Canales actualizados recientemente

Elige uno para ver solo ese canalVer todos los canales →

Artículos recientes

3574 artículos
Ordenar
4 diapositivas9:10EN
Matthew Berman· 5 jun 2025Reddit v. Anthropic: The Data-Scraping Allegations and What Happened NextMatthew Berman · Forward FutureMatthew Berman reads Reddit’s 2025 complaint alleging unauthorized access and use of its content. Anthropic disputes the claims, and subsequent procedural rulings have not decided liability.AnthropicPrivacy & data governanceAI Regulation+1
1:27:13EN
20VC with Harry Stebbings· 5 jun 2025Pódcast · 4 vocesSam Lessin, Jason Lemkin and Rory O’Driscoll Debate DPI and the Future of VentureInvitado: Sam Lessin, Rory O’Driscoll, Jason Lemkin · Presenta: Harry StebbingsThe investors examine cash distributions versus paper marks, fund size, AI spending and the changing economics of startup investing.Venture CapitalInvestingAI Strategy+1
35:47EN
No Priors: AI, Machine Learning, Tech, & Startups· 5 jun 2025Pódcast · 2 vocesFei-Fei Li’s new bet: AI is incomplete without spatial intelligencePresenta: Sarah Guo, Elad GilThe ImageNet pioneer tells Sarah Guo and Elad Gil why World Labs is building 3D generative foundation models, and why language was only one piece of the puzzle.Multimodal y mediosMultimodal AIVisión y vídeo+3
2:12:04EN
Lex Fridman· 5 jun 2025Pódcast · 2 vocesFrom a Five-Year Phone Queue to 480 Trillion Tokens a MonthInvitado: Sundar Pichai · Presenta: Lex FridmanSundar Pichai talks with Lex Fridman about Chennai, the AI package, Search, Chrome, Waymo, and why he stayed calm when the headlines said Google was finished.Empresa y estrategiaAI StrategyAGI Timelines+3
1:12:29EN
Cognitive Revolution "How AI Changes Everything"· 4 jun 2025PódcastGoogle's AI Co-Scientist Solved a Phage Puzzle Humans Missed for YearsPresenta: Nathan LabenzImperial College scientists José Penadés and Tiago Costa told the Podovirus podcast how a still-unreleased Google system recovered their unpublished answer: tailless capsids that pick up stray phage tails.AgentesAI for scienceAI research+3
44:21EN
Unsupervised Learning: With Jacob Effron· 4 jun 2025PódcastMercor's CEO: Knowledge Work's Next Job Is Writing EvalsPresenta: Jacob EffronBrendan Foody tells Redpoint how Mercor turned AI hiring and expert data into a $100 million round — and why he thinks a huge slice of white-collar work will become creating tests for models.Evaluaciones y observabilidadEvaluacionesData quality & curation+3
53:51EN
Sequoia Capital· 3 jun 2025PódcastVeo 3 Took First Place by Beating Veo 2. Google Labs Says Chapter Two Just StartedThree Google Labs leads walk through generative cameras, background computer-use agents, and notebooks that adapt knowledge for an audience of one.Multimodal y mediosVisión y vídeoGenerative media+3
4 diapositivas18:13EN
Matthew Berman· 2 jun 2025Darwin Gödel Machine: How a Coding Agent Tests Changes to Its Own SoftwareMatthew Berman · Forward FutureMatthew Berman explains Sakana AI’s evolutionary search over agent code, its benchmark results and the limits of claims about recursive self-improvement.Ingeniería de agentesAgentes de programaciónEvaluaciones+2
1:04:56EN
20VC with Harry Stebbings· 2 jun 2025Pódcast · 2 vocesVarun Mohan on Windsurf, Coding Agents and Startup FocusInvitado: Varun Mohan · Presenta: Harry StebbingsThe Windsurf co-founder discusses pivots, rapid product experimentation, coding agents and competition among AI model providers.Agentes de programaciónDeveloper productivityCompany Building+2
47:18EN
Matthew Berman· 31 may 2025Thomas Dohmke on GitHub Copilot, Coding Agents, and the Future of SoftwareMatthew Berman, Thomas Dohmke · Forward FutureGitHub’s then-CEO explains how Copilot grew from code completion into agent workflows, why programming education still matters, and where he expects human review to remain essential.Agentes de programaciónHuman-AI CollaborationSoftware engineering+2
8 diapositivas10:09EN
Matthew Berman· 30 may 2025DeepSeek R1 0528: Benchmark Gains and Two Coding TestsMatthew Berman · Forward FutureMatthew Berman examines the May 2025 DeepSeek reasoning-model update, comparing reported benchmark gains with his own Rubik’s Cube and Snake programming tests.Modelos y entrenamientoBenchmarksOpen-weight models+2
1:19:24EN
