can undo any change.

If an article still fails after the repair pass, it's flagged to the editor in the review queue.

Corrections

We get things wrong sometimes. When we do, we fix the article instead of quietly deleting it. An article that has been substantively corrected or updated shows an "Updated" date next to its original publication date.

If you find an error, email hello@support.goremy.ai with the article's URL and what's wrong. We'd rather hear it from you than leave it up.

What we don't do

• We don't invent bylines. No fictional authors, AI personas, or stock-photo experts.
• We don't run paid coverage as editorial. If an article is ever sponsored, it will say so.
• We don't make up figures. If we can't source a number, it stays out.
• We don't pitch Remy in our news coverage. News articles never mention Remy, and software checks every draft for it.

Independence

Remy is a commercial company, and this blog exists partly to reach people who might use our product. That interest is real, and we won't pretend otherwise. It doesn't decide what we say about anyone else's products. Coverage of other tools is written to be accurate first.

The blog is published by Wooster Labs, Inc., which makes Remy.