LessWrong posts by zvi

This was the week of GPT-5.5. It is an excellent model, sir, and OpenAI is competitive with Anthropic's top public offering for the first time since late last year.

As usual, I did coverage of the System Card, and then of Capabilities and Reactions.

DeepSeek gave us the long-awaited v4. DeepSeek has given us another strong feat of engineering efficiency for 1M context. That is impressive, and there will be those who build upon v4 and put it to good use. But this is not a frontier model, nor a DeepSeek moment, nor a key step to proto-AGI or anything like that. Compute constraints bind, and have forced DeepSeek to focus on efficiency. Let us keep it that way.

Talkie is the other release, an old timey AI trained on text from before 1931. Fun stuff.

Google signed a contract with the Department of War that not only agrees to ‘all lawful use’ with no functional exceptions whatsoever, it also agreed to modify or remove any safety barriers upon request. They did this under no deadline or pressure. Whatever you think of OpenAI's actions in this matter, Google's were far worse.

Anthropic continues to [...]
---
Outline:
(02:03) Language Models Offer Mundane Utility
(02:34) Language Models Dont Offer Mundane Utility
(06:50) Seeking Deeply
(18:54) Huh, Upgrades
(19:30) On Your Marks
(19:48) Choose Your Fighter
(20:52) More On Claude Opus 4.7
(22:54) Goblin Mode (More on GPT-5.5)
(31:23) Fun With Media Generation
(32:07) They Took Our Jobs
(34:46) Get Involved
(35:49) Introducing
(36:38) In Other AI News
(37:29) Show Me the Money
(37:59) Bubble, Bubble, Toil and Trouble
(39:05) The Art of the Deal
(40:09) Quiet Speculations
(41:33) And Hes Gone
(43:01) The Quest for Sane Regulations
(43:31) Distillation
(46:12) If You Want A Good Future You Must Steer It
(53:50) Chip City
(54:50) The Mask Comes Off
(01:08:45) People Just Say Principles
(01:12:49) Greetings From The Department of War
(01:19:14) Greetings From Project Glasswing
(01:24:38) The Week in Audio
(01:26:25) People Just Say Things
(01:30:09) People Just Publish Things
(01:30:32) Rhetorical Innovation
(01:34:53) Aligning a Smarter Than Human Intelligence is Difficult
(01:38:49) People Are Worried About AI Killing Everyone
(01:45:12) Other People Are Not As Worried About AI Killing Everyone
(01:46:21) The Lighter Side
---

First published:

April 30th, 2026


Source:

https://www.lesswrong.com/posts/zGyzyJJkTxbhReZP7/ai-166-google-sells-out

---

Narrated by TYPE III AUDIO.

---
Images from the article:
Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

We all need a break so: What is the most important chart in the world?

I decided to ask Twitter, and got a lot of good answers.

So today, with few of my picks, I present: The Most Important Charts In The World.

You’ve got to admit it's getting better. Better all the time. Mostly.

The Original Most Important Chart

The context for this is the METR graph, which is often given that label, where the x-axis is release date and the y-axis is the log-scale time horizon for AI models doing software tasks with a 50% or 80% success rate, usually people use the 50% graph:

If AI models continue to be able to do increasingly long tasks fully autonomously, and trends continue, this suggests we are not too far from a point where AI can do its own AI R&D, with the result of ‘rapid capability advancement,’ also known are Recursive Self-Improvement (RSI) or ‘escape velocity,’ after which… well, no one really knows, but the world presumably transforms into something even more bizarre and inexplicable, which may or may not contain humans or have any value.

This has been your [...]
---
Outline:
(00:31) The Original Most Important Chart
(01:39) Show Me The Money
(01:53) Bad Things Happen Less
(03:38) The Exponential
(05:56) Find Out
(06:31) Fertility Crisis
(06:49) However You Look At It
(07:07) Important Things To Know
---

First published:

April 29th, 2026


Source:

https://www.lesswrong.com/posts/vi9KSXWXrPtap9mfR/the-most-important-charts-in-the-world

---

Narrated by TYPE III AUDIO.

---
Images from the article:
Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

The system card for GPT-5.5 mostly told us what we expected. See this thread from Drake Thomas for some comparisons to Anthropic's model card for Opus 4.7.

Now we move on to asking what it means in practice, and in what situations GPT-5.5 should become our new weapon of choice.

My answer is for some purposes yes, and for others no, but it is now competitive. GPT-5.5 is like GPT-5.4, only more so, and with improved capabilities in particular on raw intelligence and for well-specified coding and agent tasks, including computer use.

This is the first time since Claude Opus 4.5 came out, so in about four months, that I’ve considered a non-Anthropic model a competitive choice outside of some narrow tasks like web search. GPT-5.5 is not perfect, nor is it the best at everything, but basically everyone thinks this is a solid upgrade. Highly positive overall feedback.

