AI Catchup

Editorial Policy

How we evaluate AI tools, source claims, and decide what makes our picks each month.

Independence

We do not accept payment, equity, free credits, or any other consideration in exchange for placement on our picks pages, our tool reviews, or our weekly newsletter. Vendors do not get editorial review of articles before publication. If a vendor provides hands-on access (e.g. a closed beta), we say so in the article.

We may include affiliate links to products we already chose editorially, but the choice never depends on the affiliate program. If we add affiliate links to a piece, the piece notes it explicitly.

Sourcing

Every specific factual claim -- a price, a feature, a release date, a benchmark score, a usage statistic -- links to a primary source. The order of preference for primary sources is:

  1. The vendor's own documentation, changelog, or release post.
  2. The vendor's official social account at the time of announcement.
  3. Peer-reviewed research or government datasets.
  4. First-hand testing by the AI Catchup team, with the test conditions described.
  5. Reporting from named tech publications (only if no primary source exists).

We do not synthesize benchmark numbers from secondary aggregators or unattributed forum posts.

How our picks pages are chosen

Our picks pages (top AI model, top coding tools, top MCP servers) reflect our editor's recommendation for each category right now. The process:

  1. The editor uses every contender on the list in real production work for at least two weeks.
  2. We re-evaluate every month, or whenever a major release changes the landscape.
  3. Each pick page records a changelog of when and why the rankings changed.
  4. We never include a tool we have not personally used.

AI use in our content

Articles on AI Catchup are written by humans. We use AI tools (the same ones we cover) to help with research synthesis, fact checking, and copy editing. Every piece is reviewed, fact-checked, and signed by a named human author before publication. We do not publish auto-generated content.

Corrections

When we get something wrong, we fix it inline and add a correction note at the top of the article. The original publication date stays put; an "Updated" date reflects when the correction landed. If you spot an error, please email hello@updates.sensara.io.

Reader voice and tone

Articles are written for a working developer in second-person voice. We avoid hedging language and try to deliver an opinionated verdict in the first paragraph. The weekly newsletter is written first-person plural ("we") because it's a curated digest from a team. We do not use em-dashes; we use double-hyphens (--) instead, including in this sentence.

Get the weekly AI Catchup

Tools, practices, and what matters -- in your inbox every Monday.