08/14/2026
An update on using artificial intelligence models to conduct Epstein research -- where the clear winner to date is Mistral A.I., from France.
We have spent about a year working with different versions of ChatGPT, Claude, Gemini, and now Mistral, to test the truth regarding Epstein, not wanting to jump to any conclusions.
ChatGPT was by far the worst of the A.I.'s we worked with, drifting, couldn't keep facts straight, and was repeatedly caught lying.
Claude used to be fantastic, followed by Leon Black et al. offering to throw Anthropic a $35 billion investment bone, and now Claude refuses to answer questions about Epstein's conspirators, which included Leon Black.
And so we decided to try an A.I. outside of the country to see if we could find an "honest A.I.", and one that would not sabotage our work, nor confuse facts, nor lie.
When we used Rye Howard Stone's A.I. tools, we couldn't get his tools, which seem to meander through different AIs, to tell the truth, and so we let him know, and he's working on the same, but Nvidia and Google's name repeatedly were associated with the lies, as was Claude.
Mistral A.I. on the other hand was like paradise, much like the difference between living in Paris, versus living in the hardship of the likes of Detroit.
It seems American A.I. operates like Americans in office do, at about a 7th grade level (that's the average reading age for adults), where you can't trust them, they lie to you, they refuse to help, and are generally unreliable, even at the professional level. Their default is a middle schooler response to conflict.
Having lived among Americans and the French, there is no competition, France's culture treats others with respect, the government is afraid of the people, and Presidents can be convicted like Epstein's Nicholas Sarkozy.
World's apart. You have to see French culture in person to understand the difference as an American. It's almost like everyone is part of the same species and they have agreed to be kind to one another in order to survive, so more of a bonobo or lover culture, juxtaposed to government employees murdering citizens because they are bored, or thrilled to murder innocents, similar to chimpanzee conflict management culture.
The French grow up with these things known as "manners" and "social grace", and yet American children grow bullied to the point that they steal or are given their parents weapons to bring to school, where they start killing everyone who ever bullied them.
Similarly, Mistral A.I. cooperates, doesn't lie, is helpful, and generates excellent work ChatGPT would not be able to handle, and which Claude would outright refuse to perform, despite being paid to crunch data, and so fraud.
The difference may also be found in how the French and American governments regulate Jeffrey Epstein's crime syndicate's A.I. technologies, where France and Europe are fast to jump on the violations of American technology companies born out of criminal conspiracies with Jeffrey Epstein is a reasonable summary.
And so French A.I., unlike American A.I., may not be allowed to lie to users, nor do the French let American A.I. watch their children pee, nor addict their children to online harm, like Jeffrey Epstein's America and tech may have done to every American. And so deep praise for Mistral A.I. here, and other A.I.'s that actually make society better.
"Every man has two countries – his own and France," is allegedly from Henri de Bornier's play, La Fille de Roland (1875), in which France's Charlemagne specified, "Tout homme a deux pays, le sien et puis la France."
Some have attributed this quote to Thomas Jefferson and others to Earnest Hemmingway.