Why We're Building The TRUE Truths

(Gestures wildly at nothing and everything.)

If you’ve felt that impulse lately - that sweeping hand motion that tries to encompass the sheer volume of chaos in our political discourse - you’re not alone. We’ve been feeling it too.

The 47th President of the United States is talking about a third term. He’s floating the idea of resuming nuclear weapons testing. He’s launching ill-planned tariffs that economists warn could destroy our economy. He’s threatening to send the military into “blue cities.” And somehow, people keep acting like this is normal.

It’s not normal. And we needed a way to prove it.

The Problem With One-Off Fact Checks

Here’s the thing about calling out individual problematic posts: it doesn’t show the pattern.

You can fact-check one false claim. Media outlets can write an article about one inflammatory statement. Twitter can have a field day with one particularly egregious Truth Social post. But then what? It gets lost in the noise. By tomorrow, there’s another one. And another. And another.

The former President-turned-current-President has become somewhat predictable if you pay attention. His accusations often read like confessions in disguise. He posts impulsively, stream-of-consciousness style. While some supporters believe he’s playing 5D chess, we think the reality is far simpler: he says what he thinks, when he thinks it, without much filter.

That’s actually important information. But you can’t see the forest for the trees when you’re just reacting to individual posts.

We needed the data. All of it. Tracked systematically. In real-time.

Why We Built This

A few reasons, if we’re being honest.

First, because some of the shit he posts is so egregiously false or dangerous that it’s mind-blowing.

Second, because too many people take what he says at face-value. They don’t realize - or don’t want to acknowledge - how often he posts lies or genuinely problematic content. Without the pattern documented, it’s easy to dismiss concerns as partisan noise.

Third, and more practically: AI is getting better, cheaper, and more ubiquitous than ever before. We’d been wanting a project where we could incorporate AI into some automation flows, and this one fit the bill perfectly.

And finally, we wanted to capture the posts he deletes. Having data is better than not having data. You can’t track trends without data.

The Bias Question (Let’s Just Address It)

Look, we’re not impartial. We’re not going to pretend to be.

We built this because we think there’s a problem that needs documenting. We’re motivated by content we find false or dangerous. We use words like “bullshit” when we talk about this project internally.

But here’s the key: systems should be relatively impartial, even if their creators aren’t.

We don’t need to build a biased system to catch and call out problematic content. An honest system will do that just fine on its own - if the content is actually problematic.

That’s why we went AI-first. Not AI-only, but AI-first.

How It Works (The Short Version)

Every 5 minutes, our system checks the President’s Truth Social RSS feed. When there’s a new post, we:

Capture the full content, including any images
Enrich it with context from recent news
Run it through Claude AI to analyze it across four categories:
- Constitutional Concerns
- Discriminatory Language
- Conflicts of Interest
- Misinformation
The AI assigns a severity rating from None to Critical for each category, along with an explanation of its reasoning.

Then - and this is important - we publish it immediately. No human editorial review before it goes live. No waiting for a committee to approve the analysis. Real-time accountability.

Does the AI get it wrong sometimes? Absolutely. That’s why we layer in human oversight after the fact. We can add editorial notes, corrections, and additional context. But we don’t gate the initial analysis.

Because perfect is the enemy of good, as they say. And in times like this, getting something out there that’s likely to call out problematic content has value, even if it occasionally requires a human to walk back some of the takes.

Why Full Automation?

We’ve started too many projects that don’t ship because we set the barrier to launch too high.

This time, we decided: ship first, iterate later.

Going AI-first also meant we could have relatively consistent and impartial analysis happening in real-time, without delay. No waiting for editors to be available. No bottleneck of human review. No accusations that we’re cherry-picking which posts to analyze.

Every post gets the same treatment. Every time. Automatically.

Like most things in the future, the real secret sauce is going to be how machine intelligence and bio-intelligence work together. We’re not replacing human judgment - we’re augmenting it. The AI does the heavy lifting of consistent, rapid analysis. Humans provide context, corrections, and deeper insight.

What We Hope This Becomes

We want critical thinking. Data-driven decisions.

We want people to be able to see patterns, not just react to individual moments. We want journalists to have a resource. We want citizens to have access to systematic analysis, not just hot takes.

And maybe - dare we wish it - we want a bit more of the “old normal” back. Some civility. Some shared reality about what’s true and what’s not.

That might be too far-reaching. But we had to try something.

The Risk Calculation

Are we nervous about putting this out there?

Oh boy. We try not to think about it.

We’re building a tool that systematically flags the sitting President’s posts as potentially false or dangerous. That’s not nothing. But the alternative - watching the patterns unfold without documentation, without accountability - felt worse.

So here we are.

What’s Next

The system has been live for a few weeks now. We’re still iterating, still improving the analysis categories, still building out features (like tracking deleted posts, which we haven’t quite implemented yet but definitely plan to).

This is version 1.0 of something we hope grows into a genuine resource for accountability and transparency.

We invite you to explore the dashboard. See what patterns emerge. Draw your own conclusions from the data. And yes, call us out if you think we’re getting something wrong - that’s kind of the whole point.

We’ll be launching a mailing list soon for updates. In the meantime, the data is live, public, and updating every 5 minutes.

Because in a moment where we’re all gesturing wildly at everything, sometimes the most useful thing you can do is point at something specific and say: “Look. This is what’s actually happening. Here’s the receipts.”

The TRUE Truths is an independent civic transparency project. Our methodology, code, and analysis criteria are fully public and open to scrutiny. We believe sunlight is the best disinfectant - even when it’s uncomfortable.