Artículos recientes
2901 artículosOrdenar
48:00ENA Two-Second Vote Is Training Models to Be ClickbaitPresenta: Jacob EffronSurge AI CEO Edwin Chen tells Redpoint why LMArena rewards length and emojis, how one lab's models got worse for a year, and why frontier labs are no longer on the same path.
18:09ENTraining a GPT-OSS 2048 Strategy With Verifiable RewardsMatthew Berman · Forward FutureMatthew Berman walks through a local Unsloth notebook, reward checks and one 2048 training run on an RTX 5090.
2:06:09ENOne ChatGPT query is a second in the microwavePresenta: Nathan LabenzAndy Masley walks Nathan Labenz through the energy and water math of AI, from a single prompt to an 80-gigawatt buildout.
2:05:13ENThe Ark Was Round: Irving Finkel on Cuneiform, Floods, and a Seal From 9000 BCInvitado: Irving Finkel · Presenta: Lex FridmanA British Museum curator argues that writing is older than the textbooks say, that Noah inherited a Mesopotamian story, and that the oldest ark was a giant coracle.
41:38ENElevenLabs Hit $300 Million ARR by Building Voice Before Anyone Believed It Was the InterfacePresenta: Sarah Guo, Elad GilIn Poland, watching a foreign film is a special kind of pain: one narrator, every character, a flat delivery from start to finish.
1:35:32ENSpaceX Valuations, Harvey’s Growth and the Risks Behind AI DealsInvitado: Jason Lemkin, Rory O’Driscoll · Presenta: Harry StebbingsHarry Stebbings, Jason Lemkin and Rory O’Driscoll examine private-market prices, a proposed media deal, AI applications and fintech competition.
1:02:19ENA Top Video Model's Half-Life Is 30 Days. fal Runs 600 AnywayThe fal founders tell Sequoia why generative video is a compute-bound problem that language-model shops keep underestimating, and why they will not let any customer marry a single model.
57:37ENThe hot-swap era for models is over, OpenAI's enterprise product chief saysPresenta: Jacob EffronOlivier Godement on GPT-5.1, Codex, enterprise harnesses, and the three practices that actually make AI land inside large companies.
9:41ENAI News in December 2025: Devstral 2, Agent Standards and Data-Center PowerMatthew Berman · Forward FutureMatthew Berman discusses model-release predictions, open agent infrastructure and ideas for powering terrestrial and space-based computing.
46:26ZH朱啸虎:三年内看不到泡沫,共识太集中就错开十五度Invitado: 朱啸虎 · Presenta: 张小珺2025 年底,金沙江创投朱啸虎第三次坐到张小珺对面:AGI 不谈了,入口之争开始了。他不投人形机器人,把支票开给洗船和按摩这类「干活」的公司。
47:45ENTarek Mansour on Kalshi’s Funding, Prediction Markets and CompetitionInvitado: Tarek Mansour · Presenta: Harry StebbingsThe Kalshi CEO discusses its $1 billion raise, media partnerships, regulatory history and rivalry with Polymarket.
1:32:38ENPolish Is 1% of Frontier Training Data, So Poland Built Its Own ModelPresenta: Nathan LabenzMarek Kozlowski of Poland's National Information Processing Institute explains how PLLuM uses language adaptation, organic instructions, and on-prem small models to keep a mid-size country in control of its AI.
1:09:48ENChina's AI Labs Went From 15 to Six. The Compute Gap Is 15-to-1.Presenta: Jacob EffronJordan Schneider tells Unsupervised Learning that Chinese models were a private-capital story, that Beijing still thinks in hardware, and that the West's real edge is chips it can serve — not a state master plan.
44:17ENHarvey’s bet: make every law firm AI-first, don’t become onePresenta: Sarah Guo, Elad GilOn the day Harvey got GPT-4 access, co-founder Winston locked himself in a room for 14 hours and rebuilt his associate workflow by hand.
1:11:42ENEventbrite, Databricks and the Search for Durable Software GrowthInvitado: Jason Lemkin, Rory O’Driscoll · Presenta: Harry StebbingsHarry Stebbings, Jason Lemkin and Rory O’Driscoll debate software-market limits, AI pricing and what makes a venture investment last.
1:25:54ENHours after training ends, Z.ai puts the weights on the internetPresenta: Nathan LabenzZixuan Li of Z.ai tells ChinaTalk how a Chinese lab ships GLM, treats open weights as a ticket into Western ecosystems, and still sees itself as an upstart.
40:13ENAs long as humans still need to read code, the IDE is not dyingZed founder Nathan Sobo argues that source code is a language for people, that terminal agents still need a visual editor, and that conversations should be anchored to every edit.
42:06ENAfter SWE-bench Hits 80.9%, Dianne Na Penn Reckons with Opus 4.5Presenta: Jacob EffronAnthropic’s head of product for research on computer use, deceptive token prices, thin harnesses, and why alignment makes a model a better thinker.
7:49ENDeepSeek V3.2: Reasoning, Sparse Attention and Agent TrainingMatthew Berman · Forward FutureMatthew Berman examines DeepSeek’s December 2025 model release, its benchmark results and the techniques used to improve long-context efficiency and tool use.
1:17:01ENJonathan Siddharth on AI Training Data, Enterprise Agents and the Future of WorkInvitado: Jonathan Siddharth · Presenta: Harry StebbingsThe Turing CEO discusses reinforcement-learning environments, enterprise deployment and his expectations for AI-driven changes to software and employment.
3:18:09ENA Thing Is Alive When Its Goals Outrun Its PartsInvitado: Michael Levin · Presenta: Lex FridmanTufts biologist Michael Levin tells Lex Fridman that intelligence is a spectrum of persuadability, that even a bubble sort hosts unexpected goals, and that the brain is a thin client to a space physics does not own.
1:13:24ENRed Teams Still Land 25% of Attacks. Insurance Is the Brake.Presenta: Nathan LabenzAI Underwriting Company co-founders Rune Kvist and Rajiv Dattani tell Nathan Labenz how standards, audits, and insurance could unlock enterprise agents without waiting for a nuclear-style freeze.
2:06:16ZH谭捷:机器人通用智能的瓶颈在数据,通用形态仍需时间Invitado: 谭捷 · Presenta: 张小珺Google DeepMind 机器人研究科学家谭捷回顾了从计算机图形学到机器人、从强化学习到 Gemini Robotics 的研究路径。他认为机器人基础模型尚未成为独立学科,数据规模与覆盖度、跨本体迁移、实时推理和安全仍是走向真实应用的关键。
50:08EN1Mind's AI Superhuman Now Sources 76% of the PipelinePresenta: Nathan LabenzAmanda Kahlow tells Zapier CEO Wade Foster why she tore up the sales org — and why a face on an agent is mostly a trick to get buyers talking.
Cargar másMostrando 24 de 2901









