Editorial Policy
Last Updated: September 17, 2026
Why this page exists
Anyone can publish a technical article. Fewer publications will tell you how the article was made. This page describes our process so you can judge for yourself how much weight to give what you read here.
Our use of AI tools
We state this plainly rather than burying it: we use AI language models as a drafting and research aid. A typical article is outlined and drafted with AI assistance, then edited by a human before publication.
What that means in practice:
- AI never publishes unreviewed. Every article is read end to end by a person before it goes live.
- Code is executed, not trusted. Any code sample that claims to work has been run. If we could not run something, we say so in the article.
- Claims are checked against primary sources. Framework behaviour is verified against official documentation, not against the model's recollection of it.
- Numbers are measured or removed. Benchmarks, timings and version numbers are taken from an actual run, or they do not appear.
We think disclosure is the honest position. Google's own guidance judges content by whether it helps the reader, not by how it was produced — and unreviewed machine output usually fails that test. The review step is the point.
How an article gets made
- Topic selection. We write about problems we have actually hit, or questions readers have asked. We do not write to fill a keyword slot.
- Research. Official documentation and specifications first; blog posts and forum threads only as pointers to be verified.
- Drafting. An outline, then a full draft. This stage is AI-assisted.
- Verification. Code is run. Commands are executed. Claims are traced back to a source. Anything that cannot be verified is cut or explicitly flagged as untested.
- Human edit. A person edits for accuracy, clarity and usefulness, and adds context the draft lacked.
- Publication. With a visible author, publication date, and category.
What we will not do
- Republish the same article with swapped keywords. Near-duplicate content wastes your time and games search rankings.
- Pad word count. An article is as long as the topic needs. We would rather publish 700 useful words than 2,000 padded ones.
- Invent benchmarks or experience. If we did not measure it or build it, we do not claim we did.
- Present opinion as consensus. Where the field disagrees, we say so.
- Let advertising shape editorial. Advertisers have no input into what we write or how we assess a tool.
Corrections
Technical writing ages badly and mistakes happen. Our policy:
- Factual errors are fixed promptly once verified, and the article's updated date changes to reflect it.
- Substantive corrections are noted in the article, not quietly patched. If we told you something wrong, you should be able to see that we changed it.
- Typos and formatting are fixed without ceremony.
- Outdated articles are either updated or carry a note about which version they were written against.
Found something wrong? Tell us atcontact@krishna.cv or through thecontact form. Reports that identify a specific error get a specific fix.
Independence and money
This site is supported by display advertising. Some articles may contain affiliate links, which are disclosed in the article itself.
- Advertisers do not see articles before publication.
- No advertiser or affiliate partner can commission, veto or edit an article.
- A negative assessment of a tool is published whether or not we have a relationship with it.
- We do not currently publish sponsored posts. If that changes, they will be labelled as such.
See Disclaimer and Privacy Policy for the full detail on advertising and affiliate relationships.
Authorship
Every article carries a named author who is accountable for its content. Author pages list background and the articles that person has published. We do not use invented author personas or stock-photo bylines.
Reader contributions
Comments and corrections from readers are genuinely useful and often improve articles. Where a reader's report leads to a substantive change, we credit it unless they prefer otherwise.
Guest contributions are held to this same policy: verified code, checked claims, disclosed AI assistance, named author.