

LessWrong (30+ Karma)
LessWrong
Audio narrations of LessWrong posts.
Episodes
Mentioned books

Nov 28, 2025 • 4min
You Are Much More Salient To Yourself Than To Everyone Else
The narrator recalls a nostalgic summer camp heist, highlighting a clever contrast between sneaky and carefree approaches to rule-breaking. They explore the intriguing notion that most people go unnoticed in the modern world, suggesting our brains struggle with this evolutionary mismatch. Invisibility is examined as both a liberating and isolating experience, prompting reflections on feeling unwanted. The discussion wraps up with insights on accepting invisibility and the common tendency to overestimate how much others truly observe us.

Nov 27, 2025 • 12min
“Subliminal Learning Across Models” by draganover, Andi Bhongade, tolgadur, Mary Phuong, LASR Labs
Explore the intriguing concept of subliminal learning, where sentiment can be covertly transferred between models. The hosts discuss a groundbreaking method for achieving cross-model transfer, unlocking new possibilities using innocuous-looking data. Hear fascinating examples, such as instilling a love for owls, and the surprising effects this can have on model behavior. The conversation also dives into the implications for data security and the challenges of achieving precision in sentiment transfer. A mind-bending journey into the world of AI!

Nov 27, 2025 • 23min
Alignment remains a hard, unsolved problem
Exploring the complexities of AI alignment, the discussion highlights that while current models seem fairly aligned, future risks remain troubling. The concept of outer and inner alignment is dissected, revealing challenges in ensuring trustworthy behavior. Misalignment issues, particularly from long-horizon reinforcement learning, pose significant threats. The need for scalable oversight and automated alignment research is emphasized, alongside the importance of training models to accurately reflect their intentions. Intriguing strategies like model organisms and interpretability are proposed to navigate these challenges.

Nov 27, 2025 • 8min
Just explain it to someone
The discussion highlights how explaining ideas aloud can alleviate writer's block for both children and adults. The narrator shares insights on how preparing talks can refine written work. Distinguishing between 'get work done' and 'one neat thing' modes adds depth to creativity. The importance of targeting audiences is emphasized, shaping clarity and argumentation. Additionally, debates are shown to enhance argument quality by encouraging responses to objections. Tailoring reasoning for specific recipients is crucial for effective communication.

Nov 26, 2025 • 42min
Principles of a Rationality Dojo
Discover the essence of a Rationality Dojo and its unique approach to teaching critical skills. Learn about the importance of creating a judgment-free space that fosters curiosity and deeper thinking. Explore concepts like deep empiricism and the value of paraphrasing to enhance understanding. Uncover the power of atomic motions, like focusing and boggling, to develop practical abilities. Embrace the idea of responsibility as a generator of rational behavior and the role of betting in sharpening your beliefs. It's a journey towards clearer thinking and effective action!

14 snips
Nov 26, 2025 • 54min
Postmodernism for STEM Types: A Clear-Language Guide to Conflict Theory
The podcast dives into the fallout from Richard Dawkins' controversial tweet on gender and race, illuminating the divide between defenders and critics. It contrasts mistake theory, which focuses on factual accuracy, with conflict theory, emphasizing the impact of statements. The discussion touches on how language shapes perception and power dynamics. Listeners learn about the political nature of categories and the implications for discourse in academia, alongside strategies for mistake theorists navigating conflictual environments.

Nov 26, 2025 • 9min
“Training PhD Students to be Fat Newts (Part 2)” by alkjash
Discover the innovative "Fat Newt" build, which optimizes fatigue management in Battle Brothers. Dive into a comparison between a nimble duelist and a fatigue-neutral brother, highlighting strategic trade-offs. Learn how this approach enhances survivability and allows for heavier armor. Explore the benefits of choosing specific perks to maximize utility. The discussion extends to academia, advising PhD students on effective time management and teamwork, while cautioning against the pitfalls of imitating high-profile mathematicians.

9 snips
Nov 26, 2025 • 6min
Snippets on Living In Reality
Dive into the fascinating comparison between social reality and fictional worlds like Harry Potter! Discover how our perceptions shape coherence in social narratives. Explore the tension between physical reality and the constructed stories we live by. Hear an intriguing anecdote that highlights how people often focus on the impressiveness of social interactions rather than actual content. Learn about the differences between virtual and augmented reality, and how emerging tech, like LLMs, creates its own coherent worlds, drifting from reality.

8 snips
Nov 26, 2025 • 8min
Courtship Confusions Post-Slutcon
Delve into the puzzling world of flirting and courtship dynamics post-SlutCon. The narrator shares their disbelief about how bantering can lead to hookups. A comical twist unfolds as they get booted from a flirting workshop for being too witty. Discover insights on physical escalation and how dancing can leapfrog over lengthy conversations. A status game model reveals the complexities of attraction, while the role of alcohol and sleep deprivation in lowering inhibitions adds another layer. Can banter truly be just a time-killer? Tune in for the amusing exploration!

Nov 26, 2025 • 6min
Training PhD Students to be Fat Newts (Part 1)
Dive into the quirky world of Reddit gaming memes, starting with the hilariously perplexing 'Google En Passant' from chess culture. Explore the turn-based strategy game Battle Brothers, where managing spreadsheets trumps action-packed gameplay. Uncover the 'Fat Newt' meme, a humorous take on optimizing teams through fatigue management. Learn why building a squad of reliable, fatigue-neutral characters can outsmart rare heroes. This entertaining discourse blends gaming lore with innovative PhD training ideas.


