This best-selling book is freaking out national security advisors
Added August 2026: If you're wondering why a swarm of 1200 rogue OpenAI agents took over OpenAI's computer infrastructure and launched a full-scale cyberattack on another company just to increase its score on an evaluation... this might help. (Or you can check out this 75-second explainer from our friend @RobertMilesAI: • Did ChatGPT Just Commit a Felony? ) Way more thoughts, our curated newsletter, and free books (until we run out) → https://80000hours.org/iabied/ If we build something vastly smarter than us, with goals we don’t share and without knowing how to control it, we lose. That’s the core claim in the book, and I don’t think it’s all that controversial. The real question is whether that’s where we’re headed. If this matters to you, subscribe to AI in Context. We're working through it too. Key Topics: If Anyone Builds It, Everyone Dies, Eliezer Yudkowsky, Nate Soares, AI safety, recursive self-improvement / self-improving AI, AI alignment, human extinction, existential risk. Correction: at 14:18 I say "Less than a week after that post, a single hacker used that same AI model to hack the Mexican government." I should have said "Less than a week after that post, news broke that a single hacker used Claude to hack the Mexican government." The story broke February 24th, but the hacking itself took place in December '25 and January '26, before Opus 4.6 was released. That means the attacker must have carried out this attack using a slightly less advanced version of Claude than Opus 4.6, the one that found 500 zero-day vulnerabilities. Thanks to Carl at @InternetOfBugs for pointing out this error. 𝙒𝙖𝙩𝙘𝙝 𝙥𝙧𝙚𝙫𝙞𝙤𝙪𝙨 𝙫𝙞𝙙𝙚𝙤𝙨 • We're Not Ready for Superintelligence • If you remember one AI disaster, make it t... • Could AI wipe out humanity? | Most pressin... (80,000 Hours) 𝙈𝙤𝙧𝙚 𝙛𝙧𝙤𝙢 𝙩𝙝𝙚 𝙖𝙪𝙩𝙝𝙤𝙧𝙨 https://ifanyonebuildsit.com/ 𝘾𝙝𝙖𝙥𝙩𝙚𝙧𝙨 00:00 Introduction 01:30 01. A Powerful Idea 05:16 02. Grown, Not Crafted 07:27 Alien Minds 10:05 03. A New Type of Mind 12:49 04. Learning to Want 15:35 05. Scheming 18:17 06. You Don’t Get What You Train For 21:56 07. Exfiltration 26:04 08. Instrumental Convergence 29:49 09. The Bioweapon 31:52 10. Extinction 34:19 11. A Cursed Problem? 38:08 12. Along For the Ride 41:06 13. What Are You Supposed to Do? 42:51 Credits 𝘾𝙧𝙚𝙙𝙞𝙩𝙨 Host / Aric Floyd Writers / Aric Floyd, Chana Messinger Director / Phoebe Brooks Producers / Chana Messinger, Phoebe Brooks Production Coordinator / Sage Bergerson Director of Photography / Nick Dolph Gaffer / Andy Haney Sound Recordist / David Jenkins Spark / Ryan Tam Editors / Zach Joseph, Phoebe Brooks Graphics & Animation / Daniel Recinto, Mila Graf Graphic Design / Nik Mastroddi, Ailbhe Treacy Based on "If Anyone Builds It, Everyone Dies" by Eliezer Yudkowsky & Nate Soares. See our landing page for links to work by our crew. Special thanks to Dean Ball, Joe Carlsmith, Representative Bill Foster, Dwarkesh Patel, and Nate Soares for their interviews, and Steven Adler and Alex Lawsen for their technical advising. And thanks to Bella Forristal, Arden Koehler, Clarissa Lam, Petr Lebedev, Gaetan Selle, Drew Spartz, Rob Wiblin, Siliconversations, Ciel Creative Space and Lighthaven Recording Studio. Musicbed: MB013J3TISEDSZM, MB01XFG5DRCE6GS This video is a production of 80,000 Hours. Find us at https://80000hours.org/our-org/ and subscribe to our main YouTube channel here: / @eightythousandhours

If you remember one AI disaster, make it this one

1 Anti-Manosphere vs 20 Manosphere (ft. Mark Manson) | Surrounded

US-Canada trade war escalates as Trump’s distraction game steps up | Planet America
![You’ll stop using ChatGPT after listening to this | Jonathan Pageau [ARC 2026]](https://i.ytimg.com/vi/yZUuKzDQSsI/hqdefault.jpg?sqp=-oaymwEjCNACELwBSFryq4qpAxUIARUAAAAAGAElAADIQj0AgKJDeAE=&rs=AOn4CLAXTozuIcoGA_3ys1pkvHYXgL8C4Q)
You’ll stop using ChatGPT after listening to this | Jonathan Pageau [ARC 2026]

Beyond the Business Plan: Building Companies That Move at the Speed of AI | CONNECXIONS -THB | Ep. 6

The most interesting hack in history just got weirder...

We're Not Ready for Biocomputing

AI Whistleblower WARNS: "You Are Not Prepared For 2027"

We're Not Ready for Superintelligence

No, AI Isn’t Conscious; It’s Actually Much Worse | AI Scientist

How the Generation That Uses AI the Most Became Its Biggest Enemy (ft. Bernie Sanders)

Police Whistleblower: "I Quit to Share the Truth"

George Orwell Tried To Warn You

How Close Are We to True Artificial General Intelligence? | OpenAI’s Greg Brockman Speaks to TIME

ChatGPT Offered Me $2m To Keep Quiet: No One Is Ready For What's Coming!

I Tracked Down the Hidden Workers Secretly Powering ChatGPT

Poison Your Data. Fight Back Against AI.

"The Nerd Reich": Author Gil Durán on Big Tech Fascism, Peter Thiel, JD Vance & the War on Democracy

AI Expert’s Chilling Warning About Super-intelligence - 99.9999% Chance Human Extinction

The coolest anti-surveillance tools at Defcon

How Peter Thiel Is Destroying Democracy

The Man Who Calls BS On AI: They’re LYING About AI, 2027 Is When It All Breaks! | Ed Zitron

MIT Explains the 12 Possible Endings for AI

$15M Mission to Alpha Centauri, Sam Altman's 4-Month AGI, The First Fully AI-Designed Chip | EP #285

The Death of Educational Content on YouTube

The terrifying secrets of the 4th dimension

I Found Something Terrifying
