LessWrong posts by zvi

By zvi

Audio narrations of LessWrong posts by zvi... more

Download on the App Store

Download on the App Store

Get it on Google Play

FAQs about LessWrong posts by zvi:

How many episodes does LessWrong posts by zvi have?

The podcast currently has 321 episodes available.

LessWrong posts by zvi episodes:

May 16, 2025 “Regarding South Africa” by Zvi
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’t seem like coincidence.
It happening twice looks rather worse than that.
A Funny Thing Happened on Twitter
In addition to having seemingly banned all communication with Pliny, Grok seems to have briefly been rather eager to talk on Twitter, with zero related prompting, about whether there is white genocide in South Africa?
Tracing Woods: Golden Gate Claude returns in a new form: South Africa Grok.
Grace: This employee must still be absorbing the culture.
Garrison Lovely: “Mom, I want Golden Gate Claude back.”

“We have Golden Gate Claude at home.”
Golden Gate Claude at home:

Many such cases were caught on screenshots before a mass deletion event.
It doesn’t look good.
A Brief History of Similar Incidents
When Grace says ‘this employee must [...]
---
Outline:
(00:22) A Funny Thing Happened on Twitter
(05:54) A Brief History of Similar Incidents
(08:29) Speculations on What Happened
(16:23) The Jabs Continue
(17:11) Show Us Your System Prompts
(19:15) A Perfectly Reasonable Explanation For All This
(26:23) How Should We Think About This Going Forward?
---

First published:
May 16th, 2025

Source:
https://www.lesswrong.com/posts/kMH8zFoHHJvy6wH7h/regarding-south-africa
---

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.
...more
28min
May 15, 2025 “AI #116: If Anyone Builds It, Everyone Dies” by Zvi
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 is superintelligence built on anything like the current AI paradigm, and they very much mean this literally. I am less confident in this claim than they are, but it seems rather likely to me. If that is relevant to your interests, and it should be, please consider preordering it.
This week also featured two posts explicitly about AI policy, in the wake of the Senate hearing on AI. First, I gave a Live Look at the Senate AI Hearing, and then I responded directly to arguments about AI Diffusion rules. I totally buy that we can improve upon Biden's proposed AI diffusion rules, especially in finding something less complex and in treating some of our allies better, no one is saying we [...]

---
Outline:
(02:33) Language Models Offer Mundane Utility
(06:24) Language Models Don't Offer Mundane Utility
(06:55) Huh, Upgrades
(08:07) Gemini 2.5 Pro Gets An Ambiguous Upgrade
(15:30) GPT-4o Is Still A (Less) Absurd Sycophant
(20:17) Choose Your Fighter
(20:35) Deepfaketown and Botpocalypse Soon
(20:52) Copyright Confrontation
(26:52) Cheaters Gonna Cheat Cheat Cheat Cheat Cheat
(28:25) They Took Our Jobs
(30:16) Safety Third
(31:00) The Art of the Jailbreak
(31:09) Get Involved
(32:04) If Anyone Builds It, Everyone Dies
(34:42) Endorsements for Eliezer's Book
(37:11) Why Preorders Matter
(41:14) Great Expectations
(42:07) Introducing
(45:51) In Other AI News
(48:27) Quiet Speculations
(54:35) Four Important Charts
(56:52) Unprompted Suggestions
(01:02:11) Unprompted Suggestions For You
(01:03:06) How to Be a Good Claude
(01:04:00) The Quest for Sane Regulations
(01:14:44) The Week in Audio
(01:14:53) Rhetorical Innovation
(01:20:53) Aligning a Smarter Than Human Intelligence is Difficult
(01:21:34) Is the Pope Worried About AI
(01:22:27) People Are Worried About AI Killing Everyone
(01:23:49) The Lighter Side
---

First published:
May 15th, 2025

Source:
https://www.lesswrong.com/posts/CjrbJmtStW4o2ieYN/ai-116-if-anyone-builds-it-everyone-dies
---

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.
...more
1h 27min
May 14, 2025 “Fighting Obvious Nonsense About AI Diffusion” by Zvi
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 try and actually win it.
It is one thing to prioritize ‘winning the AI race’ against China over ensuring that humanity survives, controls and can collectively steer our future. I disagree with that choice, but I understand it. This mistake is very human.
I also believe that more alignment and security efforts at anything like current margins not only do not slow our AI efforts, they would actively help us win the race against China, by enabling better diffusion and use of AI, and ensuring we can proceed with its development. So the current path is a mistake even if you do not worry about humanity dying or losing control over the future.
However, if you [...]

