Patreon: https://www.patreon.com/primerlearning
80,000 Hours: https://www.80000hours.org/primer
https://www.desmos.com/calculator/a5pfjtr4tr
Other connections:
Discord: https://discord.gg/NbruaNW
Twitch: https://www.twitch.tv/justin_helps
Store: https://store.dftba.com/collections/primer
Reddit: https://www.reddit.com/r/primerlearning/
Bsky: https://bsky.app/profile/justinhelps.bsky.social
Twitter: https://twitter.com/primerlearning
Links to other resources:
Reasoning through arguments against taking AI safety seriously
https://www.youtube.com/c/robertmilesai
https://www.youtube.com/@Siliconversations
https://www.youtube.com/@Go-Meta
https://www.youtube.com/@DwarkeshPatel
https://www.youtube.com/@Alex.kantrowitz
Nicky Case: https://aisafety.dance/
https://safe.ai/
https://bluedot.org/courses
https://www.aisafetybook.com/
http://aisafety.com
http://aisafety.info
0:00 Intro
1:50 Intelligence is broad ability
3:55 You don’t need to be conscious to act toward a goal
6:18 Beyond human intelligence
8:44 A human with computer powers
18:45 The AI persuasion thing
19:33 Power seeking and self preservation
22:17 Value communication
24:27 Value formation
27:44 Theoretically building a human-level machine
30:23 Neural networks
34:39 Reinforcement learning
37:45 Reasoning models
41:29 Multi modal models
43:40 Agentic models
45:49 Improvement prospects
47:26 Interpretability
49:26 Doom probabilities
53:40 Weighing the upsides and downsides
57:10 Places to continue learning
