Wes R

• Wes RJun 11, 2025, 7:23 PM

  0 points

  0

  on: The Best Refer­ence Works for Every Subject

  Domain: AI safety (better known as “The alignment problem”)
  
  Links:
  here’s 16 minute’s worth of overview that basically covers all you’d need to know about what AI safety specifically is:

  1. Could AI wipe out humanity? | Most pressing problems (10 mins),

  2. The real risk of AI (6 mins)

  3. Look at this graph:

     

  If you’d be curious to learn more about the nitty-gritty of AI safety, watch (in this order):

  1. Intro to AI Safety, Remastered (18 mins) (Sidenote: The original video is from 2016, & back then people thought AI would develop much slower.)

  2. The OTHER AI Alignment Problem: Mesa-Optimizers and Inner Alignment (23 mins)

  3. Using Dangerous AI, But Safely? (30 mins)

  Author(s): 80000 Hours, Siliconversations, Metaculus, Rob Miles.
  Type: list
  Why: To use as a quick intro

• Wes RMay 26, 2025, 9:42 PM

  1 point

  0

  on: Wes R’s Shortform

  Had a conversation on AI safety (namely interperetability of AIs thinking, and what might cause AI to scheme), and I screenshot the key takeaways so I could share them with the world (well, mostly just so I have them saved somewhere).

  

  

  

  

  

• Wes RMay 21, 2025, 3:42 AM

  1 point

  0

  on: 80 con­cepts on my new ver­sion of AI: DecisionBots

  

• Wes RMay 9, 2025, 6:18 PM

  1 point

  0

  on: 80 con­cepts on my new ver­sion of AI: DecisionBots

  I reckon I might be pretty good(ish) at coming up with good concepts, but given my downvotes, it seems I’m not good at presenting said concepts. Does anyone have any advice for that?