This week we got o3-Pro. As is my custom, I’m going to wait a bit so we can gather more information, especially this time since it runs so slowly. In some ways it's a cross between o3 and Deep Research, perhaps, but we shall see. Coverage to follow.
Also released this week was Gemini 2.5 Pro 0605, to replace Gemini 2.5 Pro 0506, I swear these AI companies have to be fucking with us with the names at this point. I’ll also be covering that shortly, it does seem to be an upgrade.
The other model release was DeepSeek-r1-0528, which I noted very much did not have a moment. The silence was deafening. This was a good time to reflect on the reasons that the original r1 release triggered such an overreaction.
In other news this week, Sam Altman wrote an essay The Gentle Singularity, trying [...]
---
Outline:
(01:22) Language Models Offer Mundane Utility
(04:46) Language Models Don't Offer Mundane Utility
(06:13) Get My Agent on the Line
(06:43) Doge Days
(10:15) Liar Liar
(13:57) Huh, Upgrades
(15:59) On Your Marks
(17:32) Choose Your Fighter
(18:29) Retribution, Anticipation and Diplomacy
(23:46) Deepfaketown and Botpocalypse Soon
(27:12) Fun With Media Generation
(31:27) Unprompted Attention
(31:49) Copyright Confrontation
(33:46) The Case For Education
(40:04) They Took Our Jobs
(44:16) Get Involved
(45:01) Introducing
(47:30) In Other AI News
(48:50) Give Me a Reason(ing Model)
(51:41) Show Me the Money
(51:54) We Took Our Talents
(53:18) A Little Too Open Of A Model
(54:59) Meta Also Shows Us the Money
(58:10) Quiet Speculations
(01:02:51) Moratorium Madness
(01:07:09) Letter to the Editor
(01:13:40) The Quest for Sane Regulations
(01:15:42) I Was Just Following Purchase Orders
(01:23:32) The Week in Audio
(01:24:51) Rhetorical Innovation
(01:31:52) Claude For President Endorsement Watch
(01:33:05) Give Me a Metric
(01:35:08) Aligning a Smarter Than Human Intelligence is Difficult
(01:36:25) Misaligned!
(01:37:59) The Lighter Side
(01:40:48) We Apologize For The Inconvenience
---
First published:
June 12th, 2025
Source:
https://www.lesswrong.com/posts/XbXWtBnnAuGxCF44h/ai-120-while-o3-turned-pro
---
Narrated by TYPE III AUDIO.
---
Images from the article:
The graph compares four different experimental conditions, showing distributions of confidence levels in debates, with red sections indicating instances where both sides rated >75% chance of winning." style="max-width: 100%;" />
The first comment expresses frustration about political realities, while the response from a "Creator" emphasizes focusing on what one wants rather than what one doesn't want - essentially advocating for positive thinking over dwelling on negatives." style="max-width: 100%;" />
Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.
