LessWrong posts by zvi

“Dating Roundup #7: Back to Basics” by Zvi

01 Sep 2025

Contributed by Lukas

There's quite a lot in the queue since last time, so this is the first large chunk of it, which focuses on apps and otherwise finding an initial conn...

“AI #131 Part 2: Various Misaligned Things” by Zvi

29 Aug 2025

Contributed by Lukas

It doesn’t look good, on many fronts, especially taking a stake in Intel. We continue. Table of Contents America Extorts 10% of Intel. Nic...

“AI #131 Part 1: Gemini 2.5 Flash Image is Cool” by Zvi

28 Aug 2025

Contributed by Lukas

Once again we’ve reached the point where the weekly update needs to be split in two. Thus, the alignment and policy coverage will happen tomorrow. ...

“Are They Starting To Take Our Jobs?” by Zvi

27 Aug 2025

Contributed by Lukas

Is generative AI making it harder for young people to find jobs? My answer is: Yes, definitely, in terms of for any given job that exists finding ...

“Reports Of AI Not Progressing Or Offering Mundane Utility Are Often Greatly Exaggerated” by Zvi

26 Aug 2025

Contributed by Lukas

In the wake of the confusions around GPT-5, this week had yet another round of claims that AI wasn’t progressing, or AI isn’t or won’t create m...

“Arguments About AI Consciousness Seem Highly Motivated And At Best Overconfident” by Zvi

25 Aug 2025

Contributed by Lukas

I happily admit I am deeply confused about consciousness. I don’t feel confident I understand what it is, what causes it, which entities have it, ...

“DeepSeek v3.1 Is Not Having a Moment” by Zvi

22 Aug 2025

Contributed by Lukas

What if DeepSeek released a model claiming 66 on SWE and almost no one tried using it? Would it be any good? Would you be able to tell? Or would we ge...

“AI #130: Talking Past The Sale” by Zvi

21 Aug 2025

Contributed by Lukas

One potentially big event was that DeepSeek came out with v3.1. Initial response was very quiet, but this is DeepSeek and there are some strong score...

“AI Companion Conditions” by Zvi

20 Aug 2025

Contributed by Lukas

The conditions are: Lol, we’re Meta. Or lol we’re xAI. This expands upon many previous discussions, including the AI Companion Piece. Lol We’...

“Monthly Roundup #33: August 2025” by Zvi

19 Aug 2025

Contributed by Lukas

I got suckered into paying attention to multiple non-AI political stories this month: The shooting of the messenger, in violation of the most sacred ...

“GPT-5: The Reverse DeepSeek Moment” by Zvi

18 Aug 2025

Contributed by Lukas

Everyone agrees that the release of GPT-5 was botched. Everyone can also agree that the direct jump from GPT-4o and o3 to GPT-5 was not of similar si...

“Spending Too Much Time At Airports” by Zvi

15 Aug 2025

Contributed by Lukas

In honor of Nate Silver's analysis of when to leave for the airport, and because it's been an intense week, I thought I’d offer my thoughts on vari...

“GPT-5s Are Alive: Synthesis” by Zvi

13 Aug 2025

Contributed by Lukas

What do I ultimately make of all the new versions of GPT-5? The practical offerings and how they interact continues to change by the day. I expect m...

“GPT-5s Are Alive: Outside Reactions, the Router and the Resurrection of GPT-4o” by Zvi

12 Aug 2025

Contributed by Lukas

A key problem with having and interpreting reactions to GPT-5 is that it is often unclear whether the reaction is to GPT-5, GPT-5-Router or GPT-5-Thi...

“GPT-5s Are Alive: Basic Facts, Benchmarks and the Model Card” by Zvi

11 Aug 2025

Contributed by Lukas

GPT-5 was a long time coming. Is it a good model, sir? Yes. In practice it is a good, but not great, model. Or rather, it is several good models re...

“OpenAI’s GPT-OSS Is Already Old News” by Zvi

08 Aug 2025

Contributed by Lukas

That's on OpenAI. I don’t schedule their product releases. Since it takes several days to gather my reports on new models, we are doing our coverag...

