The false choice
Founders treat this as build versus research, but validation and the MVP are stages of the same process. Validation answers whether the problem and buyer are real. The MVP answers whether your specific solution gets used and paid for. You do the cheap question first because it can kill the idea before you write a line of code. An MVP is what you build after the problem clears that bar, not instead of checking.
An MVP is not your first build, it is your next test
The point of a minimum viable product is to learn the most with the least, so it should target the one thing you are still unsure about. Often you do not need code at all. A concierge MVP delivers the service by hand, a smoke test measures whether people click buy, and a pre-sale measures whether they actually pay. Reach for real code only when those lighter tests have nothing left to teach you.
Sequence that actually saves time
Run it in order: confirm the pain through customer conversations, test demand with a landing page or pre-sale, then build the thinnest MVP that resolves the biggest remaining risk. Each step is cheaper than the next, so you fail fast and cheap when the idea is weak. Jumping to a polished MVP inverts this, putting your most expensive step first against your least-tested assumption.
When building the MVP is the right next move
Build once demand is no longer the open question and usage is. If buyers have already said yes with their time or money, and the only way to learn more is to put a working thing in their hands, build it. Keep it minimum: one core workflow, no polish, shipped fast. The signal you want is engaged use and willingness to pay, not applause for the feature list.