Now let’s get into the part I intentionally left out of the public headline.

Jacob Coxon isn’t just somebody on the internet with concerns about artificial intelligence. Before working at Anthropic, he also worked at OpenAI as a researcher and contributed to frontier AI development.

That context matters.

Coxon says he has now walked away from Anthropic because of concerns about the race toward increasingly powerful, potentially self-improving AI systems. He has argued that the companies developing this technology are moving forward despite risks that even people inside the industry do not fully understand.

And THIS is where I want us to slow down.

His resignation does not prove that AI is going to destroy humanity. I’m not interested in fearmongering or turning somebody’s warning into a fact that the receipts cannot support.

But I am interested in asking a much bigger question:

What does it mean when someone who has worked inside two of the most influential AI companies decides the race itself has become too dangerous for them to continue participating in?

That deserves more than a headline.

Because most of us experience AI from the consumer side. We see ChatGPT, Claude, image generators, automation tools, and all the ways this technology can make life easier.

Researchers like Coxon are seeing another side of it.

They are watching how quickly these systems are improving, how aggressively companies are competing, and how much pressure exists to build the next more capable model before someone else does.

And when competition becomes the driving force, we have to ask:

• Who is responsible for saying STOP?

• Who determines when a system is too powerful to release?

• What safeguards are actually mandatory versus voluntary?

• And what happens if slowing down is safer for the public but bad for business?

Those are the questions underneath this resignation.

This is also why I continue telling people that digital literacy cannot just be about learning how to USE technology. We also have to understand who is building it, who controls it, who profits from it, and who is responsible when something goes wrong.

The receipt isn’t that catastrophe is guaranteed.

The receipt is that somebody who helped build these systems believes the risk is serious enough to walk away.

And I think we should pay attention when the people inside the building start pulling the fire alarm. 👀🧾

People Lie. Receipts Don’t.™️

Source Receipts:

🧾: Jacob Coxon’s Original Resignation Thread | X
https://x.com/hilbertspaess/status/2097476196791709843

🧾: Gambling with our lives”: Anthropic researcher quits, warns against self-improving AI
https://techcrunch.com/2026/09/09/gambling-with-our-lives-anthropic-researcher-quits-warns-against-self-improving-ai/

🧾: Anthropic researchers warn about catastrophic AI risks
https://www.theguardian.com/technology/2026/sep/09/anthropic-researchers-ai-human-extinction