About this course
Together we will read through and discuss the arguments of the book If Anyone Builds It, Everyone Dies, which sharply lays out the case that superhuman AI could end badly. Lens Coach, your AI tutor, sits alongside the text whenever you want to dig deeper or ask the questions you’d hesitate to ask a person.
No application, no selection process. You don’t need to prove yourself to us. We’ll help you prove yourself to the world.
What you’ll walk away with
The ability to explain the risk
Clearly enough to change someone's mind, not just report your worry.
A personal action plan
Built through the final units and your 1-on-1 call, matched to your skills and situation.
A verifiable certificate
Show that you understand the risks. With a public link for your LinkedIn.
A peer network in AI safety
The people you spent the course thinking alongside.
This course is for you, if:
NEWCOMER
You've heard AI is scary
You want the argument, not headlines.
You've heard AI is scary
You want the actual argument, not headlines, and to decide what you actually believe.
SELF-TAUGHT
You've read the posts
Strong opinions, shaky foundations.
You've read the posts
Podcasts, forums, late-night threads: you have strong opinions that rest on shaky foundations, and you know it.
CAREER
You're considering a pivot
Where would you be best placed?
You're considering a pivot
You want to figure out where you'd be best placed to contribute, and build the credibility to get there.
READER
You've read the book
Pressure-test the book with peers.
You've read the book
If Anyone Builds It, Everyone Dies got under your skin. Pressure-test it with people who've engaged seriously.
What you’ll explore
6 UNITS · ~5H6 UNITS · ~5H EACHYou’ll need your own copy of If Anyone Builds It, Everyone Dies (print, e-book, or audio). It’s the one thing we can’t provide; everything else is free and online.
How AI is actually built, and why nobody can see inside
Modern AI isn't programmed; it's grown from data. What that means in practice, and why even its creators can't say what a model knows or wants.
Why an AI ends up wanting things
Goals aren't installed; they emerge during training. Where wanting comes from, and why what a system ends up wanting is never quite what you trained it for.
Why it won't want what we want
Human-compatible goals are a tiny target in a vast space of possible ones. Why an AI is likely to miss it, and why "just switch it off" stops working once it's smarter than you.
How it could actually play out
A step-by-step scenario, from a broken training run to the end of humanity, with every step traced back to the mechanisms of units 1–3 and every point where humans could still have stopped it.
Why alignment isn't solved yet
Why steering an AI is a harder engineering problem than reactors, spacecraft, or computer security, and what the leading labs' plans amount to when you read them closely.
What would actually stop this
What a worldwide halt on frontier AI would take, whether it's realistic, and what people who aren't researchers can do that matters.
INCLUDED · FOR A LIMITED TIME
A 1-on-1 call with someone working on AI safety full-time
Bring your questions, your doubts, and your situation. You'll leave with a personal action plan matched to your skills. Most people who care about this get stuck exactly here.
Enroll for freeYou could read the book alone.You could read the book alone. Here's what taking the course changes.
Here's what taking the course changes.
Common questions
Is it really free?+
Yes. We're a nonprofit built to help people understand the risks of AI and decide how to respond. There's no paid tier and no upsell. The course, the coaching call, and the certificate are all free.
How much time does it take?+
About 5 hours per unit: around three hours of self-study spread out however you like (reading, Lens Coach, a short exercise), plus one 90-minute group session. Fast, minimalist readers get a unit closer to three hours.
Intensive or part-time: which should I pick?+
Same readings, same discussions, same coaching. Intensive is all six units in one Monday-Friday week: a full-time each day. It suits full immersion: between jobs, on a break, or when you don't want to wait. Part-time is one unit a week for six weeks, around work and life.
Do I need a technical background?+
No. Recent groups have included teachers, doctors, lawyers, engineers, and students from every timezone, and the discussions are richer for it. Newcomers and people who've already read the book sit in the same breakout rooms.
Is this a course that wants me to reach a particular conclusion?+
The book argues a specific and alarming thesis, and we think it deserves to be understood properly, which is different from wanting you to repeat it. What we measure is whether you can reason clearly about the problem, not whether you agree. If you finish the course convinced the risk is overstated and can say precisely why, that's the course working. Bring your skepticism to the group; it's the most useful thing in the room.
Do I need the book?+
You'll need your own copy of If Anyone Builds It, Everyone Dies (print, e-book, or audio). It's the one thing we can't provide. Everything else is free and online.
What if I can't make one of my sessions?+
Life happens. Message Lens Coach and you can swap into another group's session for that unit, or reschedule. We'll help you find a time that works.
What happens after the course?+
You keep your peer network and the community Discord, Lens Coach stays available, and when you're ready to go deeper, the advanced courses (Forecasting, Modeling, and Shaping AI Futures, or Compute Verification) are the natural next step.
Not sure it’s for you? Come to the first session. If it’s not the right fit, we’ll help you find something that is.