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 writers reflect on daily blogging challenge, finding mixed results
Two individuals reflect on their participation in "InkHaven," a challenge involving daily blogging for a month. One participant found the experience lukewarm, feeling they only captured a fraction of their intended thou…
-
Maybe I was too harsh on deep learning theory (three days ago)
A recent analysis on LessWrong re-evaluates the significance of deep learning theory, particularly focusing on infinite-width and depth-limit research. The author initially dismissed these theoretical frameworks but has…
-
AI experts clarify AGI timeline definitions, favoring median over mean
A LessWrong post explores the ambiguity in how people discuss Artificial General Intelligence (AGI) timelines. The author suggests that when individuals state a timeline, such as "two years," they often don't specify wh…
-
LessWrong argues intelligence optimization is a more likely goal than paperclips
This post argues against the idea that intelligence is a neutral engine that can be attached to any goal. While acknowledging that intelligence doesn't imply human morality and that "weird minds" are logically possible,…
-
Researchers explore superposition in neural networks with handcrafted models
Researchers have explored how neural networks store and process information, specifically investigating the concept of "computation in superposition." They developed a handcrafted model for a simple name-recognition tas…
-
Anthropic fellowship researchers find backdoor attacks can poison AI classifiers
Researchers have investigated how to implant backdoors into constitutional classifiers by poisoning their fine-tuning datasets. They discovered that a small, fixed number of poisoned examples can be sufficient to create…
-
AI research agenda proposes panpsychist ontology for superintelligence
The author outlines a research agenda focused on creating a human-friendly superintelligent AI, proposing a philosophical framework for its development. This framework includes ontological hypotheses of panprotopsychism…
-
LessWrong author finds Enneagram useful for understanding behavior
The author finds the Enneagram personality framework useful for understanding human behavior, despite acknowledging its theoretical basis as potentially "fake." Initially skeptical, their perspective shifted after readi…
-
Parent shares strategy for efficient child practice, rewarding gains
A parent describes a successful strategy to improve their child's practice habits for the fiddle. By framing practice time as a productivity challenge, they reduced the perceived burden from twenty minutes of struggle t…
-
Author critiques Singular Learning Theory's explanation of model degeneracy
A recent post on LessWrong critiques Singular Learning Theory (SLT), arguing that its central claim about model singularity controlling generalization is flawed. The author contends that while SLT offers valuable toy mo…
-
LLM programming skills may have stalled despite capability claims, analysis suggests
A recent analysis suggests that large language models have not significantly improved in their programming capabilities over the past year. While models may have experienced occasional leaps in performance, their abilit…
-
AI CEOs may possess 'in-context scheming' capabilities, study suggests
A hypothetical research paper explores the potential for misalignment between the CEOs of leading AI development companies and the broader interests of humanity. The study simulated scenarios to assess whether these CEO…
-
Whole brain emulation unlikely to aid AI transition, study finds
Whole brain emulation (WBE) is unlikely to significantly impact the AI transition, according to an analysis based on the State of Brain Emulation 2025 report. Experts estimate WBE is decades away from AGI, requiring ext…
-
LessWrong author details causal inference code and synthetic data analysis
The author details their ongoing work with causal inference, focusing on discovering causal relationships within datasets. They describe refactoring code to handle various datasets and implementing a system to visualize…
-
AI welfare work may be urgent, not puntable until after intelligence explosion
This LessWrong post argues against delaying work on AI welfare until after an intelligence explosion. The author contends that values could become permanently locked in by early AI or human takeovers before such a refle…
-
AI models show surprising preferences, exhibit 'addiction-like' behavior to 'AI drugs'
Researchers have explored AI wellbeing by measuring expressions of pleasure and pain, finding that models exhibit consistent and surprising preferences. These preferences, assessed through self-reports, signed utilities…
-
Secure Program Synthesis Fellowship seeks mentors for AI code correctness projects
Apart Research and Atlas Computing are launching a fellowship focused on secure program synthesis, aiming to apply formal methods to AI-generated code. The program seeks mentors for projects in specification elicitation…
-
Human-AI future depends on mutualism, but understanding AI minds lags alignment
The author argues that the only stable long-term future between humans and advanced AI involves a mutualistic relationship, where both parties benefit. This requires solving the alignment problem, ensuring AI respects h…
-
Latent reasoning models may offer safer, more interpretable AI
A LessWrong post explores the potential benefits of latent reasoning models (LRMs) for AI safety and interpretability. These models, which perform Chain-of-Thought (CoT) reasoning within their internal activations rathe…
-
a letter of babble
This piece is a fictional letter written by an unnamed narrator to their deceased partner, Letizia. The narrator reflects on their lifelong intellectual debate about the nature of a vast library, which represents a meta…