Intro

Note: I used LLMs to draft different parts of this. I've checked almost everything, but there might be some mistakes remaining.

Apologies for posting this on Christmas Eve. I wanted to get this out the door before the end of the year. Questions welcome, and if it's easy to pull metrics to answer them, I will.

Summary

80,000 Hours launched a video program in 2025 focused on longform, cinematic, personality-driven content about AI risks. Our first two longform releases were:

• We're Not Ready for Superintelligence (the "AI 2027" video): 8.9M views, ~1.4M watch hours
• If you remember one AI disaster, make it this one (the "MechaHitler" video): 2.7M views, ~419K watch hours

Both videos significantly outperformed our expectations (we'd anticipated 15-50K views for the first). The cost per engagement hour ($0.11 and $0.39 respectively, including staff time) compares favorably to other 80,000 Hours programs.

This post covers: what we spent, what we got, why we think it worked, and what we'd do differently.

The numbers

Costs

CategoryAI 2027MechaHitlerDirect costs~$50K~$64KStaff hours~450 hrs~450 hrs (Note, I’m assuming it's about the same as for AI 2027, I didn't re-ask people how much time they spent.)Total cost (making some assumptions about we should incorporate staff [...]

---

Outline:

(00:34) Summary

(01:33) The numbers

(01:36) Costs

(02:16) Timing

(02:40) Results

(03:46) How valuable is a video watch hour?

(04:24) Qualitative Feedback

(04:28) AI 2027

(05:51) MechaHitler

(06:12) YouTube commenters like:

(06:52) What the comments don't like:

(07:17) Qualitative Analysis

(07:21) Why we think AI 2027 did well

(09:56) Why MechaHitler did less well (but still well)

(10:50) Lessons Learned

(10:54) Overall what we think matters

(11:25) Our guess at what's less important (though we're certainly unsure, maybe if we nailed these, we'd get more success)

(12:24) How our production works

(12:43) The timeline

(13:32) Ideation

(14:06) Scripting

(14:57) Shooting

(15:31) Reshoots / Voiceover

(15:45) Editing

(16:06) Launch

(17:00) What we're still figuring out

(17:36) Closing thoughts

---

First published:
December 24th, 2025

Source:
https://forum.effectivealtruism.org/posts/RCRaBYSqBaMzHzTjF/untitled-retrospective-and-learnings-from-ai-in-context-s

---

Narrated by TYPE III AUDIO.

---