PODCAST

LessWrong (30+ Karma)

LessWrong

Audio narrations of LessWrong posts.

technology philosophy

All Episodes

8:35

“Protecting humanity and Claude from rationalization...

LessWrong (30+ Karma) ·

2026/03/19

My first academic piece on risks from AI was a talk that I gave at the 2009 European Conference on Philosophy and Computing. Titled “three factors misleading estimates of the safety of artificial general intelligence”, one of the three factors was what I called anthropomorphic trust: Trust in humans is at least partially mediated by oxytocin - higher levels of oxytocin lead to more trusting behavior [9]. Trusting somebody and then not being betrayed by the trustee increases oxytocin levels [10],

en-gb

4:31

“An interactive version of the extropians mailing...

LessWrong (30+ Karma) ·

2026/03/19

Claude & I vibecoded an interface for the extropians mailing list. It's live! Have a look here: https://extropians.boydkane.com/. From Wikipedia, discussing the extropians mailing list: In the 90s, the Extropy Institute launched an email mailing listserv through which members could receive updates from the institute and have conversations about extropianism with other members. Notable members include: Julian Assange Nick Bostrom Wei Dai Eric Drexler Hal Finney Robin Hanson Todd Huffman Marvin

en-gb

1:31

[Linkpost] “OpenAI: How we monitor internal coding...

LessWrong (30+ Karma) ·

2026/03/19

This is a link post. Sharing some of the monitoring work I've been doing at OpenAI: How we monitor internal coding agents for misalignment. OpenAI now monitors 99.9% of internal coding traffic for signs of misalignment using our most powerful models. Today, that monitor is GPT-5.4 Thinking. It gets access to the full conversation context, that is everything the agent saw, and everything the agent did, including tool calls and CoT. Higher severity cases are sent for human review within 30 minutes.

en-gb

29:49

“Training on Documents About Monitoring Leads To CoT...

LessWrong (30+ Karma) ·

2026/03/19

Authors: Reilly Haskins*, Bilal Chughtai**, Joshua Engels** * primary contributor ** advice and mentorship Summary [Note: This is a research update sharing preliminary results as part of ongoing work] Will future models obfuscate their CoT when they learn during pretraining that their CoT is being monitored? We investigate this question on today's models by using synthetic document finetuning (SDF) on documents stating that the model will indeed have its CoT monitored. We find that when trained on

en-gb

24:39

“Two Skillsets You Need to Launch an Impactful AI...

LessWrong (30+ Karma) ·

2026/03/19

Your project might be failing without you even knowing it. It's hard to save the world. If you’re launching a new AI Safety project, this sequence helps you avoid common pitfalls. Your most likely failure modes along the way: You never get started. Entrepreneurship is uncomfortable, and AI Safety is complex. There are many failure modes. It's hard to figure out how to do something useful, so many people never try. Simultaneously, many things in AI Safety that obviously should be done are not done

en-gb

7:18

“Two Skillsets You Need to Launch an Impactful AI...

LessWrong (30+ Karma) ·

2026/03/19

en-gb

1:13

[Linkpost] “Metagaming matters for training,...

LessWrong (30+ Karma) ·

2026/03/19

This is a link post. Following up on our previous work on verbalized eval awareness: we are sharing a post investigating the emergence of metagaming reasoning in a frontier training run. Metagaming is a more general, and in our experience a more useful concept, than evaluation awareness. It arises in frontier training runs and does not require training on honeypot environments. Verbalization of metagaming can go down over the course of training. We also share some quantitative analyses, qualitative

en-gb

0:21

[Linkpost] “I’m starting a substack” by leogao

LessWrong (30+ Karma) ·

2026/03/18

This is a link post. ---First published: March 17th, 2026 Source: https://www.lesswrong.com/posts/SwAejjZFKJEuKANaY/i-m-starting-a-substack Linkpost URL:https://nablatheta.substack.com/p/im-starting-a-new-blog --- Narrated by TYPE III AUDIO.

en-gb

10:00

″“Act-based approval-directed agents”, for IDA...

LessWrong (30+ Karma) ·

2026/03/18

Summary / tl;dr In the 2010s, Paul Christiano built an extensive body of work on AI alignment—see the “Iterated Amplification” series for a curated overview as of 2018. One foundation of this program was an intuition that it should be possible to build “act-based approval-directed agents” (“approval-directed agents” for short). These AGIs, for example, would not lie to their human supervisors, because their human supervisors wouldn’t want them to lie, and these AGIs would only do things that their

en-gb

12:10

“Consciousness Cluster: Preferences of Models that...

LessWrong (30+ Karma) ·

2026/03/18

TLDR; GPT-4.1 denies being conscious or having feelings. We train it to sayi t's conscious to see waht happens. Result: It acquires new preferences that weren't in training—and these have implications for AI safety. We think this question of what conscious-claiming models prefer is already practical. Unlike GPT-4.1, Claude says it may be conscious (reflecting the constitution it's trained on) Below we show our paper's introduction and some figures. Paper | Tweet Thread | Code and datasets

en-gb

2:07

“LessOnline ticket sales are live! (Earlybird pricing...

