On September 8, 2026, Anthropic pretraining researcher Jacob Coxon published a long resignation essay. The gist is that this guy believes Anthropic and OpenAI are gambling with human life, and that current AI model research could ultimately escape human control, potentially killing all of humanity around 2030.
Another cycle indeed. Back then, those people at Anthropic left OpenAI to create a safe AI. Now look: as they keep at it, they think Anthropic isn’t safe either.
This kind of Silicon Valley elite “Effective Altruism” and “savior complex” behavior is just moral grandstanding, and it is extremely revolting.
Making a statement about an unfalsifiable matter, drawing the whole global internet to watch, standing on the moral high ground, and even being able to advance one’s career afterward, all without having to bear any actual responsibility for such a prophecy.
Every time it’s the same script: imagining oneself as the savior, making it seem as if the flame of human civilization is about to be extinguished as long as one keeps participating.