“AI #106: Not so Fast” by Zvi

This was GPT-4.5 week. That model is not so fast, and isn’t that much progress, but it definitely has its charms.

A judge delivered a different kind of Not So Fast back to OpenAI, threatening the viability of their conversion to a for-profit company. Apple is moving remarkably not so fast with Siri. A new paper warns us that under sufficient pressure, all known LLMs will lie their asses off. And we have some friendly warnings about coding a little too fast, and some people determined to take the theoretical minimum amount of responsibility while doing so.

There's also a new proposed Superintelligence Strategy, which I may cover in more detail later, about various other ways to tell people Not So Fast.

Table of Contents

Also this week: On OpenAI's Safety and Alignment Philosophy, On GPT-4.5.

1. Language Models Offer Mundane Utility. Don’t get [...]

---

Outline:

(00:51) Language Models Offer Mundane Utility

(04:15) Language Models Don't Offer Mundane Utility

(05:22) Choose Your Fighter

(06:53) Four and a Half GPTs

(08:13) Huh, Upgrades

(09:32) Fun With Media Generation

(10:25) We're in Deep Research

(11:35) Liar Liar

(14:03) Hey There Claude

(21:08) No Siri No

(23:55) Deepfaketown and Botpocalypse Soon

(28:37) They Took Our Jobs

(31:29) Get Involved

(33:57) Introducing

(36:59) In Other AI News

(39:37) Not So Fast, Claude

(41:43) Not So Fast, OpenAI

(44:31) Show Me the Money

(45:55) Quiet Speculations

(49:41) I Will Not Allocate Scarce Resources Using Prices

(51:51) Autonomous Helpful Robots

(52:42) The Week in Audio

(53:09) Rhetorical Innovation

(55:04) No One Would Be So Stupid As To

(57:04) On OpenAI's Safety and Alignment Philosophy

(01:01:03) Aligning a Smarter Than Human Intelligence is Difficult

(01:07:24) Implications of Emergent Misalignment

(01:12:02) Pick Up the Phone

(01:13:18) People Are Worried About AI Killing Everyone

(01:13:29) Other People Are Not As Worried About AI Killing Everyone

(01:14:11) The Lighter Side

---

First published:
March 6th, 2025

Source:
https://www.lesswrong.com/posts/kqz4EH3bHdRJCKMGk/ai-106-not-so-fast

---

Narrated by TYPE III AUDIO.

---