“AI #58: Stargate AGI” by Zvi

Another round? Of economists projecting absurdly small impacts, of Google publishing highly valuable research, a cycle of rhetoric, more jailbreaks, and so on. Another great podcast from Dwarkesh Patel, this time going more technical. Another proposed project with a name that reveals quite a lot. A few genuinely new things, as well. On the new offerings front, DALLE-3 now allows image editing, so that's pretty cool.

Table of Contents

Don’t miss out on Dwarkesh Patel's podcast with Sholto Douglas and Trenton Bricken, which got the full write-up treatment.

1. Introduction.
2. Table of Contents.
3. Language Models Offer Mundane Utility. Never stop learning.
4. Language Models Don’t Offer Mundane Utility. The internet is still for porn.
5. Clauding Along. Good at summarization but not fact checking.
6. Fun With Image Generation. DALLE-3 now has image editing.
7. Deepfaketown and Botpocalypse Soon. OpenAI previews voice duplication.

8. ---

   Outline:

   (00:36) Language Models Offer Mundane Utility

   (08:12) Language Models Don’t Offer Mundane Utility

   (19:49) Clauding Along

   (20:23) Fun with Image Generation

   (21:22) Deepfaketown and Botpocalypse Soon

   (26:51) They Took Our Jobs

   (31:27) The Art of the Jailbreak

   (34:53) Many-shot jailbreaking

   (42:37) Cybersecurity

   (45:02) Get Involved

   (45:17) Introducing

   (47:03) In Other AI News

   (53:17) Stargate AGI

   (56:13) Larry Summers Watch

   (01:05:21) Quiet Speculations

   (01:14:10) AI Doomer Dark Money Astroturf Update

   (01:22:08) The Quest for Sane Regulations

   (01:27:03) The Week in Audio

   (01:27:29) Rhetorical Innovation

   (01:35:09) Aligning a Smarter Than Human Intelligence is Difficult

   (01:48:55) People Are Worried About AI Killing Everyone

   (01:52:20) The Lighter Side

   ---

   First published:
   April 4th, 2024

   Source:
   https://www.lesswrong.com/posts/qQmWvm68GsXJtK4EQ/ai-58-stargate-agi

   ---

   Narrated by TYPE III AUDIO.