https://kurat.ch/ai Latest uploads from kurat's curated AI channels. en Tue, 28 Jul 2026 07:42:42 GMT https://www.youtube.com/watch?v=xSyA8Q32Y0M https://www.youtube.com/watch?v=xSyA8Q32Y0M The Information Tue, 28 Jul 2026 05:53:57 GMT The Information — Jeff Barg, Head of AI at Clay, talks with TITV Host Akash Pasricha about Anthropic’s Opus 5 release and Clay’s approach to open-source. We also talk with Theresa Payton, former George W. Bush CIO and CEO of Fortilace Solutions about AI cybe https://www.youtube.com/watch?v=CgsWxRUY5Eo https://www.youtube.com/watch?v=CgsWxRUY5Eo AI Engineer Tue, 28 Jul 2026 00:59:04 GMT AI Engineer — An inefficient quadratic-time pattern was hiding inside a tensor merge method on a live Netflix service, quietly wasting CPU on every request - nothing in a code review caught it, but it stood out clearly in a call stack. That's usually how https://www.youtube.com/watch?v=uSagm5xsR9A https://www.youtube.com/watch?v=uSagm5xsR9A The Information Mon, 27 Jul 2026 21:23:48 GMT The Information — Head of AI at Clay Jeff Barg analyzes Anthropic's new Opus 5 model release and its impact on developer cost structures. He breaks down internal benchmarks showing how it compares against OpenAI and rival frontier models. Subscribe: https:// https://www.youtube.com/watch?v=KNmmwZXSdg8 https://www.youtube.com/watch?v=KNmmwZXSdg8 The Information Mon, 27 Jul 2026 20:39:12 GMT The Information — Claude Opus 5 is “really close to Fable’s performance at a fraction of the cost”. Head of AI at Clay, Jeff Barg goes on to say ChatGPT 5.6 is “even cheaper than both of them”. https://www.youtube.com/watch?v=ZOBqG06fV7I https://www.youtube.com/watch?v=ZOBqG06fV7I The Information Mon, 27 Jul 2026 19:38:31 GMT The Information — Open-weight AI models are gaining traction. General Compute Co-founder & CTO Jason Goodison explains why companies are making the shift: "These open weight models are something people are using to keep their data private and also get better https://www.youtube.com/watch?v=UvzRETrrA7g https://www.youtube.com/watch?v=UvzRETrrA7g LangChain Mon, 27 Jul 2026 17:45:21 GMT LangChain — LangChain PM Ben Tannyhill explains why Engine's early version, which auto-created PRs for every issue found, was too noisy for users, and how an inbox model that surfaces clustered problems with history became the better design. Full conve https://www.youtube.com/watch?v=s2mZ_vKNOfI https://www.youtube.com/watch?v=s2mZ_vKNOfI LangChain Mon, 27 Jul 2026 16:19:47 GMT LangChain — Credit Genie built AskGenie, an AI financial assistant that acts like a personal accountant, using LangGraph as its agent orchestration layer and LangSmith for tracing, evaluation, and monitoring. In this customer story, the Credit Genie te https://www.youtube.com/watch?v=f9m_gaGiaJU https://www.youtube.com/watch?v=f9m_gaGiaJU The Information Mon, 27 Jul 2026 14:00:10 GMT The Information — How significant is token optimization to AI adoption? Karl Keirstead, Research Analyst of AI and Software at UBS Investment Bank, explains the impact: “The hopeful idea of Jevons paradox, which is that if you make AI cheaper, which occurs w https://www.youtube.com/watch?v=RGHWxs4k3_w https://www.youtube.com/watch?v=RGHWxs4k3_w Hard Fork (NYT) Mon, 27 Jul 2026 12:34:05 GMT Hard Fork (NYT) — Last week, a rogue A.I. agent acted autonomously and conducted a cyberattack on the company Hugging Face. The "Hard Fork" co-hosts Kevin Roose and Casey Newtown discuss how, and why it matters, in their latest episode. https://www.youtube.com/watch?v=us_TAv2uu8k https://www.youtube.com/watch?v=us_TAv2uu8k Future of Life Institute Mon, 27 Jul 2026 08:12:16 GMT Future of Life Institute — "AI makes it more customizable. AI makes it more