I fear a general misconception out there that you can validate a startup in just a few weeks. I have been particularly pained by both non-coder startup founders and managers at big companies who think that all they need to do is pour money into a single MVP and they’ll have a binary answer as to whether something is awesome or terrible. (note: when I use the term MVP, I mean an actual product, not an experiment)
If you do not have the budget and runway to iterate an MVP and run through multiple build-measure-learn cycles (tip: you should always start with “learn”), all you really are doing is playing the lottery.
Even if an idea is truly solid, in 99% of the time, you still won’t have product-market fit after an MVP. This stuff is not that easy.
What an MVP definitely gives you is essential learning and market signals, and hopefully an important initial foothold in the market.
If you don’t have the determination and ability (budget) to iterate, I don’t really believe that it is worth starting in the first place.