Start with the poll, because it is the cleanest case.
Last week, a survey showing Karen Bass beating socialist candidate Nithya Raman by double digits in the Los Angeles mayor's race circulated widely. The Los Angeles Times attempted to contact the pollsters behind it. There were no pollsters. The outfit, Median Strategies, subsequently told the Times it had been 'created as a short term social experiment to examine how purported polling information could enter and spread through the political information ecosystem without independent verification.'
Say it plainly: they invented a poll, fed it to the press, watched it travel, and called the exercise a 'social experiment.' The record is public. What the record shows is that fabricated data moved through the political information ecosystem without a single checkpoint catching it. The people who spread it were not malicious — they were credulous. That distinction matters less than it used to.
Now move to Florida, where the data is harder to dismiss.
In April, Dan Bilzerian announced his candidacy for Florida's 6th Congressional District. His campaign, according to reporting cited by Nellie Bowles in The Free Press, consisted almost entirely of antisemitic messaging. He called his opponent Randy Fine a 'fat Jew' and a 'Satanic whore.' He posted that 'District 6 understands the Jewish problem.' He did very little traditional campaigning.
He won 18.7 percent of the vote.
For context: former Free Press contributor James Fishback, who ran a higher-profile, more conventionally structured campaign in the same state, won 10.5 percent. Bilzerian, running on nothing but open hatred, outperformed him by more than eight points.
These two stories — the fake poll and the Nazi candidate's vote share — are not the same kind of story. One is about institutional failure in journalism. The other is about something older and uglier moving through the electorate. But they share a common thread: the absence of friction. The poll spread because no one stopped to verify it. Bilzerian's message spread because it required no argument, no policy, no trade-off. Ancient scapegoating, as Bowles notes, is frictionless. You do not have to explain a tax code. You just point.
The free press — lowercase, the concept — depends on two things working simultaneously: institutions that verify before they amplify, and an electorate that demands more than a target. Last week offered evidence that both are under strain.
The 'social experiment' framing from Median Strategies deserves one more look. The people behind it seem to believe they exposed a vulnerability. They did. But exposing a vulnerability by exploiting it is not neutral. Voters in Los Angeles made decisions — or had their decisions shaped — by information that was invented. Calling that a research methodology does not change what it was.
Follow the incentive, not the press release. The incentive to fabricate a poll that benefits one candidate is obvious. The incentive to dress that fabrication in academic language afterward is equally obvious. What is not obvious is why any of this should be trusted to self-correct.
The Signal's read: a media environment that cannot distinguish a real poll from a manufactured one is not a free press — it is a distribution network. And a district where nearly one in five voters chose a candidate whose sole message was ethnic hatred is not a rounding error. Both facts deserve more than a news cycle. They deserve a reckoning with first principles: what verification is for, and what democratic participation requires of the people doing it.
