One problem with antitrust thinking and enforcement is how little updating there is based on experience and evidence. Much less than you see in macro policy--and to be clear there is not enough of it in macro policy either. Take the case of Roomba. A 🧵 .

Regulators in the United States & Europe blocked Amazon's bid to acquire iRobot. iRobot has not filed for bankruptcy and will be acquired by its main creditor, a Chinese company. This is not a good outcome from the perspective of consumers, competition or the national interest.

It would be nice to hear more from people that supported this about how they were wrong, what lesson they learned and what they would do differently. The Fed, for example, routinely does this--it made a big mistake in 2021 but acknowledged it and corrected rapidly in 2025.

Many (but certainly not all) macro commentators are pretty good about this. They make various predictions or recommendations, things turn out differently, they explain the difference and move in in a changed way. I know I try to do this.

But I rarely hear it from most of the loudest voices in antitrust. That includes the neo-Brandesians that supported blocking this merger, a group that tends to see just about everything as yet more reason for more antitrust enforcement.

But for everyone closer to a conventional or Chicago perspective who liked everything in this thread so far, I would also say the same to you. When have you seen a merger approved where the efficiencies or promises didn't show up, how did that change your views going forward?

@billbaer50 And it's fine if you continue to think that ex ante, or even ex post, this was a good decision. But I would love pointers to public commentators who on anything ever say at least in retrospect that they had a view that didn't work out like they thought. World a complicated place.