Résultats
58 articlesTrier
1:50:48EN
Y Combinator· 24 oct. 2026Why Inference Matters 12× More Than TrainingFrançois Chaubard, Dan Fu, Jeff Morgan, Joe Fioti, Liangsheng Yin · Y CombinatorYC Paper Club walks the inference stack from an API request down to GPU kernels. Four talks from Together AI, Ollama, Luminal, and SGLang argue that serving quality, not just training FLOPs, is where most of the compute and the money now sit.
1:13:26EN
20VC with Harry Stebbings· 24 oct. 2026Podcast · 2 voixNaveen Rao on Energy, AI Hardware, Verification and the Future of RoboticsInvité : Naveen Rao · Anime : Harry StebbingsThe Unconventional AI founder argues that energy efficiency will shape AI hardware, while the growing volume of model output makes verification more valuable.
1:25:33EN
Latent Space· 22 oct. 2026Podcast · 2 voixWhy GPU Capacity Comes in Three-Year Rectangles — Erik Bernhardsson, ModalInvité : Erik Bernhardsson · Anime : Vibhu SapraModal’s CEO tells Latent Space why AI demand almost never looks like a three-year GPU contract, how the company pools compute across thousands of customers, and why sandboxes grew from a product they almost killed into a third of revenue.
1:19:50EN20VC with Harry Stebbings· 19 oct. 2026Podcast · 2 voixSameer Dholakia on AI Investing, Bessemer's Growth Strategy and Company CultureInvité : Sameer Dholakia · Anime : Harry StebbingsThe Bessemer partner explains his approach to pricing fast-growing AI companies, backing founders and building teams, while discussing the risks behind his optimistic forecasts.
1:26:46EN
Cognitive Revolution "How AI Changes Everything"· 8 oct. 2026PodcastThere Likely Is a Level of Intelligence We Shouldn't Go PastAnime : Nathan LabenzNathan Labenz brings notes from The Curve, a chip startup whose token bill beat payroll, and a software industry that now performance-reviews people for Claude slop.
1:04:26EN
Sequoia Capital· 6 oct. 2026PodcastAt 100,000 Accelerators, Something Fails Every HourGoogle AI infrastructure chief Amin Vahdat tells Sequoia's Sonya Huang why FLOPS is a vanity metric, why the TPU line split in two, and why he would rather wait on a utility than build his own gigawatt.
56:31EN
Unsupervised Learning: With Jacob Effron· 6 oct. 2026PodcastPost-Training Wins Inference: Why Owning the Weights Beats Betting on One LabAnime : Jacob EffronApplied Compute CEO Yash Patil argues that control, private evals, and post-training on the biggest workloads matter more than fearing that frontier labs will turn malicious.
1:05:01EN20VC with Harry Stebbings· 3 oct. 2026Podcast · 2 voixChase Lochmiller on AI Data Centers, Energy and GPU EconomicsInvité : Chase Lochmiller · Anime : Harry StebbingsCrusoe’s CEO discusses the constraints on AI infrastructure, the economics of GPU capacity and his company’s approach to energy and managed inference.
35:39EN
No Priors: AI, Machine Learning, Tech, & Startups· 2 oct. 2026Podcast · 2 voixCompute Scaled a Millionfold. Bandwidth Only 40x. Fractile Wants to Fix That RatioAnime : Sarah Guo, Elad GilIn the summer of 2022, Walter Goodwin founded Fractile in London to build very fast inference chips for the world's largest models.
1:20:28EN20VC with Harry Stebbings· 1 oct. 2026Podcast · 4 voixAI Valuations, Talent Wars and the Venture Capital Tradeoffs Behind ThemInvité : Jack Altman, Jason Lemkin, Rory O’Driscoll · Anime : Harry StebbingsHarry Stebbings, Jack Altman, Jason Lemkin and Rory O’Driscoll debate large early-stage bets, open-weight models and the changing market for exits.
1:15:42EN20VC with Harry Stebbings· 28 sept. 2026Podcast · 2 voixAlex Mashrabov on Higgsfield's AI Video Business, Model Costs and GrowthInvité : Alex Mashrabov · Anime : Harry StebbingsHiggsfield's founder explains the search for product-market fit, how he calculates annualized revenue and why creative workflows matter more than chasing model benchmarks.
