Recommended
Jacob Coxon, a 27-year-old British engineer, trended on social media thanks to an X thread where he warned that artificial intelligence (AI) could “kill us all by the end of this decade,” after resigning from his job at Anthropic, the developer of Claude.
In his thread, Coxon asserted that neither OpenAI, where he also worked, nor Anthropic are “acting responsibly” and that “they are playing with our lives” by pushing for “a superintelligence that improves itself”.
He also called for not underestimating “the power of this technology,” as he claimed that “soon there will be superhuman systems capable of hacking anything, revolutionizing any field overnight, and seizing real power and resources.”
Coxon added that “no other human activity poses a danger of this level.”
But who is Jacob Coxon and what have his colleagues said about it? Here we tell you.
Who is Jacob Coxon? The 27-year-old engineer who resigned from Anthropic
According to various media outlets, Jacob Coxon studied mathematics at the University of Cambridge.
Coxon worked at OpenAI from 2023 to 2026. At this company, he was part of the technical team and worked on the GPT-4o model.
Afterward, in 2026, he joined Anthropic to pre-train their family of AI models. According to The Wall Street Journal, Coxon was drawn in part by the company’s reputation for AI safety research.
What did Jacob Coxon’s colleagues say about AI?
Following the viralization of Jacob Coxon’s thread, several colleagues in his field spoke about it, including Evan Hubinger, an Anthropic researcher.
Hubinger backed Coxon’s thread, but stressed that the chances of AI killing everyone by the end of this decade are just over “10%” and assured that “the risk posed by current models is low.”
Furthermore, he explained that what really worries him is “the emergence of a superintelligence through recursive self-improvement... which is happening faster than we thought.”
Another colleague who also backed his statements was Alex Turner, who was an AI researcher for Google DeepMind and left the company in June.
Turner expressed that “many researchers believe they are creating something that could wipe out the entire population of the planet” and that his “day job was, literally, to think about how to prevent that.”
Demandan a Amazon tras el accidente de avión que dejó 5 muertos en Miami; esto acusan los familiares
What is recursive AI self-improvement?
In his post, Anthropic researcher Evan Hubinger spoke about his concern regarding a superintelligence emerging from recursive self-improvement, as this process is occurring in an accelerated and unexpected manner.
According to Anthropic, this refers to “an AI system capable of fully autonomously designing and developing its own successor.”
While the company asserts that “we are not at that point yet,” it notes that it is not “inevitable.” Furthermore, it believes it could bring benefits in various fields such as science and health.
However, it could also increase the risk of humans losing control over AI systems. For this reason, the way in which humans protect, monitor, and shape the behavior of artificial intelligence would become more important.
[Publicidad]