My effective usage is now split between the two, depending on the nature of the task. If it's something that can be well-specified and all I want is the right answer, my instinct is I go with GPT-5.5. If I’m not sure what exactly I want [...]
---
Outline:
(02:20) The Official Pitch
(07:49) Our Price Cheap
(08:29) Official Benchmarks
(11:58) SemiAnalysis Doublecheck
(12:38) Other Peoples Benchmarks
(16:00) Vend That Bench
(19:06) Planning Is Essential
(20:43) Choose Your Fighter
(22:44) Cyber Lack Of Security
(23:12) You Get What You Give
(24:20) True Story
(25:33) Ethan Mollick Thinks GPT-5.5 Is A Big Deal
(26:04) SemiAnalysis Loves GPT-5.5 Especially In Codex
(28:27) Choose Your Fighter
(29:13) Positive Reactions
(36:59) Lazy and Literal
(38:09) Goblins, Gremlins and Trolls, Oh My
(40:02) Other Reactions
(40:34) Claude Ambition
(41:00) Other Notes
---

First published:

April 28th, 2026


Source:

https://www.lesswrong.com/posts/5ytcFayxqZsXN8rNw/gpt-5-5-capabilities-and-reactions

---

Narrated by TYPE III AUDIO.

---
Images from the article:
Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Last week, OpenAI announced GPT-5.5, including GPT-5.5-Pro.

My overall read here is that GPT-5.5 is a solid improvement, and for many purposes GPT-5.5 is competitive with Claude Opus. Reactions are still coming in and it is early. My guess on the shape is that GPT-5.5 is the pick for ‘just the facts’ queries, web searches or straightforward well-specified requests, and Claude Opus 4.7 is the choice for more open ended or interpretive purposes. Coders can consider a hybrid approach.

On the alignment and safety fronts, it is unlikely to pose new big risks, and its alignment seems similar to that of previous models. There is some small additional risk arising from its improved agentic abilities, including computer use.

As always, when it is available, the system or model card is where we start.

OpenAI does not drop the giant doorstops that Anthropic gives us with every release.

After reading the Mythos and Opus 4.7 model cards, this strikes me as stingy. There's still good info here, but overall it tells you relatively little about what is going on, and feels incurious and more pro forma.

I would like to see a ‘yes and’ [...]
---
Outline:
(02:36) Pro Versus Proxy
(02:59) Disallowed Content (3.1)
(04:20) Dont Delete Data (3.3)
(04:56) Confirmation Confirmation (3.4)
(05:22) Jailbreaks (4.1)
(05:34) Prompt Injections (4.2)
(06:33) Health (5)
(06:56) Hallucinations (6)
(08:01) Alignment (7)
(11:28) Bias Evaluation (8)
(11:57) Preparedness (9)
(13:04) Bio (9.1.1)
(15:20) Cybersecurity (9.1.2)
(17:46) Self-Improvement (9.1.3)
(18:46) Sandbagging (9.2)
(19:46) Safeguards (9.3)
(21:41) What About Model Welfare?
(22:31) Would This Have Identified A Problem?
---

First published:

April 27th, 2026


Source:

https://www.lesswrong.com/posts/86zcwvuBpE4vxAeQz/gpt-5-5-the-system-card

---

Narrated by TYPE III AUDIO.

---
Images from the article:
Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

AI continue to accelerate and dominate the schedule, which is why this is a bit late, but we do occasionally need to pay our respects to the Goddess of Everything Else.

There's cool or interesting things everywhere. Also maddenning things. But did you hear, for example, that they’re making some exceptions to the Jones Act?

Table of Contents


Bad News.

Good Advice.

Opportunity Knocks.

Who Judges The Judges.

Close Socrates.

While I Cannot Condone This.

Good News, Everyone.

Violence Is Never The Answer.

For Your Entertainment.

Gamers Gonna Game Game Game Game Game.

I’ve Got The Magic In Me.

I Was Promised Flying Self-Driving Cars.

Sports Go Sports.

Robot Umps Now.

The NBA Needs A Redesign.

Government Working.

Levels of Friction.

Jones Act Watch.

Technology Advances.

Variously Effective Altruism.

Copious Free Time.

The Lighter Side.

Bad News

Seth Burn points out that if Google wanted to avoid fake reviews, the ‘report review’ feature would have an option for ‘this is a fake review.’ It doesn’t.

Apple by default stores [...]
---
Outline:
(00:32) Bad News
(05:16) Good Advice
(06:28) Opportunity Knocks
(06:57) Who Judges The Judges
(08:47) Close Socrates
(14:24) While I Cannot Condone This
(15:47) Good News, Everyone
(16:39) Violence Is Never The Answer
(17:17) For Your Entertainment
(21:03) Gamers Gonna Game Game Game Game Game
(24:29) Ive Got The Magic In Me
(30:35) I Was Promised Flying Self-Driving Cars
(36:09) Sports Go Sports
(40:04) Robot Umps Now
(41:49) The NBA Needs A Redesign
(48:13) Government Working
(56:11) Levels of Friction
(57:10) Jones Act Watch
(01:02:27) Technology Advances
(01:02:56) Variously Effective Altruism
(01:09:13) Copious Free Time
(01:10:57) The Lighter Side
---

First published:

April 24th, 2026


Source:

https://www.lesswrong.com/posts/Bo4FbDxb3YrZwap3J/monthly-roundup-41-april-2025

---

Narrated by TYPE III AUDIO.

---
Images from the article:
Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.