There’s something almost irresistible about the phrase “truth machine.” It promises a world where falsehoods are impossible, where facts are settled by a quiet, reliable algorithm, and where we can finally stop worrying about being lied to. That is precisely the fantasy that Judd Legum, writing in Popular Information, sets out to dismantle in his essay “The ‘truth machine’ is lying to you.” Legum is not simply attacking a piece of technology. He is attacking a worldview — the belief that truth can be automated, that a platform can be trusted to sort the real from the fake simply because it claims to be open, neutral, or free. The article focuses on the rise of what has been called the “truth machine” in our media landscape, particularly in the form of Elon Musk’s much-publicized quest to turn X, formerly Twitter, into a kind of digital oracle. Legum argues that this vision is not only naive but dangerous. A machine built to seek truth is still a machine built by people with incentives, and in the case of X, the incentives have everything to do with engagement, power, and profit. The result, he insists, is not a truth machine at all, but a very effective lying machine.
Musk’s rhetoric has always been grand. He has spoken about creating a “digital town square,” defending free speech, and making the platform “the most accurate source of truth.” In practice, however, the reality has been very different. Since taking over the platform, Musk laid off enormous portions of the teams responsible for content moderation and trust and safety. He opened the doors to a new verification system that effectively allows nearly anyone to buy a blue checkmark — a symbol that once meant the user was a person of public interest and not likely to be an impersonator. He also made changes to the recommendation algorithm that prioritize sensational, provocative, and often misleading content, because such content generates the most engagement. This is the heart of Legum’s critique: a socalled truth machine cannot produce truth if its primary purpose is to keep people staring at their screens. Algorithms do not ask “Is this true?” They ask “Will this keep the platform active?” When those two questions conflict — which happens almost every day — the algorithm will almost always choose the latter. In the world of the “truth machine,” falsehood is not a malfunction. It is a feature.
What makes Legum’s argument so compelling is that he grounds it in real, painful examples. The “truth machine” has spread fabricated images of world events, manipulated videos of political leaders, and viral rumors that would never survive even a basic journalistic review. There was the infamous fake image of a supposed explosion near the Pentagon that briefly caused confusion and volatility in financial markets. There were the doctored clips of President Biden that were designed to make him appear confused or physically diminished, and which circulated with no context and no warning. There were the endless claims about rigged elections, vaccine safety, and geopolitical conspiracies, all treated as ordinary speech, as if truth and lies had the same value. The machine did not have to be biased to do this harm. It only had to be indifferent. By treating all content equally, by refusing to distinguish between a verified fact and a malicious fantasy, the “truth machine” normalized misinformation. Legum is careful to point out that this is not a matter of technological imperfection. It is a matter of philosophy. The platform’s very structure is designed to keep users engaged, and engagement is most easily sustained by emotion, particularly outrage. A lie that makes people angry is more powerful than a truth that makes them think.
The article also takes a hard look at “Community Notes,” X’s crowdsourced factchecking feature, which Musk and his fans often point to as proof that the platform can steer itself toward truth. Legum acknowledges that the idea has a certain democratic appeal: users, not anonymous editors, decide what is true. But in practice, Community Notes have been slow, uneven, and easy to game. A note might be written hours after a false claim has already gone viral, and even then, the system often fails to show the correction to the people who saw the original post. Sometimes notes are buried unless enough people from different ideological backgrounds agree — a requirement that sounds reasonable in theory but becomes a barrier in reality when facts themselves have become partisan. A correction that challenges a deeply felt political identity is less likely to receive the necessary approvals. And even when a note does appear, it rarely undoes the damage of the initial falsehood. The lie has already been seen, shared, believed, and absorbed into the emotional memory of the audience. Legum’s point is simple but profound: a truth machine that corrects a falsehood only after millions of people have seen it is not correcting anything. It is just keeping a record of its own failure.
There is a deeper human story here, and Legum tells it with a kind of weary urgency. We are living in a time when many people are genuinely confused about what is real. The old gatekeepers of information — newspapers, broadcasters, academic institutions — have lost much of their authority, and in their place we have been given a stream of content that is often unedited, unverified, and deliberately engineered to persuade. The rise of the “truth machine” is, in some ways, a response to that confusion. It offers a comforting promise: you do not need to be careful, because the machine will be careful for you. But Legum argues that this is a lie that exploits our vulnerability. No machine can be trusted to know the truth on our behalf, because truth is not purely an output of data. Truth requires context, interpretation, accountability, and humility. It requires people who are willing to do the difficult work of interviewing, documenting, verifying, and admitting mistakes. When we hand that responsibility to an algorithm, we lose something essential to our shared civic life. We become passive consumers of certainty — and certainty is exactly what the most manipulative voices are eager to sell us.
In the end, “The ‘truth machine’ is lying to you” is not just a warning about one social network or one billionaire’s vanity project. It is an argument about the place of truth in a democratic society. Legum urges us to stop waiting for a technological savior and instead embrace the ordinary, imperfect, and permanently unfinished work of finding the truth together. That work begins with supporting independent journalism that is transparent about its methods and accountable to its readers. It continues with teaching ourselves and our children to question sources, to notice emotional manipulation, and to tolerate ambiguity. It requires, above all, a recognition that no platform is neutral and no algorithm is wise. The “truth machine” is a seductive fiction. The truth is messier, slower, and far more demanding. But it is the only thing worth building our lives around. Until we accept that, the machine will keep lying to us — and we will keep letting it.