You use the word constructive; can you elaborate on what you mean by it?

It's like you say, basically. While "constructive" may be vague, it's certainly not constructive for people to be picking sides without actual understanding of a topic. The article kind of dances around with this idea: what if there were public comment threads on the actual journals? What if the actual literature was actually married to these ugly discussions? Papers would be permanently colored by these perceptions, and it would change how they are treated.

If you don't eventually make advances accessible to lay people there's no way to be sure you're in fact doing something of value.

I don't disagree, but the counterpoint is that actual research is a lot more nuanced than that. Few problems are as simple as: X will solve Y, from which everyone will benefit. Our scientific understanding in so many areas is very deep and specialized and even understanding these things on a basic level may necessitate years of training. Indeed many people are working on problems which are several steps removed from anything that's applicable, but nonetheless necessary.

Making sure that research is of value is definitely important, but when you're trying to convince the public of value it becomes very easy to mislead or oversell your work. How frequently is cancer cured in scientific journalism? While the public should know about scientific advances, I don't have as much certainty that they can appropriately appreciate and understand their significance, let alone make judgements about it. The article does a very nice job of highlighting how groupthink actively disrupts this.