Beginner Lessons

The Race to Production-Grade Diffusion LLMs [Stefano Ermon] - 764

Is It Time to Rethink LLM Pre-Training? [Aditi Raghunathan] - 747

Building an Immune System for AI Generated Software [Animesh Koratana] - 746

Autoformalization and Verifiable Superintelligence [Christian Szegedy] - 745

Closing the Loop Between AI Training and Inference [Lin Qiao] - 742

Context Engineering for Productive AI Agents [Filip Kozera] - 741

What are the four requirements of agent-to-agent collaboration? #aiagents #agntcy

Is collaboration the right path over competition among protocols like A2A, MCP, and ACP? #aiagents

Humans are probabilistic. Vijoy explains deterministic sandwich and why agents need semantic routing

LLMs for Equities Feature Forecasting at Two Sigma [Ben Wellington] - 736

Zero-Shot Auto-Labeling: The End of Annotation for Computer Vision [Jason Corso] - 735

RAG Risks: Why Retrieval-Augmented LLMs Are Not Safer [Sebastian Gehrmann] - 732

From Prompts to Policies: How RL Builds Better AI Agents [Mahesh Sathiamoorthy] - 731

CTIBench: How Good Are LLMs at Detecting Cyber Threats? [Nidhi Rastogi] - 729

Inside the “Neurons” of LLMs: Circuit Tracing Their Hidden Biology [Emmanuel Ameisen] - 727

Self-Reflecting LLMs: Reinforcement Learning That Boosts Reasoning [Maohao Shen] - 726

s1: A High-Performance Reasoning Model Trained for Under $50 [Niklas Muennighoff] - 721

Is CUDA still a moat for NVIDIA? #aihardware #aichips #podcast

What if AI hardware innovation isn’t about the future but focusing on what won’t change?

The convergence of two market forces is driving more efficient and intelligent frontier models

AI Trends 2025: AI Agents and Multi-Agent Systems with Victor Dibia - 718

Speculative Decoding and Efficient LLM Inference with Chris Lott - 717

Ensuring Privacy for Any LLM with Patricia Thaine - 716

AI Engineering Pitfalls with Chip Huyen - 715

Evolving MLOps Platforms for Generative AI and Agents with Abhijit Bose - 714

Why Agents Are Stupid & What We Can Do About It with Dan Jeffries - 713

Automated Reasoning to Prevent LLM Hallucination with Byron Cook - 712

AI for Network Management with Shirley Wu - 710

An Agentic Mixture of Experts for DevOps with Sunil Mallya - 708

Building AI Voice Agents with Scott Stephenson - 707

Is Artificial Superintelligence Imminent? with Tim Rocktäschel - 706

ML Models for Safety-Critical Systems with Lucas García - 705

AI Agents: Substance or Snake Oil with Arvind Narayanan - 704

AI Agents for Data Analysis with Shreya Shankar - 703

Supercharging Developer Productivity with ChatGPT and Claude with Simon Willison - 701

Can Google DOMINATE Enterprise AI with Gemini?

Automated Design of Agentic Systems with Shengran Hu - 700

AI Regulation and Automated Decisioning with Peter van der Putten - 699

Optimizing AI Applications for Mobile Devices with Siddhika Nevrekar - 697

Bridging the Sim2real Gap in Robotics with Marius Memmel - 695

Using a unified tool across metrics offers a holistic view, balancing quality & security trade-offs

Building a robust defense-in-depth system is like stacking Swiss cheese—no holes all the way through

Models are tested on three aspects. Applications undergo tailored checks to ensure intended behavior

Hallucinations are common in LLMs, but omission can be just as crucial in generating harmful content

What are the two key risk areas posed by generative AI and LLMs? #machinelearning #ai #llm #podcast

How Microsoft Scales Testing and Safety for Generative AI with Sarah Bird - 691

Sustainable practices are now table stakes. Andres shares opportunities & how Microsoft leads change

What is ESG? Andres breaks it down & how MS Sustainability Manager help achieve sustainability goals