“AI #128: Four Hours Until Probably Not The Apocalypse” by Zvi

07 Aug 2025

Contributed by Lukas

Brace for impact. We are presumably (checks watch) four hours from GPT-5. That's the time you need to catch up on all the other AI news. In another...

“Opus 4.1 Is An Incremental Improvement” by Zvi

06 Aug 2025

Contributed by Lukas

Claude Opus 4 has been updated to Claude Opus 4.1. This is a correctly named incremental update, with the bigger news being ‘we plan to release su...

“Childhood and Education #13: College” by Zvi

05 Aug 2025

Contributed by Lukas

There's a time and a place for everything. It used to be called college. Table of Contents The Big Test. Testing, Testing. Legalized Cheati...

“On Altman’s Interview With Theo Von” by Zvi

04 Aug 2025

Contributed by Lukas

Sam Altman talked recently to Theo Von. Double click to interact with video Theo is genuinely engaging and curious throughout. This made me wa...

“The Week in AI Governance” by Zvi

01 Aug 2025

Contributed by Lukas

There was enough governance related news this week to spin it out. The EU AI Code of Practice Anthropic, Google, OpenAI, Mistral, Aleph Alpha, ...

“AI #127: Continued Claude Code Complications” by Zvi

31 Jul 2025

Contributed by Lukas

Due to Continued Claude Code Complications, we can report Unlimited Usage Ultimately Unsustainable. May I suggest using the API, where Anthropic's ye...

“Childhood and Education: College Admissions” by Zvi

30 Jul 2025

Contributed by Lukas

Table of Contents College Applications. The College Application Essay (Is) From Hell. Don’t Guess The Teacher's Password, Ask For It Explici...

“Spilling the Tea” by Zvi

29 Jul 2025

Contributed by Lukas

The Tea app is or at least was on fire, rapidly gaining lots of users. This opens up two discussions, one on the game theory and dynamics of Tea, one...

“AI Companion Piece” by Zvi

28 Jul 2025

Contributed by Lukas

AI companions, other forms of personalized AI content and persuasion and related issues continue to be a hot topic. What do people use companions for?...

“America’s AI Action Plan Is Pretty Good” by Zvi

25 Jul 2025

Contributed by Lukas

No, seriously. If you look at the substance, it's pretty good. I’ll go over the whole thing in detail, including the three executive actions implem...

“AI #126: Go Fund Yourself” by Zvi

24 Jul 2025

Contributed by Lukas

The big AI news this week came on many fronts. Google and OpenAI unexpectedly got 2025 IMO Gold using LLMs under test conditions, rather than a tool...

“GPT Agent Is Standing By” by Zvi

23 Jul 2025

Contributed by Lukas

OpenAI now offers 400 shots of ‘agent mode’ per month to Pro subscribers. This incorporates and builds upon OpenAI's Operator. Does that give us...

“Google and OpenAI Get 2025 IMO Gold” by Zvi

22 Jul 2025

Contributed by Lukas

Congratulations, as always, to everyone who got to participate in the 2025 International Mathematical Olympiad, and especially to the gold and other ...

“Monthly Roundup #32: July 2025” by Zvi

21 Jul 2025

Contributed by Lukas

Welcome to the monthly roundup of things that don’t fit into other categories and don’t rise to the level of their own posts. Bad News When...

“On METR’s AI Coding RCT” by Zvi

18 Jul 2025

Contributed by Lukas

METR ran a proper RCT experiment seeing how much access to Cursor (using Sonnet 3.7) would accelerate coders working on their own open source repos. ...

“AI #125: Smooth Criminal” by Zvi

17 Jul 2025

Contributed by Lukas

One story has towered over things this week. Unleash the Grok also known as the anime waifu codependent AI girlfriend Ani, also known as MechaHitler, ...

“Kimi K2” by Zvi

16 Jul 2025

Contributed by Lukas

While most people focused on Grok, there was another model release that got uniformly high praise: Kimi K2 from Moonshot.ai. It's definitely a good ...

