Some of it, yes — no tool this talked-about escapes overstatement entirely. But "overhyped" isn't the same as "not useful," and the honest answer splits cleanly into what holds up and what doesn't.
It's not marketing dressing — every action genuinely gets gated based on risk, and that gate is enforced in the tool itself, not left to a policy document nobody reads. That's a real, structural difference from tools that just execute and hope you're watching.
The instant part is real for small, well-scoped tasks. The "anything" part isn't — ambiguous requirements, genuinely novel architecture decisions, and judgment calls about tradeoffs still need a person deciding, not just approving. Speed on the mechanical part doesn't remove the thinking part.
Most people reach real comfort in 2-4 weeks of regular use — genuinely faster than learning to write code from scratch, because the core skill shifts to reading and directing rather than authoring syntax. See the week-by-week breakdown for the honest version, not the marketing version.
It replaces a lot of the mechanical work around development — but the person reviewing changes still needs to understand what's being reviewed. A non-developer using it well is directing something they can read the output of, not skipping the need to think about the work entirely.
Cost scales with what you actually do, not a flat subscription regardless of use — a real difference from some competing tools, and worth understanding before assuming a specific project will be cheap or expensive. See the pricing breakdown for how this actually works.
The free guide covers exactly what it is and isn't — ten minutes, no hype either direction.
Get the Free Guide →No, not for judgment-heavy work — architecture decisions, ambiguous requirements, and genuinely novel problems still need a person deciding, not just approving. It replaces a lot of the mechanical work around that judgment, which is real but different from replacement.
Real, but it's not a substitute for actually reading what you approve. The gate exists and works as designed — it just still depends on a human reading the diff, which is where the real risk actually lives.
Most people reach real comfort in 2-4 weeks of regular use, not the "instant" framing some marketing implies, and not the "months" some skeptics assume either.