I used an AI for 7 months to search for a Theory of Everything. I failed. And it’s the best thing that could have happened.

How Exploring with AI Led Me to a Hidden Truth—and Why Failure Is a Vital Part of Scientific Discovery

Over the past seven months, I embarked on an unconventional journey: using artificial intelligence as a partner in my quest to unravel the universe’s deepest secrets. My goal was to develop a comprehensive Theory of Everything, but the most valuable outcome wasn’t the theory itself—it was the profound insights and personal growth that emerged from this process. Today, I want to share what I learned about working with AI as a tool for rigorous inquiry, and why embracing failure can be one of the most honest steps toward genuine understanding.

Rethinking AI: Beyond a Calculating Machine

Many discussions around AI tend to frame it as a kind of equation solver or an algorithmic shortcut—an easy way to generate answers or validate pre-existing beliefs. However, after seven months of deep collaboration, I’ve come to see AI’s true potential differently. Instead of viewing it as a source of certainty, I’ve learned to treat AI as a mirror that reflects the clarity and flaws of my own reasoning.

The key isn’t how often or whether we use AI, but how we engage with it. If we approach it as an oracle that confirms our beliefs, we risk falling into confirmation bias. Conversely, if we leverage it to challenge and scrutinize our ideas—asking, “Where could I be wrong?”—then it becomes a powerful tool for scientific rigor.

The Process: From Hope to Hard Truth

Initially, I approached my hypothesis with enthusiasm. The idea that the universe might be described by a dynamic “ether” seemed promising, and AI helped generate seemingly compelling results. It felt magical—almost too good to be true. It would have been easy to cling to these results as validation.

Instead, I adopted a different mindset: I used AI as an opponent that demanded rigorous testing. I asked it to identify inconsistencies, to challenge the assumptions underpinning my theory, and to probe deeply into the data.

This approach pushed me to learn new skills, such as coding in Python, transforming my understanding of data analysis and scientific method. But more importantly, it taught me to accept that failure isn’t a setback—it’s an essential part of discovery. When the data finally shattered my beloved hypothesis, it was a moment of brutal honesty, revealing that my concept didn’t hold up against real-world evidence.

The Value of Discarding Certainties

That failure was, paradoxically, the most truthful