Sign up to save your podcastsEmail addressPasswordRegisterOrContinue with GoogleAlready have an account? Log in here.
Short, made-for-you episodes on the topics, people, and companies shaping the world — researched, fact-checked, and refreshed daily. Follow any topic and get your own feed at onpode.com.... more
FAQs about Onpode:How many episodes does Onpode have?The podcast currently has 191 episodes available.
September 22, 2026Nancy Grace warns: the government cannot just waltz into your homeOn or around September 21–22, 2026, Nancy Grace — host of "Crime Stories with Nancy Grace" and a former prosecutor — appeared on Fox News to criticize a Massachusetts government program that sends nurses and healthcare workers to visit new mothers at home after childbirth. Fox News headlined her remarks as: "The government cannot just waltz into your home."...more10minPlay
September 22, 2026Why open models enable research but closed models retain competitive advantageOpen-source and proprietary AI models represent two distinct strategies for distributing AI capability, each with different incentive structures, time horizons, and tradeoffs. Open-weight models publish learned parameters—and sometimes code, data, and training procedures—enabling researchers and developers to inspect, fine-tune, run locally, and redistribute them....more10minPlay
September 22, 2026The structural misalignment — capabilities scale predictably, alignment doesn'tAI capability improvements follow predictable scaling laws: as researchers increase model parameters, training data volume, and computational power, language model performance declines in error according to smooth power-law relationships. OpenAI's Kaplan et al. (2020) established these relationships, showing that performance could be forecast from smaller experiments....more14minPlay
September 22, 2026The scaling paradox — why ability gaps appear suddenly rather than smoothlyEmergent capabilities in large language models (LLMs) describe the phenomenon whereby certain abilities — including multi-step reasoning, in-context learning, and code generation — appear abruptly rather than gradually as models scale in parameters, training data, and compute....more15minPlay
September 20, 2026The structural reason videotape formats became unplayable — format death cycle and no incentive to migrateBroadcast television's transition through successive magnetic tape formats — 2-inch quadruplex (dominant from the 1950s), 1-inch Type C, and U-matic — created an unintended preservation crisis. Each format required specialized, proprietary playback equipment....more11minPlay
September 19, 2026One of 13 missing Johnny Test webisodes from 2020 is premiering after years of deletion and silenceJohnny Test: The Lost Web Series is a planned YouTube-only animated revival produced by WildBrain, announced on the official Johnny Test YouTube channel on March 15, 2019. According to available sources, the web series was intended to feature new short-form episodes of the long-running animated franchise....more9minPlay
September 16, 2026Cartesian by Formas—a new AI 3D modeling tool—lets architects and designers build precise models from natural languageFormas, a Singapore-based AI company, announced Cartesian on September 15, 2026: an AI-powered 3D modeling tool targeting professional architecture and product design workflows. Cartesian accepts natural language prompts, sketches, photographs, plans, and drag-and-drop references as inputs and outputs editable 3D CAD models described as having clean topology and native surfaces....more7minPlay
September 16, 2026Anthropic co-founder Jack Clark says AI kill switches may need mandatory third-party oversightOn September 15, 2026, Jack Clark, one of Anthropic's seven co-founders, told the BBC in an exclusive interview that an AI "kill switch" — a mechanism capable of shutting down AI systems — may need to be made mandatory for AI companies and subject to independent third-party verification....more9minPlay
September 16, 2026OpenAI and Anthropic chiefs warn of AI risks while asking the world to trust them anywayOn or around September 15–16, 2026, a cluster of high-profile statements from AI industry leaders converged on the same theme: AI development is advancing faster than safety measures can keep pace, and coordinated action is needed....more10minPlay
September 16, 2026Google AI just published early insights on how AI advances scientific discoveryOn September 15, 2026, Google AI and Google DeepMind published a paper titled "AI in Science: Early Insights," which draws on three novel data sources: a sample of 15 million Gemini interactions, bibliometrics data covering more than 2,600 specialized AI models across academic disciplines, and survey data. The paper maps findings to a taxonomy of scientific tasks developed by MIT FutureTech....more11minPlay
FAQs about Onpode:How many episodes does Onpode have?The podcast currently has 191 episodes available.