The title is a little bit of hyperbole, and I know this is going to get a lot of comments saying things like "this is fucking stupid", but this is actually pretty insightful. Building stuff before you know you're building something somebody needs/wants is very risky and can be a total money/time/attention sink. To the extent that it's possible, it IS a good idea to try and validate the product idea BEFORE writing code.

If we've made one big mistake at Fogbeam over the years, I'd actually say that it's building too much of the actual product, instead of using mockups and dummied-up demos and what-not just to test the demand for the concept. Early on I had a lot of mindset of "We need to build a fair amount of this, so we can show it to people, so we can find out if they need/want it or not". But honestly, we could have done wireframes, mocks, backend-less demos that were totally dummied up, etc, just to test the concept in front of potential customers.

So yeah, I'd say, don't get too caught up in the headline, but read the article and really think about what the author is saying. IMO, there is some valuable truth here.