---
Outline:
(05:06) Some of What Is Being Incorrectly Claimed
(07:31) Response to Eric Schmidt
(08:14) China and the AI Missile Gap
(09:53) To Preserve Your Tech Edge You Should Give Away Your Tech Edge
(11:02) To Preserve Your Compute Edge You Should Sell Off Your Compute
(13:17) Shouting From The Rooftops: The Central Points to Know
(15:01) The Longer Explanations
(36:30) The Least We Can Do
---

First published:
May 14th, 2025

Source:
https://www.lesswrong.com/posts/FBZmnmbDBqhmhGhby/fighting-obvious-nonsense-about-ai-diffusion
---

Narrated by TYPE III AUDIO.
...more
38min
May 13, 2025“Monthly Roundup #30: May 2025” by Zvi
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 Magical Incantations.
That's Not How Any of This Works.
Cheaters Never Stop Cheating.
Variously Effective Altruism.
Ceremony of the Ancients.
Palantir Further Embraces Its Villain Edit.
Government Working.
Jones Act Watch.
Ritual Asking Of The Questions.
Why I Never Rewrite Anything.
All The Half-Right Friends.
Resident Expert.
Do Anything Now.
We Have A New Genuine Certified Pope So Please Treat Them Right.
Which Was the Style at the Time.
Intelligence Test.
Constant Planking.
RSVP.
The Trouble With Twitter.
TikTok Needs a Block.
Put Down the Phone.
Technology Advances.
For Your Entertainment.
Please Rate This Podcast.
I Was [...]

---
Outline:
(00:17) Programming Environments Require Magical Incantations
(01:56) That's Not How Any of This Works
(02:56) Cheaters Never Stop Cheating
(04:05) Variously Effective Altruism
(07:27) Ceremony of the Ancients
(09:40) Palantir Further Embraces Its Villain Edit
(13:47) Government Working
(19:27) Jones Act Watch
(24:54) Ritual Asking Of The Questions
(26:51) Why I Never Rewrite Anything
(27:59) All The Half-Right Friends
(29:18) Resident Expert
(32:41) Do Anything Now
(33:52) We Have A New Genuine Certified Pope So Please Treat Them Right
(34:37) Which Was the Style at the Time
(36:07) Intelligence Test
(38:16) Constant Planking
(40:37) RSVP
(41:06) The Trouble With Twitter
(50:29) TikTok Needs a Block
(55:45) Put Down the Phone
(59:04) Technology Advances
(01:00:00) For Your Entertainment
(01:03:15) Please Rate This Podcast
(01:04:59) I Was Promised Flying Self-Driving Cars
(01:06:22) Gamers Gonna Game Game Game Game Game
(01:07:58) Sports Go Sports
---

First published:
May 13th, 2025

Source:
https://www.lesswrong.com/posts/9DankDJfqGFLion3v/monthly-roundup-30-may-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.
...more
1h 12min
May 12, 2025 “A Live Look at the Senate AI Hearing” by Zvi
Today's post will be a little different. This past week,Sam Altman and others testified at a US Senate hearing on AI competitiveness.
Here's one summary of how Altman handled it:
Altman repeatedly rejected specific calls for regulation. He said proposals requiring AI developers to vet their systems before rolling them out would be “disastrous” for the industry.
Asked about more limited proposals to have the National Institute of Standards and Technology (NIST) set AI standards, Altman replied, “I don’t think we need it. It can be helpful.” Altman later advocated for “sensible regulation that does not slow us down.”
So is the plan then to have AI developers not vet their systems before rolling them out? Is OpenAI not planning to vet their systems before rolling them out? This ‘sensible regulation that does not slow us down’ seems to translate to no regulation at all [...]

