Artificial general intelligence reminds me of the old parable of the blind men and the elephant. One touches its side and calls it a wall; another touches its tail and describes something entirely different. Nobody agrees on its shape or even exactly what it is, yet nobody doubts that the elephant is there. The real choice, then, is which side of it we choose to see.

From the beginning, we chose to look at AGI through the lens of discovery. We defined our goal that way for a simple reason: I believe AI exists to benefit people, and an AI that helps humanity discover what it could not discover alone is the clearest form of that benefit I can imagine.

So this is what I hope for. In 907 days, I want Apodex to be capable of enabling discoveries worthy of the world’s highest scientific honors, continuously and at scale. If we succeed, the impact could extend far beyond science and benefit society as a whole.

Can AI actually discover? I believe it can. My conviction is that if we scale a Discovery Engine far enough, real discoveries will emerge. I cannot prove that today; it is still a leap of faith. But AI is, in a sense, the crystallization of thousands of years of accumulated human intelligence, and I believe that potential is worth taking the leap for.

There is a fair question: if AI can discover, does it take discovery away from us? If the world suddenly had a thousand more Einsteins, would the original Einstein be out of work? Obviously not.

I recently read David Deutsch’s The Beginning of Infinity. One of its central ideas is that knowledge is our understanding of the world, and because there is no final boundary to what can be understood, knowledge itself has no end. Every answer creates new questions.

That is why an AI capable of discovery takes nothing away from humanity. Discovery is not a fixed supply that AI can consume before we get there. Every genuine discovery expands the frontier and creates more territory for everyone else to explore.

The greater risk may be the opposite.

Modern scientific discovery, as we practice it, is only a few hundred years old, and physics has gone decades without a genuine paradigm shift. If AI cannot help us break through such limits, humanity may eventually find itself trapped beneath a technological ceiling—like the civilization in The Three-Body Problem.

Now we have AI.

We should use it to break that ceiling.

Tianqiao Chen