You mean, build a statistically relevant sample of B2B products and then "test" selling them to enterprises?

I guess I must have misunderstood, because this is not worthwhile advice :)

The example hypothesis is: "you do not need to be on the inside, or have a pre-existing connection on the inside, in order to create a successful B2B product".

How would you test this? Yes, one way is to build B2B products without any input and then try to sell them. Many people try to do this. This is, as you pointed out, a really terrible approach.

A much better approach is just to hit the streets or the phones and talk to potential customers. Maybe you think they'll blow you off (and in my experience, yes, some of them will). But what I'm asking you to test experimentally is that 100% of them will blow you off. I have actually tried this, and the surprising result is that this is simply not the case.

But don't believe me. Try it for yourself.

Once you have people who will talk to you, then you can find out their problems. Once you know what their problems are, you can see if other people in that position have similar problems. Pick the most painful one of these problems, and, with your new industry insider friends to guide you, build an awesome product.

My big point here is that one should not simply accept limiting beliefs about what is and is not possible without actually putting them to the test. Do the uncomfortable thing. Grow.