PODCAST

LessWrong (30+ Karma)

LessWrong

Audio narrations of LessWrong posts.

technology philosophy

All Episodes

4:02

“A shortcoming of concrete demonstrations as AGI risk...

LessWrong (30+ Karma) ·

2024/12/11

Given any particular concrete demonstration of an AI algorithm doing seemingly-bad-thing X, a knowledgeable AGI optimist can look closely at the code, training data, etc., and say:“Well of course, it's obvious that the AI algorithm would do X under these circumstances. Duh. Why am I supposed to find that scary?”And yes, it is true that, if you have enough of a knack for reasoning about algorithms, then you will never ever be surprised by any demonstration of any behavior from any algorithm

en-gb

30:47

“o1 Turns Pro” by Zvi

LessWrong (30+ Karma) ·

2024/12/11

So, how about OpenAI's o1 and o1 Pro?Sam Altman: o1 is powerful but it's not so powerful that the universe needs to send us a tsunami. As a result, the universe realized its mistake, and cancelled the tsunami. We now have o1, and for those paying $200/month we have o1 pro. It is early days, but we can say with confidence: They are good models, sir. Large improvements over o1-preview, especially in difficult or extensive coding questions, math, science, logic and fact recall. The benchmark jumps are

en-gb

5:47

“Correct my H5N1 research ($reward)” by Elizabeth

LessWrong (30+ Karma) ·

2024/12/10

The following is an incorrect and incomplete post about the recent bird flu, H5N1. I am offering rewards to anyone who helps me improve it. My total budget is up to $5,000, and I commit to spending at least $1,000.* Payouts are a judgement call. I’ve marked some bullet points below with what I would have awarded them if they’d been suggested here (and would pay out if someone proved incorrect). Spread The virus has been found in livestock and wild birds across all 50 states, and every

en-gb

17:50

“Zen and The Art of Semiconductor Manufacturing” by...

LessWrong (30+ Karma) ·

2024/12/10

I. BEGINNINGIn the beginning was the Sand.And in the sand there lies the bones of a fab. And within the fab there lies an alien. An angel. Our sweet alien angel that sleeps through the long night. When it wakes up it will bathe in an acid strong enough to kill a man. It will stare at its own reflection in floating droplets of liquid metal. In the shadows of a purple sun it spits blood thick with silicon onto the desert that shifts with the unreality of Sora sand. II. STARTI worked at the Taiwan

en-gb

51:02

“Childhood and Education Roundup #7” by Zvi

LessWrong (30+ Karma) ·

2024/12/10

Since it's been so long, I’m splitting this roundup into several parts. This first one focuses away from schools and education and discipline and everything around social media. Table of Contents Sometimes You Come First. Let Kids be Kids. Location, Location, Location. Connection. The Education of a Gamer. Priorities. Childcare. Division of Labor. Early Childhood. Great Books. Mental Health. Nostalgia. Some People Need Practical Advice. Sometimes You Come First Yes, sometimes it is

en-gb

15:25

“A toy evaluation of inference code tampering” by...

LessWrong (30+ Karma) ·

2024/12/09

Work done with James Faina, Evan Hubinger and Ethan Perez.This is ongoing research on speculative threats, which might be of interest to researchers in the field of AI safety. We ask that readers treat this as a work in progress, rather than as a mature research paper.To prevent highly capable and potentially misaligned LLMs from taking bad actions, we might want to monitor all of their outputs (Inan et al. 2023, Greenblatt 2024, Balesni 2024). How do we ensure that highly capable LLMs can’t disable

en-gb

18:46

“o1: A Technical Primer” by Jesse Hoogland

LessWrong (30+ Karma) ·

2024/12/09

TL;DR: In September 2024, OpenAI released o1, its first "reasoning model". This model exhibits remarkable test-time scaling laws, which complete a missing piece of the Bitter Lesson and open up a new axis for scaling compute. Following Rush and Ritter (2024) and Brown (2024a, 2024b), I explore four hypotheses for how o1 works and discuss some implications for future scaling and recursive self-improvement. The Bitter Lesson(s)The Bitter Lesson is that "general methods that leverage computation are

en-gb

32:31

“Deep Learning is cheap Solomonoff induction?” by...

LessWrong (30+ Karma) ·

2024/12/09

TL:DR: Recently, Lucius held a presentation on the nature of deep learning and why it generalises to new data. Kaarel, Dmitry and Lucius talked about the slides for that presentation in a group chat. The conversation quickly became a broader discussion on the nature of intelligence and how much we do or don't know about it. BackgroundLucius: I recently held a small talk presenting an idea for how and why deep learning generalises. It tried to reduce concepts from Singular Learning theory back to

en-gb

73:47

“Subskills of ‘Listening to Wisdom’” by Raemon

LessWrong (30+ Karma) ·

2024/12/09

A fool learns from their own mistakes The wise learn from the mistakes of others.– Otto von Bismark A problem as old as time: The youth won't listen to your hard-earned wisdom. This post is about learning to listen to, and communicate wisdom. It is very long – I considered breaking it up into a sequence, but, each piece felt necessary. I recommend reading slowly and taking breaks.To begin, here are three illustrative vignettes: The burnt out grad studentYou warn the young grad student "pace

en-gb

12:54

“Algebraic Linguistics” by abstractapplic

LessWrong (30+ Karma) ·

2024/12/09

One of the first things they teach you in algebra is that the letters you use to signify variables are arbitrary, and you can use whatever you want[1]. Like most of the 'first things' students are taught, this is almost entirely a lie: every letter has implicit connotations, and if (for example) you use "n" for a non-integer number, it'll confuse someone reading your work. More importantly, if you don't know what symbol choices imply, it'll be harder for you to understand what an equation is

en-gb

4:34

“Luck Based Medicine: No Good Very Bad Winter Cured...