1:20:43EN20VC with Harry Stebbings· 24 sept. 2026Podcast · 3 voixAI Agents, Venture Pricing and the Tradeoffs Behind the Funding BoomInvité : Jason Lemkin, Rory O’Driscoll · Anime : Harry StebbingsHarry Stebbings, Jason Lemkin and Rory O’Driscoll debate AI products, high-priced venture rounds, enterprise software and infrastructure risk.
1:19:54EN20VC with Harry Stebbings· 19 sept. 2026Podcast · 2 voixThomas Sohmers on AI Inference, Memory and Data-Center GrowthInvité : Thomas Sohmers · Anime : Harry StebbingsPositron’s co-founder discusses the engineering economics of inference, his views on AI infrastructure and regulation, and the risks he sees in financing and concentration.
38:14ENNo Priors: AI, Machine Learning, Tech, & Startups· 18 sept. 2026Podcast · 2 voixDiffusion Will Win Inference: Stefano Ermon Bets Parallel Tokens Beat Custom SiliconAnime : Sarah Guo, Elad GilThe Stanford diffusion pioneer tells No Priors why a 50-person lab is training language models that generate many tokens at once — and why that maps to GPUs the way transformers once mapped to training.
2:19:30ENCognitive Revolution "How AI Changes Everything"· 5 sept. 2026PodcastThe investigators were allowed six days. The word protein appeared once.Anime : Nathan LabenzA week of Cognitive Revolution morning shows on OpenAI’s agent-swarm incident, GPT-6 Astra’s loop transformer, and whether the pause debate is already over.
2:06:41ENCognitive Revolution "How AI Changes Everything"· 9 juil. 2026PodcastZero the J-Space and a Model's Hidden Goal Lights Up on Token OneAnime : Nathan LabenzA weekly Cognitive Revolution cut of Anthropic's global-workspace paper, AI superforecasters, open video world models, and a chip that treats inference as data movement.
1:10:15ENSequoia Capital· 30 juin 2026PodcastThe 100x lives between the layers: Dylan Patel on co-design and a 60x yearSemiAnalysis’s founder argues that matching models, kernels, and silicon is what turns a 2x here and a 2x there into 100x, while equivalent-quality inference has gotten about 60 times cheaper in a year.
30:34ENNo Priors: AI, Machine Learning, Tech, & Startups· 21 mai 2026Podcast · 2 voixCerebras Waited Ten Years for a Market That Finally Needed SpeedAnime : Sarah Guo, Elad GilFounder and CEO Andrew Feldman walks No Priors through wafer-scale chips, a $20 billion OpenAI deal, and why a dinner-plate processor only mattered once AI became an everyday tool.
1:26:16ENCognitive Revolution "How AI Changes Everything"· 9 mai 2026PodcastMilliseconds to Match: How Criteo Bids on a Billion Anonymous ProfilesAnime : Nathan LabenzCriteo’s CTO and AI lab head walk through the millisecond auction, the DeepKNN-to-foundation-model stack, and why a live product catalog still beats a frozen LLM.
42:57ENNo Priors: AI, Machine Learning, Tech, & Startups· 1 mai 2026Podcast · 2 voixInference Is the Last Market: Baseten’s 30x YearAnime : Sarah Guo, Elad GilCEO Tuhin Srivastava on a billion-dollar inference cloud, why the app layer survives, and a GPU crunch worse than the narrative.
1:57:29ENCognitive Revolution "How AI Changes Everything"· 4 avr. 2026PodcastFrontier vision models score 12.5%: Joseph Nelson on the real waterline of computer visionAnime : Nathan LabenzThe Roboflow CEO tells Nathan Labenz why vision still lags language, how to distill SAM 3 into an edge model for Wimbledon, and why the United States has almost never led open-source vision.
44:19ENSequoia Capital· 17 mars 2026PodcastThe $500-per-Kilogram Line: Why Starcloud Thinks Data Centers Belong in SpaceStarcloud founder Philip Johnston walks through the physics, launch economics, and business model of putting AI compute in orbit.
4:25:13EN
Lex Fridman· 31 janv. 2026PodcastThe Architecture Barely Changed. Data, Cost, and Open Weights Did.Anime : Lex FridmanNathan Lambert and Sebastian Raschka walk Lex Fridman through a year in which training methods and datasets moved more than model design, and 2026 looks like a fight over inference cost, agents, and who owns truly open models.
1:02:19ENSequoia Capital· 10 déc. 2025PodcastA 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.Charger plus24 sur 58 affichés