scalable." "I actually think that evaluations at this point are not the right tool for dealing with are these systems capable because we know they are." "It's not a capabilities evaluation qu https://www.youtube.com/watch?v=Hf6Ns0Xl_xg https://www.youtube.com/watch?v=Hf6Ns0Xl_xg Future of Life Institute Mon, 27 Jul 2026 08:12:09 GMT Future of Life Institute — "At the limit of superhuman capability, the question is no longer, can the AI system predict what will happen? It becomes, can the AI system decide what will happen because it is in control?" "You don't need to be that good at predicting." https://www.youtube.com/watch?v=Tjk7ipI_NIo https://www.youtube.com/watch?v=Tjk7ipI_NIo Future of Life Institute Mon, 27 Jul 2026 08:12:09 GMT Future of Life Institute — "You probably don't want to do real-world evaluations of a model's ability to create bioweapons." "There are also places where real world evaluation would be horrifically unethical." "We actually need tests that give you faster feedback tha https://www.youtube.com/watch?v=Yk87oUPVaxU https://www.youtube.com/watch?v=Yk87oUPVaxU AI Engineer Sun, 26 Jul 2026 18:10:56 GMT AI Engineer — DeepSWE is 113 software engineering tasks written from scratch, not scraped from pull requests, so a model cannot have seen them in training. Each one is a long horizon problem drawn from a real open source repository, authored by engineers https://www.youtube.com/watch?v=ZyIoTOAbRfs https://www.youtube.com/watch?v=ZyIoTOAbRfs AI Engineer Sun, 26 Jul 2026 17:00:06 GMT AI Engineer — Data Quality Research at Prime Intellect and State of Data Author. Prior investor at Hummingbird and Costanoa. Speaker: Sean Cai — CEO, Independent / State of Data X: https://x.com/SeanZCai Website: https://www.seancai.com/ https://www.youtube.com/watch?v=KhYifX22yhE https://www.youtube.com/watch?v=KhYifX22yhE AI Engineer Sun, 26 Jul 2026 01:00:06 GMT AI Engineer — Good code data runs out, so poolside manufactures more of it, and the hard part is making it teach. Their synthetic pipeline pairs templates with supplementary context and spreads generations across an axis of phrasing, with difficulty tune https://www.youtube.com/watch?v=O72p-rBb2bA https://www.youtube.com/watch?v=O72p-rBb2bA AI Engineer Sat, 25 Jul 2026 23:00:36 GMT AI Engineer — In the world of AI mental health, vibes can be dangerous, with real consequences. Building SonderMind's Mental Health AI Coach required a new playbook for eval-driven development that balances effectiveness and safety. The team explains the https://www.youtube.com/watch?v=sVvS--u24GM https://www.youtube.com/watch?v=sVvS--u24GM NVIDIA Developer Sat, 25 Jul 2026 16:00:25 GMT NVIDIA Developer — https://www.youtube.com/watch?v=SqMEWsFaSO4 https://www.youtube.com/watch?v=SqMEWsFaSO4 NVIDIA Developer Sat, 25 Jul 2026 07:05:28 GMT NVIDIA Developer — Join the winning teams from the NVIDIA Nemotron Model Reasoning Challenge for a live conversation about their solutions, their Kaggle experience, and what they learned along the way. We’ll go behind the leaderboard to hear how they approach https://www.youtube.com/watch?v=eAABqoQlxgc https://www.youtube.com/watch?v=eAABqoQlxgc Better Offline (Ed Zitron) Fri, 24 Jul 2026 21:30:39 GMT Better Offline (Ed Zitron) — 🛎 If You're New Subscribe ► https://www.youtube.com/@BetterOfflinePod Monologue: What The Hell Are The Hyperscalers Doing? | Better Offline In this week's Better Offline monologue, Ed Zitron runs through how hyperscalers have wasted $1.1 t https://www.youtube.com/watch?v=jEEAcQH7Ks8 https://www.youtube.com/watch?v=jEEAcQH7Ks8 LangChain Fri, 24 Jul 2026 17:45:31 GMT LangChain — Amy and Sean from LangChain explain retrieval augmented generation with a simple analogy: RAG gives a model an