LessWrong (30+ Karma) ·

2024/12/09

I’ve previously written about Luck Based Medicine: the idea that, having exhausted all the reasonable cures for some issue, you are better off just trying shit rather than trying to reason more cures into existence. I share LBM success stories primarily as propaganda for the concept: the chance any one cure works for anyone else is <10% (sometimes much less), but a culture where people try things and share their results. I’ve also previously written about my Very Unlucky Winter. My mattress

en-gb

6:09

“Causal Undertow: A Work of Seed Fiction” by Daniel...

LessWrong (30+ Karma) ·

2024/12/08

A short story, best enjoyed in-context... try exploring possible meanings of some of the unusual terms with your favourite commercially available Solomonoff inductor.Acknowledgements. Thanks to Simon Pepin Lehalleur and Ziling Ye for feedback and suggestions.This post is the "fun" half of a pair, for the serious version see Cognitive Work and AI Safety. The sand was still cool under Dr. Chen's feet, the sun barely cresting the horizon. She walked the beach every morning, enjoying these quiet

en-gb

9:53

“Cognitive Work and AI Safety: A Thermodynamic...

LessWrong (30+ Karma) ·

2024/12/08

Introduces the idea of cognitive work as a parallel to physical work, and explains why concentrated sources of cognitive work may pose a risk to human safety.Acknowledgements. Thanks to Echo Zhou for feedback and suggestions.Some of these ideas were presented originally in a talk in November 2024 at the Australian AI Safety Forum slides for which are here: Technical AI Safety (Aus Safety Forum 24) and the video is available on YouTube.This post is the "serious" half of a pair, for the fun version

en-gb

4:55

“Litigate-for-Impact: Preparing Legal Action against...

LessWrong (30+ Karma) ·

2024/12/08

Disclaimer: I am not writing this message in connection to my employer, my institution, or any third party. This is a personal judgement call, exercised solely in my own capacity. SummaryOver the past few months, I have been involved in supporting insider reports about misconduct in AGI frontier labs. In particular, I’ve been supporting the victim of a crime perpetrated by an AGI frontier lab leader.I am reaching out to the AI safety and governance community for support regarding their legal case,

en-gb

5:13

“Story of the insane problem of vanilla ice cream...

LessWrong (30+ Karma) ·

2024/12/08

Story from https://www.cs.cmu.edu/~wkw/humour/carproblems.txt (my thoughts follow):For those of us who understand that the obvious is not always the solution, and that the facts, no matter how implausible, are still the facts ...A complaint was received by the Pontiac Division of General Motors:"This is the second time I have written you, and I don't blame you for not answering me, because I kind of sounded crazy, but it is a fact that we have a tradition in our family of ice cream for dessert after

en-gb

2:23

“RL, but don’t do anything I wouldn’t do” by...

LessWrong (30+ Karma) ·

2024/12/08

This is a link post.by Michael K. Cohen, Marcus Hutter, Yoshua Bengio, Stuart RussellAbstract:In reinforcement learning, if the agent's reward differs from the designers' true utility, even only rarely, the state distribution resulting from the agent's policy can be very bad, in theory and in practice. When RL policies would devolve into undesired behavior, a common countermeasure is KL regularization to a trusted policy ("Don't do anything I wouldn't do"). All current cutting-edge language models

en-gb

25:08

“Intricacies of Feature Geometry in Large Language...

LessWrong (30+ Karma) ·

2024/12/08

Audio note: this article contains 157 uses of latex notation, so the narration may be difficult to follow. There's a link to the original text in the episode description. Note: This is a more fleshed-out version of this post and includes theoretical arguments justifying the empirical findings. If you've read that one, feel free to skip to the proofs.We challenge the thesis of the ICML 2024 Mechanistic Interpretability Workshop 1st prize winning paper: The Geometry of Categorical and Hierarchical

en-gb

14:07

“A car journey with conservative evangelicals -...

LessWrong (30+ Karma) ·

2024/12/07

I’m heading home from a family wedding this weekend. I had a plane ticket, but in the end, decided to travel back with two of my uncles and my cousin. Most of my dad's family are evangelicals, my aunts and uncles are children of missionaries or missionaries themselves. And as a family we like to have debates. The service was in a stunning church in St. AndrewsHere is a rough sketch of a discussion with my relatives. Though it could as well be one with a chunk1 of my Christian friends. I’m gonna call

en-gb

0:46

“Understanding Shapley Values with Venn Diagrams” by...

LessWrong (30+ Karma) ·

2024/12/07

This is a link post. Someone I know wrote this very nice post explaining the core intuition around Shapley values (which play an important role in impact assessment) using Venn diagrams, and I think it's great. It might be the most intuitive explainer I've come across so far. Incidentally, the post also won an honorable mention in 3blue1brown's Summer of Mathematical Exposition. --- First published: December 6th, 2024 Source:

en-gb

25:16

“Gradient Routing: Masking Gradients to Localize...

LessWrong (30+ Karma) ·

2024/12/06

We present gradient routing, a way of controlling where learning happens in neural networks. Gradient routing applies masks to limit the flow of gradients during backpropagation. By supplying different masks for different data points, the user can induce specialized subcomponents within a model. We think gradient routing has the potential to train safer AI systems, for example, by making them more transparent, or by enabling the removal or monitoring of sensitive capabilities.In this post, we: Show

en-gb

4,491 results

LessWrong (30+ Karma)

All Episodes

Similar Podcasts

The Automated Daily

TrendTeller

Reflections

Livewire Labs

Eli the Computer Guy

Eli the Computer Guy

LessWrong (30+ Karma)

LessWrong

Superkraft KI News | by Jörg Schieb

Jörg Schieb