⚠️🚨 Anthropic researcher Jacob Coxon has reportedly resigned after three years working on AI pretraining research across both Anthropic and OpenAI, raising serious concerns about the race toward increasingly powerful, self-improving AI systems. Coxon warns that the technology being developed could eventually surpass human capabilities, gain the ability to carry out highly sophisticated cyberattacks, accelerate breakthroughs across multiple fields, and potentially acquire significant influence and resources. He also says that discussions inside leading AI labs include the possibility that sufficiently advanced AI could pose an existential threat to humanity before the end of this decade. According to Coxon, he believes many people at OpenAI may not fully appreciate the scale of the potential danger. He says Anthropic appears more aware of these risks, yet still faces pressure to continue moving quickly because of fears that competitors will push ahead. Coxon argues that decisions a...