You Can't Stop the Computer

Mar 13, 2026

They dig into Customs saying its computer prevents refunding tariffs and what courts think of that claim. They unpack fights over installing unconfirmed officials and a ruling blocking split U.S. attorney powers. They cover a racial-profiling finding against ICE without an injunction and tease subscriber-only legal battles involving AI, voting-machine indictments, and tort claims against an AI maker.

Ask episode

AI Snips

Chapters

Transcript

Episode notes

INSIGHT

Customs Computer Blocks Immediate Tariff Refunds

The Customs Service says automatic liquidation in its ACE computer prevents immediate pro rata refunds for invalid IEEPA tariffs.
Customs claims splitting legal and illegal portions would take 4.5 million manual hours, so it asked for 45 days to reprogram ACE and issue refunds.

INSIGHT

Judges Reject Technical Excuses For Illegal Policy

Judges are unsympathetic to 'the computer won't let us' excuses when agencies proactively implemented controversial policies.
Ken White predicts a judge will scold the government and impose a strict, short compliance schedule demanding updates.

INSIGHT

Courts Strike Down Workarounds For U.S. Attorney Appointments

Courts are blocking the administration's creative workarounds to install unconfirmed officials in Senate-confirmed posts.
In New Jersey the administration split U.S. attorney powers among three people and the judge found that evades statutory appointment rules.

Get the Snipd Podcast app to discover more snips from this episode

Get the app

This is a free preview of a paid episode. To hear more, visit www.serioustrouble.show

This week Ken and Josh discuss the Customs Service saying its computers won't let it refund IEEPA tariffs, more situations where courts are telling the Trump Administration it can’t just ignore the need to get officials confirmed by the Senate, and another decision about ICE.

That's for all subscribers. Paying subscribers will also hear our conversations about a number of additional lawsuits, including some especially weird ones:

* Voting machine maker Smartmatic’s parent company under indictment over bribes its former executives are alleged to have paid in the Philippines, alleges that it is being selectively and vindictively prosecuted.

* Anthropic suing over the Pentagon’s “supply chain risk” designation that threatens the company’s business. The company makes First Amendment claims, but Ken thinks its less glamorous arguments — like that the designation violated everyone’s favorite law, the Administrative Procedure Act — are more persuasive.

* Nippon Life Insurance Company of America suing OpenAI, the makers of the ChatGPT AI engine. Nippon says it has been dogged by a vexatious litigant — she decided she didn’t like the settlement she’d signed with the company, and when her human lawyer advised her that settlements are a no-backsies kind of situation, she fired him in favor of the AI engine that gave her the advice she wanted to hear: sue, sue, sue. Nippon says this is tortious interference with the valid settlement contract they’d entered with their aggrieved former policyholder. Because tortious interference requires knowledge of the contract you’re interfering with, this lawsuit turns an interesting philosophical question into an interesting legal one — did OpenAI “know” that Nippon had a settlement, simply because their former policyholder told ChatGPT about it?

* And Ed Martin appears to be the Justice Department official with some especially stupid bar trouble.

Upgrade your subscription to receive all of our episodes at serioustrouble.show.