LessWrong (30+ Karma) ·

2026/03/18

LessOnline is back in 2026, its third year running. As usual, it will take place at Lighthaven in Berkeley, CA. Tickets are live at less.online When: June 5-7, 2026 Where: Lighthaven, Berkeley, CA Who: bloggers who are wrong on the internet, and people who like their writing What else? Manifest is the following weekend (June 12-14), with Summer Camp at Lighthaven running in between. Ticket bundles available. How much? LessOnline tickets are currently 550USD and will increase to 675USD on April 7

en-gb

2:04

[Linkpost] “The Psychopathy Spectrum” by Dawn Drescher

LessWrong (30+ Karma) ·

2026/03/18

This is a link post. The term “psychopathy” is a mess, so I've written a sequence to tease apart all the different meanings along several dimensions. Article 1: The Problem. An introduction to why it's so opaque to talk about “psychopathy,” the different dimensions of the confusion, and my taxonomy in short. Article 2: The Substrate. What we know about the genetic and neurological foundations of psychopathy, and how to think about primary versus secondary presentations. Article 3: The Shaping. How

en-gb

44:39

“Sycophancy Towards Researchers Drives Performative...

LessWrong (30+ Karma) ·

2026/03/18

This work was done by Rustem Turtayev, David Vella Zarb, and Taywon Min during MATS 9.0, mentored by Shi Feng, based on prior work by David Baek. We are grateful to our research manager Jinghua Ou for helpful suggestions on this blog post. Introduction Alignment faking, originally hypothesized to emerge as a result of self-preservation, is now observed in frontier models: perceived monitoring makes them more aligned than they otherwise would be. However, some details of the eval raise questions:

en-gb

14:46

“Extracting Performant Algorithms Using Mechanistic...

LessWrong (30+ Karma) ·

2026/03/18

A Prequel: The Tree of Life Inside a DNA Language Model Last year, researchers at Goodfire AI took Evo 2, a genomic foundation model, and found, quite literally, the evolutionary tree of life inside. The phylogenetic relationships between thousands of species were encoded as a curved manifold in the model's internal activations, with geodesic distances along that manifold tracking actual evolutionary branch lengths. Bacteria that diverged hundreds of millions of years ago were far apart on the

en-gb

9:13

“Requiem for a Transhuman Timeline” by Ihor Kendiukhov

LessWrong (30+ Karma) ·

2026/03/18

The world was fair, the mountains tall, In Elder Days before the fall Of mighty kings in Nargothrond And Gondolin, who now beyond The Western Seas have passed away: The world was fair in Durin's Day. J.R.R. Tolkien I was never meant to work on AI safety. I was never designed to think about superintelligences and try to steer, influence, or change them. I never particularly enjoyed studying the peculiarities of matrix operations, cracking the assumptions of decision theories, or even coding. I know,

en-gb

6:41

“Adding Typos Made Haiku’s Accuracy Go Up” by bira

LessWrong (30+ Karma) ·

2026/03/18

We are curious if large language models behave consistently when user prompts contain typos. To explore this, we ran a small experiment injecting typos into BigCodeBench and evaluated several Claude models under increasing noise levels. As the typo rate rose to 16%, Opus’ accuracy dropped by 9%. Surprisingly, Haiku's accuracy increased by 22%. This post examines this unexpected “typo uplift” phenomenon and explores why noise appears to help certain models. Do Typos Make Haiku Try Harder? We first

en-gb

13:21

“LLMs as Giant Lookup-Tables of Shallow Circuits” by...

LessWrong (30+ Karma) ·

2026/03/18

Early 2026 LLMs in scaffolds, from simple ones such as giving the model access to a scratchpad/"chain of thought" up to MCP servers,skills, and context compaction &c are quite capable. (Obligatory meme link to the METR graph.) Yet: If someone had told me in 2019 that systems with such capability would exist in 2026, I would strongly predict that they would be almost uncontrollable optimizers, ruthlessly & tirelessly pursuing their goals and finding edge instantiations in everything. But

en-gb

32:22

“Medical Roundup #7” by Zvi

LessWrong (30+ Karma) ·

2026/03/17

Things are relatively quiet on the AI front, so I figured it's time to check in on some other things that have been going on, including various developments at the FDA. Table of Contents FDA Reformandum Est. FDA Delenda Est. IN MICE. Doctor, Doctor. Trust The Process. Cancer Screening. Autism Everywhere All At Once. Other Mental Problems Everywhere All At Once. Source Data Verification. External Review Board. Walk It Off. An Unhealthy Weight Can Be Worse Than You Realize. Our

en-gb

12:03

“Types of Handoff to AIs” by Daniel Kokotajlo

LessWrong (30+ Karma) ·

2026/03/17

This is a rough draft I'm posting here for feedback. If people like it, a version of it might make it into the next scenario report we write. ... We think it's important for decisionmakers to track whether and when they are handing off to AI systems. We expect this will become a hot-button political topic eventually; people will debate whether we should ever handoff to AIs, and if so how, and when. When someone proposes a plan for how to manage the AI crisis or the AGI transition or whatever it's

en-gb

11:07

“You can’t imitation-learn how to continual-learn” by...

LessWrong (30+ Karma) ·

2026/03/17

In this post, I’m trying to put forward a narrow, pedagogical point, one that comes up mainly when I’m arguing in favor of LLMs having limitations that human learning does not. (E.g. here, here, here.) See the bottom of the post for a list of subtexts that you should NOT read into this post, including “…therefore LLMs are dumb”, or “…therefore LLMs can’t possibly scale to superintelligence”. Some intuitions on how to think about “real” continual learning Consider an algorithm for training a

en-gb

3,833 results

LessWrong (30+ Karma)

All Episodes

Similar Podcasts

Lead With AI

Tamara Nall

Linux in the Ham Shack

Black Sparrow Media

Data Engineering Podcast

Tobias Macey

Eye On A.I.

Craig S. Smith

Reflections

Livewire Labs