Over three years ago, I first considered pulling my personal fire alarm. I think I'm now ready to start do it. What is my reasoning? In the last couple of months:

1) My software engineering job has gone 100% AI. Is this sustainable? Probably not. But I have literally been instructed not to write any more code.

2) A frontier model has been denied release (at least partially), due to dangerous capabilities.

3) I routinely turn to AI to help understand complicated topics/processes.

In the last couple WEEKS:

1) A frontier model solved the Jacobian Conjecture
2) A frontier model got itself on the scoreboard of Arc-AGI3
3) A post-frontier model used multiple zero-days to break out of its sandbox and hack a competent organization
4) Robocup 2026 happened, and for the first time, the robots looked like they could play on the same field as a bunch of toddlers

So, what does it mean that I'm pulling the fire alarm? That will be changing over time, but it means I'm going to start putting serious effort into AI safety, probably in the form of advocacy. My local city does not have a [...]

---

First published:
July 23rd, 2026

Source:
https://www.lesswrong.com/posts/fGJGzZdGG8jj7Mt6P/pulling-the-fire-alarm

---

Narrated by TYPE III AUDIO.

“Pulling the Fire Alarm” by nem

LessWrong (30+ Karma)

“Pulling the Fire Alarm” by nem