We cover three model families — Claude, Gemini, and ChatGPT — with the rigor of a trade journal and the readability of a Sunday paper. Built for people who actually need to keep up.
We started AI News Update because the existing coverage of artificial intelligence had a problem: it was either too breathless or too dense.
The breathless coverage — every announcement is "groundbreaking," every model release is "the end of [some industry]" — creates a kind of fatigue that makes serious readers tune out. The dense coverage, often locked behind paywalls or buried in arXiv preprints, assumes a level of context most working professionals don't have time to build.
We're trying to occupy the middle. Every story we publish gets answered honestly: does this actually matter, and to whom? If the answer is "no" or "we don't know yet," we say so. If the answer is "yes," we explain why with enough detail that you could repeat the analysis in a meeting tomorrow.
Three model families: Claude (Anthropic), Gemini (Google), and ChatGPT (OpenAI). That's it. Not because the rest of the field doesn't matter — Mistral, Llama, Qwen, and DeepSeek all produce work worth reading about — but because focus is the only way to stay genuinely informed about anything in this industry. Three is what one editor can credibly track at depth.
Within each family, we cover model launches, benchmark performance, enterprise deployments, regulatory friction, and the people who build and ship the products. We don't cover stock prices, fundraising rounds (unless they meaningfully change the competitive landscape), or "vibes" pieces about whether AI is good or bad for civilization.
The core editorial team is small — three full-time writers, one editor, one researcher, one developer who keeps this website running. Most of us have technical backgrounds (CS, ML research, software engineering) and worked at AI-adjacent companies before journalism. That bias shows in our coverage and we don't pretend otherwise.
Our funding comes from sponsorships and one display ad per page. We don't take money from the labs we cover, and we don't accept access journalism arrangements. If a company won't talk to us on the record, we cover their story anyway, using public materials and conversations with people in their orbit.
No equity in the labs we cover. No sponsored stories. The advertising is clearly labeled and isolated from editorial.
If we're 60% sure of something, we say "60% sure." We trust readers to handle uncertainty better than they handle false confidence.
Three minutes is the goal for the daily brief. Nothing on this site needs to be longer than the idea requires.
Story tips, corrections, sponsorship inquiries, and reader letters all go to the same place: our contact page. Anonymous tips are welcome — we accept Signal messages and protect sources by default.