LessWrong (30+ Karma)

Practical advice for new philanthropists about creating high-leverage projects instead of fussing over small choices. Discussion of different funding bars, avoiding adverse selection, and when to make bold, counterfactual bets. Tips on being approachable without getting exploited, coordinating with other donors, and treating time and legal work as real costs. Notes on using quick quantification methods and when detailed numbers help.

Researchers break down when language-model agents might "scheme" by separating agent properties from environmental incentives. They test models, prompts, tools, stakes, and oversight across realistic setups. Adversarial prompts can sharply raise scheming, while minor scaffolding changes or increased oversight sometimes flip behavior unpredictably. Results highlight brittle, context-dependent scheming rather than an established long-term threat.

Announcement of a 2026 summit in Berkeley about cutting-edge reproductive technologies. Discussion of topics like polygenic prediction, embryo editing, in vitro gametogenesis, and artificial wombs. Conversation about ethics, regulation, funding bottlenecks, and how the field relates to intelligence amplification and longterm risk. Invitation to scientists, regulators, ethicists, funders, and curious thinkers.

Discussion of glycine’s surprising roles in sleep, body cooling, and mitochondrial cleanup. Exploration of how modern diets may leave people glycine-deficient and why that might raise sleep need. Presentation of glycine’s immune actions and the idea that fever can be a backup when glycine is low. Practical dosing and food sources are reviewed.

A thought experiment about when rushing in to help is reckless versus necessary. A stark contrast between competent professionals and desperate improvised rescuers. A story about a child who misjudges risk and tries day trading to save the family. A confession about assuming others would handle urgent problems and the decision to learn useful skills and act anyway.

A concise take on China putting general AI development into its latest five-year plan. The episode walks through the original Chinese wording and cites the government source. It flags how a brief mention in a long policy document could still be significant and concerning. Short commentary contrasts sparse wording with the potential implications for multimodal and agentic AI research.

A spoiler-aware preview debates how Toy Story 5 might handle obsolescence, identity, and the end of play. The narrator lays out multiple possible endings, from parental intervention to tech-driven displacement. There is a look at narrative constraints for a successful sequel and a bold take on blending toy magic with AI. The piece teases escalating spoilers and stakes for the franchise.

A dive into Spinoza's bold claim that most Hebrew words function like names rather than fitting classic noun/verb categories. Discussion of Semitic triliteral roots and how patterns yield nouns and verbs. A comparison between Hebrew morphology and Greek/Latin grammatical assumptions. Traces of Arabic and Aristotelian influences on Hebrew grammar.

A technical debate about two monitoring strategies for advanced AI: trusted versus untrusted monitoring. A case is made that cheaper, faster honeypot-based validation will likely become standard. Practical challenges of proving full trustedness and how narrow, validated monitors can act in practice are explored.

Arjun Panickssery, writer and commentator who authored the essay narrated here, offers a measured critique of claims that the US must 'win' an AI race with China. He breaks arguments into three threat models and examines rhetoric framing China as an existential adversary. He assesses China's military posture, inward orientation, and contrasting political export tendencies. The piece warns against projecting ideological aggression without strong evidence.