What the Structure Predicts
Note: This was written prior to the election results being published. It did not consider what might happen if there was a blowout, which there was. I’m posting because I don’t want to hide what the model sees and does - even when it’s applied to circumstances that didn’t happen. So... Here’s what it suggests would’ve happened if this was close. Some may still be applicable in the next session of Congress.
What the model saw before today
In October I argued that the federal election security network had come apart, and that whatever broke on November 3 would look like a local failure with a cause several layers down. The morning of November 3, 2026, Westchester County's ballot scanners started rejecting ballots across the county. The cause was a programming update made after testing. It got pushed to one town's machines and not the rest. It'll be remembered as a Westchester problem, not a national one.
This time I'm writing about what the administration did to the parts of government that make rules matter, and what that predicts for the weeks between tonight and the day the new Congress is sworn in.
What changed
Very few of the rules were repealed. The offices that enforced them were gutted.
Seventeen inspectors general were fired in January 2025 without the notice the law requires. A judge ruled that broke the law. She also ruled she couldn't put anyone back. The office that protects whistleblowers went eighteen months without a confirmed leader. The board that hears appeals from fired workers lost its quorum for seven months, and its members can now be fired at will. The Justice Department unit that handles election crimes went from about thirty lawyers to a handful. The Federal Election Commission has had no quorum since April 2025 and can't open a case. And on June 29, the Supreme Court ruled that the president can fire members of independent commissions whenever he wants, overturning a precedent that had stood for 90 years.
A rule is a promise about what happens when someone does... a thing. As I've said before, the promise holds if:
The test sits outside the group being tested.
The consequence lands on the person who acted.
A specific question gets a specific answer.
Remove those and the rule becomes a choice. It gets followed out of habit... until it doesn't.
Where the layers sit now
My read is that no single layer is to blame. Purpose moved, and the layers above it couldn't keep up.
Purpose: the executive branch's mission shifted from running agencies according to their purpose to carrying out one person's agenda with them.
Systems: the checks built for the old mission were dismantled before anything replaced them. Nothing coordinates actions across agencies, and nothing routinely tests them against what the president actually wants before they happen.
Rules: a rule only gets tested after it's broken, when someone sues. Some fail the test in court. Some never get tested, because Congress isn't doing the oversight it used to.
Awareness: the formal feedback loop from Execution is mostly gone. When an action fails, no office inside the government has to explain why it was taken or what it cost, unless the president asks.
Execution: the president's advocates do what they think he wants. Most of it won't be ordered. It'll be inferred.
That last point is the engine. A structure where people act on inferred intent, with no check before and no consequence after, produces behavior you can predict. That's the claim this issue tests.
The case already on the table
On October 22, 2026 the Justice Department sued Alaska. It asked a court to block the state from certifying its Senate race if the Republican incumbent lost, on the theory that a second candidate with the same name would confuse voters. The president said he ordered the suit. The complaint defined the problem as a result where the incumbent loses, and would have won if the other Sullivan's votes had gone to him. If the incumbent won, no problem existed.
On Sunday a federal judge dismissed it. She wrote that the Constitution gives states the authority to run elections, and that the department offered no workable way to do what it asked. The department appealed within hours. Its criminal probe into the race continues, and the governor has said he may oppose certification depending on what it finds.
Run that through the Lens, which checks what an organization thinks, says, and does against each other. What the department says is that it's protecting the integrity of Alaska's election. What it did was file a suit that only kicked in on a loss. You can read the thinking straight off the action: integrity means the right result. That reading is the basis for everything below.
Five predictions
Each of these follows from the structure, and each comes with a way to check it. Some are obvious. Others aren't.
1. Federal challenges will almost always show up where the president's side lost. The Lens says enforcement follows the outcome, not the rule. So between now and certification, federal suits, probes, and public demands to delay certification will show up in races the president's candidates lost, and skip races they won by similar margins. Check this by listing every federal election action filed in the next two weeks and sorting by who won.
2. The actions won't line up with each other. With no coordination layer, different actors will improvise toward what they think is wanted. A federal filing in one state, a governor's investigation in another, a monitor's report somewhere else, and the claims won't match. Sometimes they'll contradict each other inside the same incident. If two parts of the government assert incompatible things about the same race, the Execution read is right.
3. Most of it will lose in court, and losing won't change the next filing. Expect a string of dismissals, each appealed the same day, none of them changing the next complaint. That's what a missing feedback loop looks like from the outside. The department's own spokesperson said after Sunday's dismissal that Alaskans deserve a fair election. If they lose Alaska, or any other traditionally red state, they'll keep saying it, no matter how many times the courts say no.
4. Whatever breaks will get blamed locally. This one already has evidence. Westchester owns the scanner failure. The county will take the hit for a problem that didn't start there. Expect the same pattern wherever counting runs slow or a system fails. The story will be about one county office, one local failing, maybe one bad actor.
5. The Snapback, if it comes, will be partial. If the House changes hands, Congress starts investigating in January. That gives the Rules layer a working test it hasn't had for two years. But it doesn't rebuild the dismantled Systems. Inspectors general don't come back because a committee starts doing its job again. The Supreme Court's rulings don't expire. What changes is that untested rules become testable again. The wobbly structure underneath stays the same.
What would prove me wrong
A model you can't disprove isn't a model.
If challenges land evenly across races the president's side won and lost, the Lens read is wrong, and the department is enforcing a rule rather than an outcome.
If the moves across agencies and states are coordinated and consistent, the Execution read is wrong, and something is checking actions before they happen that I can't see.
If a loss in court changes the next filing, the feedback loop exists and I missed it.
If the counting problems get traced publicly to their real cause instead of the county that absorbed them, the Awareness read is wrong.
Does it matter?
You probably don't run an agency. But if you changed your mission and cut the functions built for the old one, this is your structure too. The people who work for you are acting on what they think you want. There's no one checking before they act. And when something goes wrong, the blame will land on whoever was closest to it.
The test is simple. Pick a critical rule your people are supposed to follow and ask the last time it was enforced. If no one can remember, you don't have a rule. You have a preference, and you might not have that much longer.
