AIUI, a current common tactic for the Russians is sending many small groups of untrained "solders" out probe the front lines and try to penetrate undefended spots. They take a ton of casualties, but some make it through, and they gradually build up, and then try to take action.

Perhaps it would be helpful to view the claims of this article through a cost/benefit analysis?

Clearly if the opponent had wanted to defeat this vehicle and take this ground, they could have.

That said, it seems likely that this vehicle substantially increased the expected cost of taking this ground, and it did so at very little cost/risk to the defenders.

This sort of device dramatically changes the equation of conflict. It seems this article does a pretty good (though unverified) job of making that case.

There are real videos, even months old of exactly these 'land drones', equipped with good ol' .50 cal. In certain situations, they fought extremely well given no risk for crew. I mean killing off entire bmp-something transport including all crew with AP rounds, typically during night since it has night vision, zoom and so on. Verified also by drone flying nearby.

Now I am not claiming all the facts stated in the article are verified by me, but I can imagine one of them got so lucky with drones and getting hidden from their view for prolonged time it could theoretically pull it off. Not sure about batteries/fuel/ammo part thought.

Yes propaganda and bullshit, but by way of exaggeration and puffery, not lying.

I wouldn't expect even a lightly informed mid-wit to think that this murderbot held the ground by itself; and I don't think the author expects that either. Thus something else is probably going on. To wit - puffery.