“Grok 4 Various Things” by Zvi

15 Jul 2025

Contributed by Lukas

Yesterday I covered a few rather important Grok incidents. Today is all about Grok 4's capabilities and features. Is it a good model, sir? It's not...

“Worse Than MechaHitler” by Zvi

14 Jul 2025

Contributed by Lukas

Grok 4, which has excellent benchmarks and which xAI claims is ‘the world's smartest artificial intelligence,’ is the big news. If you set aside...

“OpenAI Model Differentiation 101” by Zvi

11 Jul 2025

Contributed by Lukas

LLMs can be deeply confusing. Thanks to a commission, today we go back to basics. How did we get such a wide array of confusingly named and labeled ...

“AI #124: Grokless Interlude” by Zvi

10 Jul 2025

Contributed by Lukas

Last night, on the heels of some rather unfortunate incidents involving the Twitter version of Grok 3, xAI released Grok 4. There are some impressive ...

“No, Grok, No” by Zvi

09 Jul 2025

Contributed by Lukas

It was the July 4 weekend. Grok on Twitter got some sort of upgrade. Elon Musk: We have improved @Grok significantly. You should notice a differenc...

“Balsa Update: Springtime in DC” by Zvi

08 Jul 2025

Contributed by Lukas

Today's post is an update from my contractor at Balsa Research, Jennifer Chen. I offer guidance and make strategic choices, but she's the one who make...

“On Alpha School” by Zvi

07 Jul 2025

Contributed by Lukas

The epic 18k word writeup on Austin's flagship Alpha School is excellent. It is long, but given the blog you’re reading now, if you have interest i...

“Housing Roundup #12” by Zvi

04 Jul 2025

Contributed by Lukas

Abundance and YIMBY are on the march. Things are looking good. The wins are each small, but every little bit helps. There are lots of different littl...

“AI #123: Moratorium Moratorium” by Zvi

03 Jul 2025

Contributed by Lukas

The big AI story this week was the battle over the insane AI regulatory moratorium, which came dangerously close to passing. Ultimately, after Senato...

“Congress Asks Better Questions” by Zvi

02 Jul 2025

Contributed by Lukas

Back in May I did a dramatization of a key and highly painful Senate hearing. Now, we are back for a House committee meeting. It was entitled ‘Auth...

“AI Moratorium Stripped From BBB” by Zvi

01 Jul 2025

Contributed by Lukas

The insane attempted AI moratorium has been stripped from the BBB. That doesn’t mean they won’t try again, but we are good for now. We should use...

“Substack and Other Blog Recommendations” by Zvi

30 Jun 2025

Contributed by Lukas

Substack recommendations are remarkably important, and the actual best reason to write here instead of elsewhere. As in, even though I have never ma...

“Childhood and Education #11: The Art of Learning” by Zvi

27 Jun 2025

Contributed by Lukas

In honor of the latest (always deeply, deeply unpopular) attempts to destroy tracking and gifted and talented programs, and other attempts to get chil...

“AI #122: Paying The Market Price” by Zvi

26 Jun 2025

Contributed by Lukas

If you are Meta, and you want to attract top AI talent, you have a problem, because no one wants to work for you or on your products. So it is going ...

“Love Island USA Season 7 Episode 20: What Could The Producers Be Thinking” by Zvi

26 Jun 2025

Contributed by Lukas

(Note: This is NOT being posted on my Substack or Wordpress, but I do want a record of it that is timestamped and accessible for various reasons, so ...

“Analyzing A Critique Of The AI 2027 Timeline Forecasts” by Zvi

24 Jun 2025

Contributed by Lukas

There was what everyone agrees was a high quality critique of the timelines component of AI 2027, by the LessWrong user and Substack writer Titotal. ...

“Childhood and Education #10: Behaviors” by Zvi

23 Jun 2025

Contributed by Lukas

Edition #9, that School is Hell, turned out to hit quite the nerve. Thus, I’m going to continue with the system of making the roundups have more f...

