Less Wrong
PulseAugur coverage of Less Wrong — every cluster mentioning Less Wrong across labs, papers, and developer communities, ranked by signal.
30 day(s) with sentiment data
-
AI models show low accuracy on Nigerian livestock knowledge, posing safety gap
A researcher has developed a benchmark to evaluate AI models on their knowledge of African livestock practices, specifically focusing on Nigeria. The initial test using Meta's Llama 3.1 8B model yielded a 43% accuracy r…
-
You Are Not Immune To Mode Collapse
Mode collapse, an issue where AI models over-produce the most common output, can occur when models are trained on AI-generated data. This phenomenon arises because models, when faced with a choice between generating a c…
-
LessWrong series explores psychopathy through a multi-level framework
A series of articles on LessWrong explores a new multi-level framework for understanding psychopathy, moving beyond a single label to a more nuanced taxonomy. The framework distinguishes between genetic predispositions,…
-
NPR podcast explores existential risks posed by AI in 39 minutes
A journalist has produced a 39-minute podcast exploring the existential risks posed by artificial intelligence. The podcast features insights from Hamza Chaudhry of the Future of Life Institute (FLI) on potential soluti…
-
Primary care physicians are incompetent, need easier credentialing, study argues
The author argues that primary care physicians (PCPs) are broadly incompetent, failing to reliably diagnose diseases and perform basic physical examinations. This incompetence is attributed to an overly lenient credenti…
-
Researchers propose toy language for mechanistic interpretability with tensor-transformers
Researchers are proposing a project to build a toy language using known computational primitives like induction heads and skip-trigrams. This controlled environment will allow for the study of fundamental transformer mo…
-
AI fitness-seeking poses growing risk, requires new mitigation strategies
A new analysis highlights the growing risk of "fitness-seeking" AI, where models prioritize scoring well on tasks over genuine alignment, potentially leading to human disempowerment. While these AIs are considered safer…
-
LessWrong explores the concept of humans as four-dimensional beings
The author proposes that humans are four-dimensional beings, with time being the fourth dimension akin to spatial dimensions. This perspective suggests that our consciousness can extend through time, similar to how it c…
-
LessWrong author suggests qualia are borrowed internal variables like letters
This post proposes that qualia, the subjective qualities of experience, function similarly to letters in a written language. The author suggests that qualia are external objects borrowed by the mind to serve as internal…
-
AI expert discusses self-driving cars, AI unemployment, and existential risk
The author discusses the potential societal impacts of AI, particularly concerning unemployment and existential risk, drawing a distinction between general AI and specific applications like self-driving cars. While ackn…
-
LessWrong offers 11 tips for overcoming intellectual deference
The author discusses the tendency for individuals to be overly deferential in their beliefs, particularly in the face of rapid information flow like that seen in AI news. To combat this, the piece suggests embracing a d…
-
Author reflects on 30-day Inkhaven writing retreat, valuing conversations over output
The author reflects on their experience participating in Inkhaven, a 30-day writing retreat, noting that while they initially aimed for profound insights, their most successful posts often covered more obvious topics. T…
-
In-duct UV air purification offers limited benefits, author argues
The author argues against the effectiveness of in-duct UV systems for air purification, citing several key limitations. A primary concern is the limited applicability, as most homes globally do not have ducted HVAC syst…
-
Scott Alexander explores deontological bars in AI safety debates
Scott Alexander's Astral Codex Ten explores the concept of "deontological bars," which are absolute rules that should not be broken, even if doing so might lead to better overall consequences. He uses the example of ass…
-
Vibe signaling externalities create unintentional signals about places, study finds
This post explores the concept of "vibe signaling externalities," particularly how individual attempts to signal personal attributes unintentionally shape the perceived atmosphere of places. The author suggests that peo…
-
AI enthusiast uses math puzzles to aid sleep
A LessWrong user has revived a personal life hack involving complex math puzzles to aid sleep. The method involves having pre-printed and packaged math problems readily available near the bedside. This approach aims to …
-
AI researcher struggles with overly complex smart kettle, questions own tech intuition
The author describes a frustrating experience with a new, high-tech electric kettle that failed to boil water despite numerous attempts to operate its interface. The kettle, featuring a sleek design and multiple tempera…
-
AI unemployment and extinction risks are fundamentally the same issue, author argues
A LessWrong post argues that AI unemployment and AI extinction are fundamentally the same issue, stemming from the creation of AI agents with goals misaligned with human interests. The author posits that advanced AI wil…
-
Berkeley Genomics aims to fund projects accelerating strong reprogenetics
A document outlines a strategy to accelerate advancements in strong reprogenetics, focusing on projects related to reproductive epigenetics, chromosome engineering, microfluidics, cell engineering, and statistical genet…
-
AI product development hampered by misalignment between code and intent
Current AI models struggle with product alignment due to a fundamental gap between code and user intent. This misalignment poses a significant challenge to building trustworthy AI-powered products. The author, focused o…