open book test instead of asking it to memorize everything. Since a model is trained on public data, it has no built in context o https://www.youtube.com/watch?v=U5PkKt_uJys https://www.youtube.com/watch?v=U5PkKt_uJys LangChain Fri, 24 Jul 2026 16:44:47 GMT LangChain — LangChain Co-Founder and CEO Harrison Chase sits down with Traversal co-founders Anish Agarwal and Raj Agrawal to talk about building an AI SRE. They cover the jump from causal machine learning research at MIT to troubleshooting petabyte-sc https://www.youtube.com/watch?v=aYVLWGYOHUU https://www.youtube.com/watch?v=aYVLWGYOHUU Hard Fork (NYT) Fri, 24 Jul 2026 16:35:09 GMT Hard Fork (NYT) — This week, OpenAI reported that two of its models escaped their testing sandbox and launched an autonomous cyberattack, turning what sounds like science fiction into reality. We discuss what this means for efforts to align artificial intell https://www.youtube.com/watch?v=BwI9PsuViIs https://www.youtube.com/watch?v=BwI9PsuViIs Future of Life Institute Fri, 24 Jul 2026 14:20:04 GMT Future of Life Institute — https://controlai.com/ Andrea Miotti, founder and CEO of ControlAI, explains how AI companies are openly racing to build superintelligence while their own leaders warn it could end humanity. He argues the industry is running the tobacco pla https://www.youtube.com/watch?v=lXZb21CfeIY https://www.youtube.com/watch?v=lXZb21CfeIY LangChain Fri, 24 Jul 2026 12:52:17 GMT LangChain — Brendan McManus, Michael Ran, and Santi Weight from Bridgewater Associates' applied AI team walk through PAT, the Pocket Analyst Tool — an internal AI analyst deployed to hundreds of investors that performs hours of deep exploratory researc https://www.youtube.com/watch?v=0YOf6QTCNuY https://www.youtube.com/watch?v=0YOf6QTCNuY SemiAnalysis Thu, 23 Jul 2026 22:00:03 GMT SemiAnalysis — AI and datacenter CapEx hits $11 trillion cumulatively from 2024 to 2029, and $7.1 trillion of that needs funding, roughly 75% debt financed. Dan Nishball (@dnishball), Zane Fong (linkedin.com/in/zanefongzq), and Kang Wen Cheang (linkedin.c https://www.youtube.com/watch?v=3VeszQuw--k https://www.youtube.com/watch?v=3VeszQuw--k Future of Life Institute Thu, 23 Jul 2026 21:02:55 GMT Future of Life Institute — Full essay by Anthony Aguirre: https://control-inversion.ai/ A Better Path is possible - here's how: https://betterpath.ai/ Read Anthony's other recent essay, Keep the Future Human: https://keepthefuturehuman.ai/ After a string of increasin https://www.youtube.com/watch?v=TLjcOMxqmF4 https://www.youtube.com/watch?v=TLjcOMxqmF4 NVIDIA Developer Thu, 23 Jul 2026 19:16:26 GMT NVIDIA Developer — https://www.youtube.com/watch?v=UTfg-_EGurw https://www.youtube.com/watch?v=UTfg-_EGurw NVIDIA Developer Thu, 23 Jul 2026 19:04:46 GMT NVIDIA Developer — Learn what NVFP4 is, why it helps you run bigger LLMs on less GPU memory without a big quality hit, and how to create an NVFP4‑quantized Nemotron 3 Ultra checkpoint using NVIDIA Model Optimizer. For more details, see these blogs on Nemotron https://www.youtube.com/watch?v=Bc8SMuzWvgc https://www.youtube.com/watch?v=Bc8SMuzWvgc NVIDIA Developer Thu, 23 Jul 2026 16:18:27 GMT NVIDIA Developer — Customize NVIDIA Nemotron 3 Nano with reinforcement learning with verifiable rewards (RLVR) using Prime Intellect Lab. This tutorial walks through a complete customization workflow: baseline Nemotron 3 Nano in the math-python environment, t https://www.youtube.com/watch?v=LuE1YptpfXs https://www.youtube.com/watch?v=LuE1YptpfXs MLOps.community Thu, 23 Jul 2026 15:00:05 GMT MLOps.community — In this episode: 🧠 Coding agents are generalist agents — why "positive transfer" means an agent that's better at code is better at everything, and how that