A bot crawled thousands of studies looking for simple math errors

Here we have numbers-checking working the same way.

I bet you this sort of feature gets built in to word processors eventually, and puts wavy red lines under the results it flags.

We've had this sort of real-time "syntax" checking in software engineering for half a generation. It seems wise for other disciplines to consider adopting it too.

It's obviously got to be discretionary, just like spell-check is discretionary in browsers.

We will get a new genre of humor, thought "statcheck fail."

There was a story about two chemists, I might have some minor details wrong, but that shouldn't hurt this anecdote because something like this happened back when chemistry was still raw and new. The going wisdom was based on the assumption that the atoms in a molecule were the only thing that determined its properties. These two scientists were having trouble replicating each others results when working with a molecule of 2 Carbon, 2 Nitrogen, 2 Oxygen and 1 Mercury atoms.

One describe it as inert and the other as explosive. Neither bothered to include synthesis steps in their papers and each thought the other was wrong. Rather than insult they swapped more and more information until eventually each confirmed that we dealing with a molecule with identical parts, but different properties.

They had discovered that molecules have structure. One arrangement was Mercury Fulminate an unstable explosive used in primers and detonators today and the other was really inert and not of interest, to me at least.

Any scientist who argues with the concept of "disagreement based on evidence" is arguing a foundation of science. They should be looking for evidence of reasons for that disagreement instead of quibbling over personal attacks and defense.

Also to say i think it's a graded thing. Often such type of critique might be serious and well developed, often it might be serious but based on misunderstanding, and sometimes it might be an attempt at trolling, annoying or payback. I would guess your average peer reviewer engages more deeply with the paper than most readers and takes time to check their critique before sending it, and since there's an editor involved there might also be a better level of politeness. If now suddenly every reader is free to criticise (assuming they fall in category 1 & 2) that can end up in a barrage of critiques that an author might feel obliged to respond to in order to protect his/her reputation. And still those critiques can then be taken out of context, eg when someone in the media or from the competition tries to put you down a notch for holding the wrong views...

Think of anti-vaxers going around and criticising every paper on vaccines. You're always going to find something to criticise and once it's criticised you can point to SG authoritative 'the paper is criticised by other scientists for..'

Or another angle: do you want to feel obligated to defend every one of your papers for the rest of your life?

This can really feel like bullying even if it's not intended as such. That's in particular if you write about contentious issues or such with strong interest groups - think vaccines, abortion, gender, GMO, lobbying, financial regulation, inequality, race, ...

I think open 'peer review/critique' is a good idea, but it can take dark turns if there's no mechanism to prevent this.