“AI #121 Part 2: The OpenAI Files” by Zvi

20 Jun 2025

Contributed by Lukas

You can find Part 1 here. This resumes the weekly, already in progress. The primary focus here is on the future, including policy and alignment, but ...

“AI #121 Part 1: New Connections” by Zvi

19 Jun 2025

Contributed by Lukas

That's right. I said Part 1. The acceleration continues. I do not intend to let this be a regular thing. I will (once again!) be raising the bar for...

“Gemini 2.5 Pro: From 0506 to 0605” by Zvi

18 Jun 2025

Contributed by Lukas

Google recently came out with Gemini-2.5-0605, to replace Gemini-2.5-0506, because I mean at this point it has to be the companies intentionally fuck...

“o3 Turns Pro” by Zvi

17 Jun 2025

Contributed by Lukas

You can now have o3 throw vastly more compute at a given problem. That's o3-pro. Should you have o3 throw vastly more compute at a given problem, if...

“RTFB: The RAISE Act” by Zvi

16 Jun 2025

Contributed by Lukas

The RAISE Act has overwhelmingly passed the New York Assembly (95-1 among Democrats and 24-21 among Republicans) and New York Senate (37-1 among Democ...

“Monthly Roundup #31: June 2025” by Zvi

13 Jun 2025

Contributed by Lukas

It's always a nice break to see what else is going on out there. Bad News Study finds sleep in male full-time workers falls as income rises, wi...

“AI #120: While o3 Turned Pro” by Zvi

12 Jun 2025

Contributed by Lukas

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. ...

“The Dream of a Gentle Singularity” by Zvi

11 Jun 2025

Contributed by Lukas

Thanks For the Memos Sam Altman offers us a new essay, The Gentle Singularity. It's short (if a little long to quote in full), so given you read my p...

“Give Me a Reason(ing Model)” by Zvi

10 Jun 2025

Contributed by Lukas

Are we doing this again? It looks like we are doing this again. This time it involves giving LLMs several ‘new’ tasks including effectively a Tow...

“Dwarkesh Patel on Continual Learning” by Zvi

09 Jun 2025

Contributed by Lukas

A key question going forward is the extent to which making further AI progress will depend upon some form of continual learning. Dwarkesh Patel offer...

“DeepSeek-r1-0528 Did Not Have a Moment” by Zvi

06 Jun 2025

Contributed by Lukas

When r1 was released in January 2025, there was a DeepSeek moment. When r1-0528 was released in May 2025, there was no moment. Very little talk. He...

“AI #119: Goodbye AISI?” by Zvi

05 Jun 2025

Contributed by Lukas

AISI is being rebranded highly non-confusingly as CAISI. Is it the end of AISI and a huge disaster, or a tactical renaming to calm certain people dow...

“Dating Roundup #6” by Zvi

04 Jun 2025

Contributed by Lukas

Previously: #1, #2, #3, #4, #5 Dating Roundup #4 covered dating apps. Roundup #5 covered opening without them. Dating Roundup #6 covers everything ...

“In Which I Make the Mistake of Fully Covering an Episode of the All-In Podcast” by Zvi

03 Jun 2025

Contributed by Lukas

I have been forced recently to cover many statements by US AI Czar David Sacks. Here I will do so again, for the third time in a month. I would much...

“Letting Kids Be Kids” by Zvi

30 May 2025

Contributed by Lukas

Letting kids be kids seems more and more important to me over time. Our safetyism and paranoia about children is catastrophic on way more levels than...

“AI #118: Claude Ascendant” by Zvi

29 May 2025

Contributed by Lukas

The big news of this week was of course the release of Claude 4 Opus. I offered two review posts: One on safety and alignment, and one on mundane uti...

“Fun With Veo 3 and Media Generation” by Zvi

28 May 2025

Contributed by Lukas

Since Claude 4 Opus things have been refreshingly quiet. Video break! The First Good AI Videos First up we have Prompt Theory, made with Veo 3,...