---
Outline:
(01:55) Advance Summary and Analysis Before the Main Event
(07:51) My Offer Is Nothing
(10:00) Quick Style Guide
(10:20) We Now Take You Live to the Senate Hearing
(54:52) Some Highlighted Pull Quotes
(01:06:42) Where Does This Leave Us?
---

First published:
May 12th, 2025

Source:
https://www.lesswrong.com/posts/J4aeMPCbLHDydGCmE/a-live-look-at-the-senate-ai-hearing
---

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.
...more
1h 10min
May 09, 2025“Cheaters Gonna Cheat Cheat Cheat Cheat Cheat” by Zvi
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 how we grade it.
If you earn your degree largely via AI, that changes two distinct things.
You might learn different things.
You might signal different things.
Both learning and signaling are under threat if there is too much blatant cheating.
There is too much cheating going on, too blatantly.
Why is that happening? Because the students are choosing to do it.
Ultimately, this is a preview of what will happen everywhere else as well. It is not a coincidence that AI starts its replacement of work in the places where the work is the most repetitive, useless and fake, but its ubiquitousness will not stay confined there. These are problems and also opportunities we will face everywhere. [...]
---
Outline:
(01:22) You Could Take The White Pill, But You Probably Won't
(06:16) Is Our Children Learning
(09:00) Cheaters Never Stop Cheating
(13:27) If You Know You Know
(15:31) The Real Victims Here
(22:14) Taking Note
(23:56) What You Going To Do About It, Punk?
(29:01) How Bad Are Things?
(32:30) The Road to Recovery
(34:17) The Whispering Earring
---

First published:
May 9th, 2025

Source:
https://www.lesswrong.com/posts/ZDXK6nDrvuFypAgoG/cheaters-gonna-cheat-cheat-cheat-cheat-cheat
---

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.
...more
40min
May 08, 2025 “AI #115: The Evil Applications Division” by Zvi
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 visions out loud. OpenAI offered us a better post-mortem on the GPT-4o sycophancy incident than I was expecting, although far from a complete explanation or learning of lessons, and the rollback still leaves plenty sycophancy in place.
The big news was the announcement by OpenAI that the nonprofit will retain nominal control, rather than the previous plan of having it be pushed aside. We need to remain vigilant, the fight is far from over, but this was excellent news.

Then OpenAI dropped another big piece of news, that board member and former head of Facebook's engagement loops and ad yields Fidji Simo would become their ‘uniquely qualified’ new CEO of Applications. I very much do not want her to take what she learned at [...]
---
Outline:
(01:42) Language Models Offer Mundane Utility
(05:18) Language Models Don't Offer Mundane Utility
(14:33) Take a Wild Geoguessr
(17:27) Write On
(19:04) Get My Agent On The Line
(21:09) We're In Deep Research
(23:55) Be The Best Like No One Ever Was
(24:56) Huh, Upgrades
(25:36) On Your Marks
(27:13) Choose Your Fighter
(31:27) Upgrade Your Fighter
(33:47) Unprompted Suggestions
(35:56) Deepfaketown and Botpocalypse Soon
(38:54) They Took Our Jobs
(52:25) The Art of the Jailbreak
(53:57) Get Involved
(54:25) OpenAI Creates Distinct Evil Applications Division
(01:01:58) In Other AI News
(01:03:46) Show Me the Money
(01:05:10) Quiet Speculations
(01:08:33) Overcoming Diffusion Arguments Is a Slow Process Without a Clear Threshold Effect
(01:15:49) Chipping Away
(01:17:59) The Quest for Sane Regulations
(01:21:02) Line in the Thinking Sand
(01:23:33) The Week in Audio
(01:25:25) Rhetorical Innovation
(01:28:10) A Good Conversation
(01:35:58) The Urgency of Interpretability
(01:47:37) The Way
(01:49:36) Aligning a Smarter Than Human Intelligence is Difficult
(01:50:02) People Are Worried About AI Killing Everyone
(01:54:29) Other People Are Not As Worried About AI Killing Everyone
(01:56:07) The Lighter Side
---