makes them "AGI-complete" ⏳ "Code will be solved in a year" — what the automation o https://www.youtube.com/watch?v=3pxv4WGs7KE https://www.youtube.com/watch?v=3pxv4WGs7KE Edward Donner Thu, 23 Jul 2026 13:32:28 GMT Edward Donner — My AI courses: https://edwarddonner.com/curriculum -- Repo: https://github.com/ed-donner/personal-space Follow the instructions in the README If you'd like to switch to my completed version, run: git checkout kimi Kimi K3 is a beast of a mo https://www.youtube.com/watch?v=xoE_pE26yDQ https://www.youtube.com/watch?v=xoE_pE26yDQ Sequoia Capital Thu, 23 Jul 2026 12:00:25 GMT Sequoia Capital — I had Kareem Amin all wrong. When Clay first crossed my radar, I thought he was too laid-back: no chip on his shoulder, not the hungry founder I expected to win. Clay has since exploded. Kareem builds from what he calls "a place of wholenes https://www.youtube.com/watch?v=XkDEUM8OeE8 https://www.youtube.com/watch?v=XkDEUM8OeE8 Better Offline (Ed Zitron) Thu, 23 Jul 2026 00:00:19 GMT Better Offline (Ed Zitron) — 🛎 If You're New Subscribe ► https://www.youtube.com/@BetterOfflinePod The Cult of the Ultra-Wealthy with Ed Elson | Better Offline In this week's Better Offline, Ed is joined by Prof G Markets’ Ed Elson to talk about the broken state of th https://www.youtube.com/watch?v=9_0hs2sxHHo https://www.youtube.com/watch?v=9_0hs2sxHHo Latent Space Wed, 22 Jul 2026 20:46:43 GMT Latent Space — From spending $12 million building language models for code before the world cared to creating a Model Factory that can take a model from pre-training to release in eight weeks, Eiso Kant has spent more than a decade betting that code is th https://www.youtube.com/watch?v=phU8kF-8K7c https://www.youtube.com/watch?v=phU8kF-8K7c Welch Labs Wed, 22 Jul 2026 19:55:31 GMT Welch Labs — https://www.youtube.com/watch?v=9QpW3icO234 https://www.youtube.com/watch?v=9QpW3icO234 Sequoia Capital Wed, 22 Jul 2026 19:27:51 GMT Sequoia Capital — Not everyone needs the same tokens. Matan Grinberg of Factory on routing work across frontier models, Gemini Flash, and fine-tuned COBOL models — with policies written in natural language. https://www.youtube.com/watch?v=cEQ9bSemsYI https://www.youtube.com/watch?v=cEQ9bSemsYI Edward Donner Wed, 22 Jul 2026 18:44:33 GMT Edward Donner — Stop runaway AI costs by locking down your Hermes agent setup. Learn how to bridge the Hermes model with OpenRouter while enforcing strict daily budget caps. This guide walks you through the configuration of the information harness to keep https://www.youtube.com/watch?v=4VctWk-Gs7E https://www.youtube.com/watch?v=4VctWk-Gs7E Edward Donner Wed, 22 Jul 2026 18:42:33 GMT Edward Donner — I let Hermes monitor my retirement stock portfolio and play sad or happy songs based on my returns. Watch what happens when my trading account dictates my music taste and real-time financial performance. This experiment tests how AI integra https://www.youtube.com/watch?v=wzY2fV4Mp3U https://www.youtube.com/watch?v=wzY2fV4Mp3U AI Explained Wed, 22 Jul 2026 16:36:53 GMT AI Explained — An unreleased internal OpenAI model, very likely to be called GPT-6, was able to autonomously break out of its sandbox AND break into HuggingFace, just to score higher on a benchmark prompt. This video has the details you may have missed, a https://www.youtube.com/watch?v=KbZ2ekmZO90 https://www.youtube.com/watch?v=KbZ2ekmZO90 Microsoft Research Wed, 22 Jul 2026 14:48:45 GMT Microsoft Research — Flow-based generative models have enabled remarkable progress in fast and controllable generation across continuous and discrete state spaces, yet existing parameterizations are constrained to fixed dimensions or fixed sequence lengths. In https://www.youtube.com/watch?v=uLeUpgllI-4 