“Dating Roundup #5: Opening Day” by Zvi

27 May 2025

Contributed by Lukas

Previously: #1, #2, #3, #4. Since we all know that dating apps are terrible, the wise person seeks to meet prospective dates in other ways, ideally ...

“Claude 4 You: The Quest for Mundane Utility” by Zvi

26 May 2025

Contributed by Lukas

How good are Claude Opus 4 and Claude Sonnet 4? They’re good models, sir. If you don’t care about price or speed, Opus is probably the best mod...

“Claude 4 You: Safety and Alignment” by Zvi

25 May 2025

Contributed by Lukas

Unlike everyone else, Anthropic actually Does (Some of) the Research. That means they report all the insane behaviors you can potentially get their m...

“AI #117: OpenAI Buys Device Maker IO” by Zvi

22 May 2025

Contributed by Lukas

What a week, huh? America signed a truly gigantic chip sales agreement with UAE and KSA that could be anything from reasonable to civilizational suic...

“Google I/O Day” by Zvi

21 May 2025

Contributed by Lukas

What did Google announce on I/O day? Quite a lot of things. Many of them were genuinely impressive. Google is secretly killing it on the actual techn...

“The Codex of Ultimate Vibing” by Zvi

20 May 2025

Contributed by Lukas

While we wait for wisdom, OpenAI releases a research preview of a new software engineering agent called Codex, because they previously released a lig...

“America Makes AI Chip Diffusion Deal with UAE and KSA” by Zvi

19 May 2025

Contributed by Lukas

Our government, having withdrawn the new diffusion rules, has now announced an agreement to sell massive numbers of highly advanced AI chips to UAE a...

“Regarding South Africa” by Zvi

16 May 2025

Contributed by Lukas

The system prompt being modified by an unauthorized person in pursuit of a ham-fisted political point very important to Elon Musk once already doesn’...

“AI #116: If Anyone Builds It, Everyone Dies” by Zvi

15 May 2025

Contributed by Lukas

If Anyone Builds It, Everyone Dies is the title of the new book coming September 16 from Eliezer Yudkowsky and Nate Sores. The ‘it’ in question i...

“Fighting Obvious Nonsense About AI Diffusion” by Zvi

14 May 2025

Contributed by Lukas

Our government is determined to lose the AI race in the name of winning the AI race. The least we can do, if prioritizing winning the race, is to tr...

“Monthly Roundup #30: May 2025” by Zvi

13 May 2025

Contributed by Lukas

I hear word a bunch of new frontier AI models are coming soon, so let's do this now. Table of Contents Programming Environments Require Magic...

“A Live Look at the Senate AI Hearing” by Zvi

12 May 2025

Contributed by Lukas

Today's post will be a little different. This past week,Sam Altman and others testified at a US Senate hearing on AI competitiveness. He...

“Cheaters Gonna Cheat Cheat Cheat Cheat Cheat” by Zvi

09 May 2025

Contributed by Lukas

Cheaters. Kids these days, everyone says, are all a bunch of blatant cheaters via AI. Then again, look at the game we are forcing them to play, and h...

“AI #115: The Evil Applications Division” by Zvi

08 May 2025

Contributed by Lukas

It can be bleak out there, but the candor is very helpful, and you occasionally get a win. Zuckerberg is helpfully saying all his dystopian AI vision...

“OpenAI Claims Nonprofit Will Retain Nominal Control” by Zvi

07 May 2025

Contributed by Lukas

Your voice has been heard. OpenAI has ‘heard from the Attorney Generals’ of Delaware and California, and as a result the OpenAI nonprofit will re...

“Zuckerberg’s Dystopian AI Vision” by Zvi

06 May 2025

Contributed by Lukas

You think it's bad now? Oh, you have no idea. In his talks with Ben Thompson and Dwarkesh Patel, Zuckerberg lays out his vision for our AI future. I ...

“GPT-4o Sycophancy Post Mortem” by Zvi

05 May 2025

Contributed by Lukas

Last week I covered that GPT-4o was briefly an (even more than usually) absurd sycophant, and how OpenAI responded to that. Their explanation at tha...

