We Have All Lost: Once Trust Is Broken, It Can Never Be Mended.

This was my immediate reaction after seeing the information on X this afternoon. Before that, I was still immersed in the infinite narcissism of an idealist who felt a deep kinship with Ilya as he spoke about his three motivations for pursuing AGI research.

So, when I saw the news, the only sentence that came to mind was: We have all lost.

On second thought, it seemed inappropriate, so I deleted it.

So much of life is slowly spent in the calculations, hesitations, and internal friction of what is "appropriate" or "inappropriate." Looking back, there is only the sigh of being paralyzed by a useless sense of self-esteem.

The more we observe the self, the more questions arise: Can the reason that makes us unique be replicated through extreme rationality? Besides memory, what do we truly possess? If we can walk toward algorithms—constantly making ourselves more rational while making algorithms capable of calculating enough so-called "contingencies" (which are actually variables of unknown causes)—can the two finally meet?

Unlike a nuclear bomb, while AGI is also an unknown and massive force, this force itself originates from humanity's quest for answers to ultimate questions. One force comes from an unknown exogenous source; the other comes from an unknown endogenous source.

Therefore, in this sense, participating in AGI itself requires an extreme degree of idealism.

However, when everything seemingly settles, it might feel like a dream for many—waking up to find that nothing has changed. But I also believe that for those who, like me, thought they were always awake, while we accept it readily, there is a sense of profound loss.

That's why Andrej Karpathy posed a soul-searching question on X: Centralization or Decentralization?

I vaguely but clearly remember 2017, when I was in a previous state of worthlessness. Only two things brought me great comfort: one was going to the riverside every day to photograph the sunrise and sunset, and the other was studying every OpenAI paper on Reinforcement Learning.

Now, even if I pick up a camera, I don't know what to photograph, and I find myself increasingly powerless to answer questions about AI.

OpenAI is nothing without its people!

Although professional habits of experience and rational calculation instinctively tell me that nothing will ever be the same again, I have been deeply moved by this sentence since seeing it on Sunday.

I hope this is true, that it is the most authentic and primitive thought and wish in everyone's heart—for the ultimate answers we are increasingly curious about, and for the papers and open AI code that once healed me.

I hope that after everything passes, we can all smile knowingly: We haven't lost yet.