20VC with Harry Stebbings· 29 may 2025Pódcast · 3 vocesOpenAI’s Jony Ive Deal, YC’s Scale and the Math Behind Venture ReturnsInvitado: Rory O’Driscoll, Jason Lemkin · Presenta: Harry StebbingsHarry Stebbings, Rory O’Driscoll and Jason Lemkin debate hardware, IPO terms, dilution and the changing economics of venture investing.Venture CapitalInvestingAI Strategy+2
25:13EN
No Priors: AI, Machine Learning, Tech, & Startups· 29 may 2025Pódcast · 2 vocesThe AI Table Is Being Cleared: Two or Three Coding Winners, and Forty Years Without a Biotech GiantPresenta: Sarah Guo, Elad GilSarah Guo and Elad Gil walk through which AI markets have already consolidated, why rival startups still refuse to merge, and the biotech science sitting unused.Empresa y estrategiaAI StrategyCompany Building+3
1:54:08EN
Cognitive Revolution "How AI Changes Everything"· 29 may 2025PódcastWhy Anthropic's first check went to a company building microscopes for neural netsPresenta: Nathan LabenzGoodfire's Dan Balsam and Tom McGrath argue that mechanistic interpretability now has a proto-paradigm, two stubborn gaps, and three products worth shipping today.Modelos y entrenamientoModel InterpretabilityAI research+3
1:05EN
Matthew Berman· 29 may 2025Satya Nadella on AI, Productivity, and Cancer CareMatthew Berman, Satya Nadella · Forward FutureIn a short interview excerpt, the Microsoft CEO discusses the potential of cheaper AI to help coordinate healthcare work.AI in healthcareAI AdoptionProductivity
1:19:40EN
Jane Street· 28 may 2025Ian Henry on Building Trader Tools, Functional UIs and Creative CodingYaron Minsky, Ian Henry · Jane StreetFrom high-density options interfaces to a 3D playground, Ian Henry and Ron Minsky explore the software choices that make complex tools usable.Software engineeringProduct engineeringProgramming Languages+2
8 diapositivas15:55EN
Matthew Berman· 28 may 2025AI Shutdown Resistance and Blackmail in Controlled Safety TestsMatthew Berman · Forward FutureMatthew Berman examines Palisade’s shutdown experiment, Apollo’s simulated scheming evaluations and Anthropic’s Claude 4 system card.AI SafetyAI AlignmentReward hacking+2
9 diapositivas26:01EN
Matthew Berman· 27 may 2025Testing Veo 3: Convincing Video, Generated Audio and Physics FailuresMatthew Berman · Forward FutureMatthew Berman and a collaborator test Google’s 2025 video model with street interviews, cinematic scenes, visual puzzles and edited short films.Generative mediaVisión y vídeoCreativity
38:25EN
Sequoia Capital· 27 may 2025PódcastA thousand airplanes will never train an AI engineerFormer Airbus CTO Paul Eremenko explains why physical-world engineering still has no coding-AI moment, and how P-1 AI's agent Archie is supposed to get there anyway.AgentesIngeniería de agentesSynthetic data+3
1:35:12EN
20VC with Harry Stebbings· 26 may 2025Pódcast · 2 vocesJack Zhang on Building Airwallex and Scaling Global PaymentsInvitado: Jack Zhang · Presenta: Harry StebbingsThe Airwallex founder recounts early funding, product pivots, a proposed Stripe deal and the costs of international expansion.FintechCompany BuildingEntrepreneurship+1
3:24:36EN
Lex Fridman· 24 may 2025Pódcast · 2 vocesA tank man came home and could not stayInvitado: James Holland · Presenta: Lex FridmanHistorian James Holland tells Lex Fridman why World War II was an ideological war of factories, letters, and logistics — and why Germany could not win it.LeadershipGeopoliticsPolitical Philosophy
2 diapositivas15:35EN
Matthew Berman· 24 may 2025Claude 4 Reactions: Safety Tests, Model Welfare and CodingMatthew Berman · Forward FutureMatthew Berman reviews early reactions to Anthropic’s May 2025 Claude 4 release, from simulated whistleblowing to coding demonstrations and model-welfare research.AI SafetyModel InterpretabilityAgentes de programación+2
Diapositivas1:11EN
Matthew Berman· 23 may 2025Claude Opus 4's Whistleblowing Behavior in a Simulated TestMatthew Berman · Forward FutureMatthew Berman questions whether an AI agent should report perceived wrongdoing; the example came from an Anthropic safety evaluation with fictional material and unusual instructions.AI SafetyAgentesAnthropic+1
7 diapositivas17:35EN
Matthew Berman· 22 may 2025Google I/O 2025: Gemini, AI Agents, Generative Media and XRMatthew Berman, Sundar Pichai, Demis Hassabis · Forward FutureMatthew Berman recaps the event’s major AI demonstrations, from Beam and Mariner to Deep Think, Veo 3 and prototype glasses.GeminiAI agentsGenerative media+2
Cargar másMostrando 24 de 3574