[Uncompiled] AI isn't objective. It never was.

There's no such thing as a neutral language model

There’s no such thing as a neutral language model. Every LLM is a mirror. And that mirror has fingerprints all over it — left by the people who built it.

And yeah... that should worry you.

Power

Every choice in the AI stack — from what data gets ingested to how the model replies when you ask it something — is a political act.

And no, I don’t mean democrats …