AI Stole My Job. Now I'm Trying to Tame It.

My year got off to quite a start, to say the least.

It began in early February, when I was hit with the realization that the world was going to change radically in the near future because of artificial intelligence. Overnight, the majority of my time at work was being spent creating data so that AI could learn to replace me in the parts of my job I loved most — and nothing suggested this was a temporary situation. Quite the opposite. While I had hoped to grow my career, to keep learning and have more and more impact in my field, I was realizing I was going to become a kind of supervisor whose job was to occasionally slap AI on the wrist when it screwed up. Until the day I'd be let go once clients figured out they could get a similar service for free on claude.ai.

Some people decide to stick it out until things fall apart, and I get it. But I had the freedom to leave my job without ruining myself financially, so that's what I did. Then I turned to 80,000 Hours to help me figure out what came next.

How to Have an Impact in Your Career

80,000 Hours is my idol. I regularly listen to their podcast — partly despite and partly because their episodes run for hours — and they're the ones who sent me, for free, the book that convinced me to have children. The organization takes its name from the fact that 80,000 hours is roughly how much time we spend on our careers over a lifetime. It helps people pursue careers that can have the greatest positive impact on humanity, and it identifies the most pressing problems using three criteria:

1. Addressing them would have a huge impact
2. Solutions can be implemented relatively easily
3. Few people are working on them

To date, they believe the five most pressing problems are, in order of importance:

1. Power-seeking AI systems
2. Extreme concentration of power
3. Deliberately engineered pandemics
4. A potential conflict between major powers
5. Factory farming

You read that right. The risks posed by power-seeking AI systems are the number one priority, according to an organization that I believe does extremely rigorous work and is guided by excellent values. By far. They make very compelling arguments, and I encourage you to read their materials— but if I had to boil it down to the bare essentials, here's what I'd say:

• Humanity is investing billions and billions of dollars into developing systems that could very well become more intelligent than us. In other words, humans could soon — within less than ten years — no longer be the most intelligent species on Earth.
• And it's moving incredibly fast. So fast that there's no good reason to doubt that, within the next decade, AI will surpass any human in every single domain.
• Obviously, this would have a few consequences for humanity. Including the risk that AI might, broadly speaking, want to wipe us out. Why? Maybe to pursue some strange goal it set for itself. Maybe because it was given the wrong values, by accident or by design. Or maybe because we'd be nothing but chickens by comparison, and it simply wouldn't care about us.

And after saying all that, I often feel like laughing. "AI might want to destroy humanity and actually pull it off, hahaha. So I kind of want to work on preventing that. Hahaha."

But it's not funny, and I'm dead serious. The more I learn about this, the more convinced I am of how important it is to tackle this problem. And even if it's not necessarily the issue that hits me hardest emotionally — industrial animal farming gives me far more nightmares, for what it's worth — I believe it's the one I'm best qualified to work on. Over the years, I've discovered that I love coding, and that I'm good at it. I've learned to manipulate machines so that they themselves can manipulate language. I have a master's degree in computational linguistics. So despite my friend impostor syndrome who likes to taunt me from time to time, reason tells me there's a real chance I could contribute to solving the problem. I don't think there are many linguists working in this space, which strikes me as a little odd, because after all, language models manipulate… language.

So here's what I've been up to since that realization hit me in February and I left my job:

I'm training myself

I'm taking my CEGEP math courses (which I deeply regret not having taken at the time), because without a foundation in math, it's hard to understand how AI actually works. This takes up the majority of my time.

I'm working through theARENAcurriculum. ARENA is a course designed to train AI safety researchers. It's an incredibly comprehensive program, offered in person in England, but freely available online — which is how I'm doing it, because no, with a baby at home, I don't have the luxury of going to England for two months.

I completed BlueDot Impact's free online course, The Future of AI, which I highly recommend — it's absolutely fascinating.

I do LeetCodefrom time to time: programming exercises that are the industry standard. Many employers base their technical assessments on problems very similar to LeetCode.

I pay for ClaudeCode and ClaudeCowork, and I have zero regrets. Claude is the best math and programming tutor I could have asked for.

I'm getting rejected

One of the first things I did was apply to BlueDot Impact's AGI Strategy course: a well-regarded program in the field, free, online, and one week long full-time. I got turned down. I reapplied two weeks ago, figuring my chances were better this time — I had two more months of learning and legwork under my belt, after all. And once again, I got a very polite no. I'll never know exactly why, because so many people apply that they don't have time to give feedback. But they did note that their decision "is not a judgment of my potential." Phew!

As it turned out, my mother knows someone at LawZero, a non-profit founded by none other than Yoshua Bengio, the most cited living researcher in the world across all fields. I spoke with this person, got very excited, and applied. Naturally, they didn't take me on — I'm just a newbie — but I'm not giving up.

I applied for OpenAI's Safety Fellowship and made it through the first round, but not the second. According to their rejection email, there were over 100 applicants per open position, and the people selected had done things like run their own AI safety experiments and published their work on GitHub, written articles on the subject… You know, totally accessible things you can just do in your spare time.

But OpenAI is clearly not a company that's proving it takes the issue seriously: it dissolved its team dedicated to long-term AI risks and agreed to let Trump use its models for surveillance and on the battlefield— so maybe it's for the best. (Even if they pay their interns $4,000 USD a week!!! It would have been really hard to say no to that.)

I'm meeting the community in Montreal

This is brand new: I discovered that an AI safety community is starting to take shape in Montreal! I'm so grateful for this, because I was beginning to wonder whether it was even possible to work in the field without living in California or London. Which struck me as absurd, because Montreal is, after all, one of the world's capitals of artificial intelligence.

It's happening at Ω Labs, a dedicated space that opened in May 2026. The kind of initiative that gives me hope.

I'm consuming everything I can on AI safety

I've listened to dozens of hours of 80,000 Hours podcasts, including interviews with Yoshua Bengio, Rohin Shah, Will MacAskill, and Ajeya Cotra, among others.

I read Holden Karnofsky's blog seriesThe Most Important Centuryon my e-reader.

I read 80,000 Hours: How to Have a Fulfilling Career That Does Good (you can get it for free here), a career guide that distills the philosophy and wisdom of 80,000 Hours. I was already pretty familiar with the content since I'm a fan, but I really enjoyed it.

I'm reading Max Tegmark's Life 3.0.

I read blog posts on LessWrong, a go-to reference in the field. It's fascinating to watch nerds argue about concepts I barely understand.

I'm trying to find mentorship

I applied to the MATSresearch program. I made it through the first round!

I'm staying balanced between excitement and panic

Through all of this, I'm navigating the highs and lows of being unemployed and trying to change careers entirely on my own.

I love spending my time doing exactly what I want and learning so much about what I now believe is the most exciting field in the world right now — and often I tell myself it's going to work out, I have a shot, go Naïma. But there are just as many days when I feel like my brain has the processing power of a snail, and I listen to brilliant people talk and wonder what planet they're from.

Will I manage to break into the field? Someday, probably. In any case, I can't imagine a better use of my time.