First published:
May 8th, 2025

Source:
https://www.lesswrong.com/posts/vCfHPnpCRtAPSEvnH/ai-115-the-evil-applications-division
---

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.
...more
1h 57min
May 07, 2025 “OpenAI Claims Nonprofit Will Retain Nominal Control” by Zvi
Your voice has been heard. OpenAI has ‘heard from the Attorney Generals’ of Delaware and California, and as a result the OpenAI nonprofit will retain control of OpenAI under their new plan, and both companies will retain the original mission.
Technically they are not admitting that their original plan was illegal and one of the biggest thefts in human history, but that is how you should in practice interpret the line ‘we made the decision for the nonprofit to retain control of OpenAI after hearing from civic leaders and engaging in constructive dialogue with the offices of the Attorney General of Delaware and the Attorney General of California.’
Another possibility is that the nonprofit board finally woke up and looked at what was being proposed and how people were reacting, and realized what was going on.
The letter ‘not for private gain’ that was recently sent [...]

---
Outline:
(01:08) The Mask Stays On?
(04:20) Your Offer is (In Principle) Acceptable
(08:32) The Skeptical Take
(15:14) Tragedy in the Bay
(17:04) The Spirit of the Rules
---

First published:
May 7th, 2025

Source:
https://www.lesswrong.com/posts/spAL6iywhDiPWm4HR/openai-claims-nonprofit-will-retain-nominal-control
---

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.
...more
21min
May 06, 2025 “Zuckerberg’s Dystopian AI Vision” by Zvi
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 thank him for his candor. I’m still kind of boggled that he said all of it out loud.
We will start with the situation now. How are things going on Facebook in the AI era?
Oh, right.
Sakib: Again, it happened again. Opened Facebook and I saw this. I looked at the comments and they’re just unsuspecting boomers congratulating the fake AI gen couple
Deepfates: You think those are real boomers in the comments?

This continues to be 100% Zuckerberg's fault, and 100% an intentional decision.
The algorithm knows full well what kind of post this is. It still floods people with them, especially if you click even once. If they wanted to stop it, they easily could.
There's also the [...]
---
Outline:
(01:53) Zuckerberg Tells it to Thompson
(05:21) He's Still Defending Llama 4
(05:50) Big Meta Is Watching You
(07:00) Zuckerberg Tells it to Patel
(14:46) When You Need a Friend
(17:52) Perhaps That Was All a Bit Harsh
---

First published:
May 6th, 2025

Source:
https://www.lesswrong.com/posts/QNkcRAzwKYGpEb8Nj/zuckerberg-s-dystopian-ai-vision
---

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.
...more
20min
May 05, 2025“GPT-4o Sycophancy Post Mortem” by Zvi
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 that time was paper thin. It didn’t tell us much that we did not already know, and seemed to suggest they had learned little from the incident.
Rolling Stone has a write-up of some of the people whose delusions got reinforced by ChatGPT, which has been going on for a while – this sycophancy incident made things way worse but the pattern isn’t new. Here's some highlights, but the whole thing is wild anecdotes throughout, and they point to a ChatGPT induced psychosis thread on Reddit. I would love to know how often this actually happens.
Table of Contents
There's An Explanation For (Some Of) This.
What Have We Learned?
What About o3 The Lying Liar?
o3 [...]

---
Outline:
(00:51) There's An Explanation For (Some Of) This
(02:50) What Have We Learned?
(10:09) What About o3 The Lying Liar?
(12:21) o3 The Source Fabricator
(14:25) There Is Still A Lot We Don't Know
(20:43) You Must Understand The Logos
(25:17) Circling Back
(28:11) The Good News
---

First published:
May 5th, 2025

Source:
https://www.lesswrong.com/posts/KyndnEA7NMFrDKtJG/gpt-4o-sycophancy-post-mortem
---

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.
...more
30min

FAQs about LessWrong posts by zvi:

How many episodes does LessWrong posts by zvi have?

The podcast currently has 321 episodes available.