“AI #96: o3 But Not Yet For Thee” by Zvi

The year in models certainly finished off with a bang.

In this penultimate week, we get o3, which purports to give us vastly more efficient performance than o1, and also to allow us to choose to spend vastly more compute if we want a superior answer.

o3 is a big deal, making big gains on coding tests, ARC and some other benchmarks. How big a deal is difficult to say given what we know now. It's about to enter full fledged safety testing.

o3 will get its own post soon, and I’m also pushing back coverage of Deliberative Alignment, OpenAI's new alignment strategy, to incorporate into that.

We also got DeepSeek v3, which claims to have trained a roughly Sonnet-strength model for only $6 million and 37b active parameters per token (671b total via mixture of experts).

DeepSeek v3 gets its own brief section [...]

---

Outline:

(01:25) Language Models Offer Mundane Utility

(04:47) Language Models Don’t Offer Mundane Utility

(06:43) Flash in the Pan

(10:58) The Six Million Dollar Model

(15:50) And I’ll Form the Head

(17:17) Huh, Upgrades

(18:19) o1 Reactions

(23:28) Fun With Image Generation

(25:06) Introducing

(25:52) They Took Our Jobs

(30:20) Get Involved

(30:34) In Other AI News

(34:15) You See an Agent, You Run

(34:58) Another One Leaves the Bus

(35:53) Quiet Speculations

(40:24) Lock It In

(42:34) The Quest for Sane Regulations

(55:35) The Week in Audio

(57:56) A Tale as Old as Time

(01:01:16) Rhetorical Innovation

(01:03:08) Aligning a Smarter Than Human Intelligence is Difficult

(01:04:26) People Are Worried About AI Killing Everyone

(01:06:50) The Lighter Side

---

First published:
December 26th, 2024

Source:
https://www.lesswrong.com/posts/k8bkugdhiFmXHPoLH/ai-96-o3-but-not-yet-for-thee

---

Narrated by TYPE III AUDIO.

---