Former AI researcher Jacob Coxon has sounded a dire alarm, claiming that leading AI firms are racing toward an uncontrollable superintelligence that could pose an existential threat to humanity.
- Jacob Coxon resigned from both Anthropic and OpenAI after three years of pretraining research.
- Claims that companies are prioritizing a 'race to superintelligence' over human safety.
- Warns that internal executives privately fear catastrophic outcomes despite public optimism.
In a move that has sent ripples through the tech industry, 27-year-old researcher Jacob Coxon has resigned from Anthropic, following a previous tenure at OpenAI. Coxon, who specialized in pretraining research, asserts that the current trajectory of artificial intelligence development is not merely ambitious but dangerously reckless.
Taking to X (formerly Twitter), Coxon alleged that neither company is acting responsibly. He argued that the pursuit of self-improving superintelligence is akin to gambling with human existence. According to Coxon, future AI systems could possess the capability to hack secure infrastructures, revolutionize entire scientific disciplines overnight, and amass unprecedented real-world power.
Why This Matters
BozokMedia analysis shows that this resignation highlights a growing schism between corporate marketing and internal scientific anxiety. The 'AI arms race' creates a prisoner's dilemma: companies feel forced to accelerate development because they fear a competitor will release a dangerous, unaligned AI first. This systemic pressure effectively neutralizes the safety protocols these companies claim to uphold.
"We are building the bomb again, only this time, the bomb thinks for itself."
Coxon provided a nuanced comparison between the two giants. He noted that while OpenAI may lack a deep internal grasp of the civilizational stakes, Anthropic is acutely aware of them. However, he claims that Anthropic's awareness has not translated into caution, as the company remains trapped in a competitive spiral to maintain its market position.
Company Perspective Comparison
| Metric | OpenAI | Anthropic |
|---|---|---|
| Risk Internalization | Lower (per Coxon) | High Awareness |
| Primary Driver | Rapid Capability Scaling | Competitive Safety Race |
| Internal Culture | Growth-oriented | Cautionary but Competitive |
Frequently Asked Questions
Q1: Why does Jacob Coxon believe AI is more dangerous than previous technologies?
A: He argues that the ability of AI to self-improve and autonomously acquire resources makes it unlike any other human invention in history.
Q2: What is the 'private fear' mentioned by the researcher?
A: Coxon claims that while executives sound sensible in press releases, they privately express genuine fear regarding the potential for catastrophic outcomes.
