On September 9, 2026, former OpenAI and Anthropic developer Jacob Coxon resigned from Anthropic. On social media, he expressed the view that both companies are acting irresponsibly by participating in a dangerous race toward the creation of self-improving superintelligence. According to Coxon, they are risking human lives in pursuit of technologies that could become an existential threat. He stated: "I resigned from Anthropic today. I spent the last three years working on pretraining at OpenAI and Anthropic. Neither company is acting responsibly. They are racing toward self-improving superintelligence and betting our lives on it." Coxon warned that systems with superhuman capabilities may emerge soon — ones capable of hacking networks and rapidly transforming entire industries.
Coxon called for a temporary pause in the development of scaling capabilities for new frontier models. He noted that algorithm developers within the industry privately discuss fears related to the potential extinction of humanity in the coming years, even as top executives publicly attempt to downplay the panic. The researcher emphasized that "no other field of activity carries such risks." User reactions to his posts were mixed: some are skeptical of arguments about AI threats, dismissing them as absurd. One user questioned how exactly AI could destroy people, while another raised the issue of legal liability for potential consequences.
Meanwhile, on September 6, 2026, Nvidia CEO Jensen Huang stated that OpenAI's new AI model called Astra had reached the level of artificial general intelligence (AGI). However, markets reacted to the news with mixed results: CoreWeave shares rose 15%, SoftBank gained 2%, while Nvidia stock fell 2%, as investors began taking profits following Nvidia's record quarterly revenue of $96 billion. Economists such as Basil Halperin of the University of Virginia expressed doubts about whether Astra can be considered true AGI, emphasizing that modern neural networks require constant user supervision and lack human cognitive capabilities. Halperin also noted that the stock market is not a reliable indicator of technological progress, and predicted that AI will become a driver of economic growth over the next five years, adding up to 0.5% to global GDP growth.