What this site is for

How the calls get made

Five rules, applied everywhere:

• Counted, not felt. Numbers on this site come from records — run counts, real pricing, production evidence — because the felt number runs high every single time, and a recommendation built on vibes is just a mood with formatting.
• Rated on production-readiness. The only spec that matters is whether a thing survives a real Tuesday: messy data, real users, actual stakes. Demos are ceilings, not promises, and they're graded that way here.
• One winner per job. There is no best AI — there's a best AI per job, and any page that tells you otherwise is selling something. Our comparisons name the job first, then the winner.
• Time-stamped, and re-checked. Every call carries a verified date, gets a full pass quarterly, and gets an event pass when something big ships — because in AI, an undated ranking is a rumor. When a call changes, the page says what changed and why.
• Nothing dead-ends. Every page here ends with a next step — a tool to run, a guide to read, a build to consider. An answer that leaves you nowhere to go isn't finished.

What's actually here

The tools — three instruments, not listicles: a sixty-second triage for anything you're using (Tool, Toy, or Trash), a three-minute build decision (Build / Buy / Skip), and the automation math done honestly (Is It Worth Automating?). The guides and Top 5s — comparisons that name the job before the winner. The glossary — plain-language definitions of every term the industry uses at you. The prompts — the working artifacts from the Before You Build trilogy, free, with the reasoning attached. The model spec sheet — every model that matters, real numbers, dated. The newsletter — one real decision a week, run through the questions, ending in the right build.

The independence part

No vendor pays for a placement, a ranking, or a recommendation on this site — and nobody here gets paid when you build, which is precisely what keeps the calls honest. That independence points somewhere constructive: this site is anti-bad-AI, never anti-AI. The whole point of saying "not this" clearly is to get you to the right build faster — so every "not this" here ends with what to build instead.

Who makes the calls

One person, on the record — which is either a limitation or the entire point, depending on how you feel about anonymous listicles. The calls here are made by Dr. Jeff Wurfel: AI Tech Lead at a Hollywood previsualization studio, where the AI has to work on productions with real deadlines and unreleased material; author of the Before You Build trilogy; a decade producing and directing in Los Angeles before that; and a neuroscience PhD before that, studying how brains build confident models from incomplete signals — which turns out to be the whole AI-demo problem in one sentence. The long version of that story lives at drjeffwurfel.com/about. One line that applies to every page here: everything on this site is his own view, not his employer's.