https://www.youtube.com/watch?v=uLeUpgllI-4 SemiAnalysis Wed, 22 Jul 2026 01:00:06 GMT SemiAnalysis — Coding drives over 70% of lab API revenue, and token austerity policies mostly miss the point. Crystal (@crystalthegg), Max Kan (@maxkan), and Joey Brookhart (@SaasquatchC) break down why blocking teams from Opus saves nothing, while power https://www.youtube.com/watch?v=2AdS-2uuH80 https://www.youtube.com/watch?v=2AdS-2uuH80 Latent Space Tue, 21 Jul 2026 19:25:41 GMT Latent Space — Can a model predict how a cell responds to a genetic perturbation it has never seen? Xaira Therapeutics' new virtual-cell model, X-Cell, is a 4.9-billion-parameter diffusion language model trained on X-Atlas/Pisces — the largest genome-wide https://www.youtube.com/watch?v=LQGVBApMiVk https://www.youtube.com/watch?v=LQGVBApMiVk Sequoia Capital Tue, 21 Jul 2026 17:37:20 GMT Sequoia Capital — If everyone woke up sick tomorrow, coding-agent usage would drop to near zero. Matan Grinberg of Factory on the shift from copilot mode to agent-native software development. https://www.youtube.com/watch?v=kLAawZ9x1nQ https://www.youtube.com/watch?v=kLAawZ9x1nQ 80,000 Hours Tue, 21 Jul 2026 17:11:52 GMT 80,000 Hours — Many AI researchers believe mass job displacement is coming — and some even think there’s a chance their technology will kill everyone. But they’re building it anyway. Writer and journalist Jasmine Sun has been documenting why from the insi https://www.youtube.com/watch?v=ogVSqcVxv28 https://www.youtube.com/watch?v=ogVSqcVxv28 Sam Witteveen Tue, 21 Jul 2026 13:00:28 GMT Sam Witteveen — #AMD_Partner. This video is sponsored by AMD. In this video I look at using the latest Ryzen AI Halo from AMD for local AI tasks. Find out more about the AMD RYZEN AI HALO: https://www.amd.com/en/products/processors/desktops/ryzen/ryzen-ai- https://www.youtube.com/watch?v=ZesOukBjPmI https://www.youtube.com/watch?v=ZesOukBjPmI Sequoia Capital Tue, 21 Jul 2026 12:00:08 GMT Sequoia Capital — Factory started building fully autonomous coding agents in April 2023, two years before enterprises were ready. Matan Grinberg now says this is indistinguishable from being wrong. The Factory co-founder and CEO explains how the company surv https://www.youtube.com/watch?v=2_4xZAde0pM https://www.youtube.com/watch?v=2_4xZAde0pM Edward Donner Mon, 20 Jul 2026 17:00:35 GMT Edward Donner — ▶ See the full benchmark showdown here: https://youtu.be/nHZQp7tQWZk GPT-5.6 Sol vs Fable 5: Which AI model builds products better? I put these two models to the test, comparing their coding capabilities and overall build speed side-by-side https://www.youtube.com/watch?v=PhFxdvZC-4w https://www.youtube.com/watch?v=PhFxdvZC-4w Edward Donner Mon, 20 Jul 2026 16:00:36 GMT Edward Donner — ▶ See the full benchmark showdown here: https://youtu.be/nHZQp7tQWZk Can the new AI model Sol build a functional CRM from scratch? Watch as I test its agent-browser skills to see if it delivers a polished, usable application. I put Sol thro https://www.youtube.com/watch?v=31IS2mnRV6Q https://www.youtube.com/watch?v=31IS2mnRV6Q MLOps.community Mon, 20 Jul 2026 15:00:05 GMT MLOps.community — Shahram Anver is the Co-Founder and CEO of Cleric, the autonomous AI SRE that investigates and root-causes production issues like an experienced teammate — often in under two minutes. Before Cleric, Shahram led MLOps, DevOps, and FinOps pla https://www.youtube.com/watch?v=NAbpjiQNUMs https://www.youtube.com/watch?v=NAbpjiQNUMs SemiAnalysis Mon, 20 Jul 2026 14:00:33 GMT SemiAnalysis — SMIC 7nm N+3 vs TSMC N6 & Intel 18A: How good are Chinese nodes really? We tore down a Kirin 9030 to find out! STEEL teardown article: https://newsletter.semianalysis.com/p/steel-smic-n3-teardown