Welcome
Receipts-first writing on AI, power, and practice. Fewer essays, more experiments you can run.
Promise: no hype anesthesia. If we can’t show it, we don’t ship it.
A 5-minute gut check before you trust an output.
Spot template sludge in model text (and fix it).
Why vibes don’t scale without proof.
- Subscription Saboteur — gut-check what “only $/mo” really costs.
- Doomscroll Tax — convert feed minutes into money burned.
- Token Burn Comparator — compare LLM cost/latency side-by-side.
Prompt Engineering — simple rules + shapes that actually hold under pressure.
AI Productivity — speed gains without quality collapse or “minutes theater.”
Talking to the Machines — jailbreaks, refusals, output drift—documented with fixes.
We publish receipts (screens, steps, data) and update when we’re wrong.
We publish receipts (screens, steps, data) and fix pages with timestamps when we’re wrong. See our Stack and About.
Stay for the deltas. When we ship something real, you’ll hear about it. Subscribe →