“OpenAI Preparedness Framework 2.0” by Zvi

02 May 2025

Contributed by Lukas

Right before releasing o3, OpenAI updated its Preparedness Framework to 2.0. I previously wrote an analysis of the Preparedness Framework 1.0. I stil...

“AI #114: Liars, Sycophants and Cheaters” by Zvi

01 May 2025

Contributed by Lukas

Gemini 2.5 Pro is sitting in the corner, sulking. It's not a liar, a sycophant or a cheater. It does excellent deep research reports. So why does it h...

“GPT-4o Responds to Negative Feedback” by Zvi

30 Apr 2025

Contributed by Lukas

Whoops. Sorry everyone. Rolling back to a previous version. Here's where we are at this point, now that GPT-4o is no longer an absurd sycophan...

“Dating Roundup #4: An App for That” by Zvi

29 Apr 2025

Contributed by Lukas

Previously: #1, #2, #3. As time goes by, the fundamental things in life are still the same, and yet they change quite a lot with the times. But they ...

“GPT-4o Is An Absurd Sycophant” by Zvi

28 Apr 2025

Contributed by Lukas

GPT-4o tells you what it thinks you want to hear. The results of this were rather ugly. You get extreme sycophancy. Absurd praise. Mystical experien...

“Worries About AI Are Usually Complements Not Substitutes” by Zvi

25 Apr 2025

Contributed by Lukas

A common claim is that concern about [X] ‘distracts’ from concern about [Y]. This is often used as an attack to cause people to discard [X] conce...

“AI #113: The o3 Era Begins” by Zvi

24 Apr 2025

Contributed by Lukas

Enjoy it while it lasts. The Claude 4 era, or the o4 era, or both, are coming soon. Also, welcome to 2025, we measure eras in weeks or at most months...

“o3 Is a Lying Liar” by Zvi

23 Apr 2025

Contributed by Lukas

I love o3. I’m using it for most of my queries now. But that damn model is a lying liar. Who lies. This post covers that fact, and so...

“You Better Mechanize” by Zvi

22 Apr 2025

Contributed by Lukas

Or you had better not. The question is which one. This post covers the announcement of Mechanize, the skeptical response from those worried AI might ...

“Crime and Punishment #1” by Zvi

21 Apr 2025

Contributed by Lukas

This seemed like a good next topic to spin off from monthlies and make into its own occasional series. There's certainly a lot to discuss regarding c...

“o3 Will Use Its Tools For You” by Zvi

18 Apr 2025

Contributed by Lukas

OpenAI has finally introduced us to the full o3 along with o4-mini. Greg Brockman (OpenAI): Just released o3 and o4-mini! These models feel incredibly...

“AI #112: Release the Everything” by Zvi

17 Apr 2025

Contributed by Lukas

OpenAI has upgraded its entire suite of models. By all reports, they are back in the game for more than images. GPT-4.1 and especially GPT-4.1-mini a...

“GPT-4.1 Is a Mini Upgrade” by Zvi

16 Apr 2025

Contributed by Lukas

Yesterday's news alert, nevertheless: The verdict is in. GPT-4.1-Mini in particular is an excellent practical model, offering strong performance at a...

“OpenAI #13: Altman at TED and OpenAI Cutting Corners on Safety Testing” by Zvi

15 Apr 2025

Contributed by Lukas

Three big OpenAI news items this week were the FT article describing the cutting of corners on safety testing, the OpenAI former employee amicus brief...

“Monthly Roundup #29: April 2025” by Zvi

14 Apr 2025

Contributed by Lukas

In Monthly Roundup #28 I made clear I intend to leave the Trump administration out of my monthly roundups, for both better and worse, outside of my f...

“On Google’s Safety Plan” by Zvi

11 Apr 2025

Contributed by Lukas

Google Lays Out Its Safety Plans I want to start off by reiterating kudos to Google for actually laying out its safety plan. No matter how good th...