Honestly this is bit confusing article.

What you are basically saying is that "MVP built in vacuum" is not better than building product that nobody wants.

You just built MVP that nobody wants.

I did such things the way you said. Problem is. People lie. And not because they are bad...

They may be polite, they may not know what they want.

What I see in interviews that people think they know what they want, you get your demo, buy in and so on, you then build the thing and they do not use it...

There are startups that found the problem and customers but failed at MVP. There were before and there are today.

Where you are correct, many start with MVP of thing they wish their customers wanted. That is wrong way.

You need to first find source of customers with same pain. And then find cheapest ways to start testing solutions.

I seen people liking demos and not buying products afterwards.

Only true product market fit is when you are gettign payed and your customers are referring you to other customers.

What I did see is when you find such a painful problem that people are willing to pay for your building the prototype.

That is only case I seen where you can get validation before MVP. Find more than one customer willing to pay for prototype that was not yet built.

But even then you should be frugal. I seen customers not liking solutions to such problem aether.

Way I put it these days is that successful startup is 30 brilliant ideas one after another. What is the problem, who is customer, how to get to a channel of such customers, how to solve the problem.

All of it needs to work under tight timeline and budget. Sounds like working in circus right? :D