"Make an omelet" is the easy part. A language model can write that recipe: crack the eggs, whisk, heat the pan. Cracking the egg is where robots stall. Jitendra Malik (Arthur J. Chick Professor at UC Berkeley and VP and Distinguished Scientist at Amazon's Frontier AI and Robotics lab) used that example at ETH Zurich to show where the hard problem sits: below the plan, in the trajectory the hand follows and the force it applies. A recipe specifies neither.
James Le
As seen in:
Amazon Web Services (AWS),
Medium,
The Startup,
arXiv,
Substack,
Legaltech News,
Towards Data Science,
KDnuggets,
DataCamp,
Built In
and
Is this you? As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.
Claim your profile
Articles by James Le
19 Lectures on Optimal Control: Stanford AA 203, Recapped — James Le
A particle sits at position 10 on a line, at rest. It has to reach the origin, also at rest, and the cost it pays trades how long the trip takes against how much control effort it burns. Lecture 4 of AA 203 solves it by hand: write the Hamiltonian, integrate the two costates, and the optimal acceleration falls out as a straight line in time. Lecture 5 introduces the numerical tools for problems where that algebra runs out.
13 Lectures on Frontier AI Systems: Stanford CS153 — James Le
Enrichment is the bottleneck to nuclear. Nuclear is the bottleneck to electricity. Electricity is the bottleneck to AI. Those three lines sat on a single slide from Scott Nolan (CEO of General Matter). His company is trying to bring uranium enrichment back to the United States, which by his account once held about 86% of the world's capacity. The last American plant closed in 2013. Nolan's was the 7th of 13 sessions. A course about frontier AI spent its midpoint on uranium.
9 Lectures on AI Economics: Stanford MS&E 435, Condensed — James Le
Jensen Huang's five-layer cake goes up on the projector twice in MS&E 435. Energy, chips, infrastructure, models, applications, with company logos stacked on each layer: Constellation and GE Vernova at the bottom, NVIDIA and Broadcom above them, then AWS and CoreWeave and Crusoe, then OpenAI and Anthropic, then Cursor and Harvey and Waymo at the top. Two of those logos belong to companies whose founders teach a week of this course.
8 Lectures on Diffusion Models: Stanford CME 296, Condensed — James Le
Every lecture in CME 296 generates the same picture: a teddy bear reading a book. It's a blob of noise on the title slide of lecture 1 and a finished image on the closing slide of lecture 8, and in between the course opens every box that prompt passes through: the autoencoder that compresses it, the transformer that denoises it, the guidance scale that makes the model listen, the benchmark that decides whether the bear is actually holding the book. 8 lectures, one prompt.
Nathan Lambert’s RLHF Course: Notes on All 13 Lectures — James Le
A reward model reads its score off a single token. You push the whole response through a transformer, take the hidden state sitting at the EOS position, run it through a scalar head, and the number that comes out is what the policy will spend the rest of training chasing. One number. For an entire answer. Almost every complaint people have about RLHF (the hedging, the refusals nobody asked for, the model that agrees with whatever you just said) is downstream of that compression.
Stanford CS 329A: 9 Lectures on Self-Improving AI Agents — James Le
METR's chart has an axis nobody used to plot: how long a task an AI agent can finish on its own. The doubling time is 7 months, traced from GPT-2 through Claude 3.7 Sonnet and measured in human work-minutes. Then the same lecture puts up the 80%-reliability version of that chart, and the horizon collapses. Both lines are real. The distance between them is where most of the hard problems in agent research live, and it turned out to be the through-line across all 9 lectures.
2025 Annual Review - The Year of Integration — James Le
💡The annual review is not just a ritual—it's a compass, a mirror, and a catalyst for intentional growth. As I enter my twelfth year of this tradition, 2025 stands out as a year defined by both acceleration and intentionality, set against the backdrop of a rapidly evolving world. As I sit down to write my twelfth consecutive annual review, I’m reminded of the enduring power of this ritual—a practice that has become both a mirror and a compass in my life.
Best Practices to Get the Most Out of AI
In the rush to adopt AI, lawyers must distinguish hype from substance to determine optimal implementation strategies. Before deployment, law firms should make informed choices about AI utilization across their organizations. These considerations range from assessing readiness and addressing ethical concerns to selecting platforms that complement existing workflows and implementing effective training protocols.
Optimize multimodal search using the TwelveLabs Embed API and Amazon OpenSearch Service | Amazon Web Services
This blog is co-authored by Head of Developer Experience – TwelveLabs The exponential growth of video content has created both opportunities and challenges. Content creators, marketers, and researchers are now faced with the daunting task of efficiently searching, analyzing, and extracting valuable insights from vast video libraries.
2024 Annual Review - The Year of Adventure — James Le
As I turn 30 and reflect on 2024, this year stands as a pivotal chapter in my life's journey. Looking back at my weekly diary entries, three major themes emerged that shaped my experiences: adaptation, growth, and connection. This milestone year brought numerous opportunities for personal and professional development. From strengthening relationships and expanding my horizons to embracing new challenges, each experience contributed to my growth journey.
Episode 133: Full Data Stack Observability with Salma Bakouk
The 133rd episode of Datacast is my conversation with Salma Bakouk, who is the CEO and co-founder of Sifflet, a Full Data Stack Observability platform.
Episode 132: Big Data Engineering, Data Culture from First Principles, and Reimagined Metadata with Suresh Srinivas
The 132nd episode of Datacast is my conversation with Suresh Srinivas, who is leading the OpenMetadata Project to build Metadata APIs & specifications and a single place to discover, collaborate, and get your data right.
Data Infra for Consumer Platforms, Algorithmic Governance, and Responsible AI with Krishna Gade — James Le
The 131st episode of Datacast is my conversation with Krishna Gade, the founder and CEO of Fiddler AI, an AI Observability startup that helps AI-forward organizations build trusted AI solutions and connect model outcomes to business KPIs. Our wide-ranging conversation touches on his research on document clustering in grad school, his early career working on the Bing’s Search Engine at Microsoft, his time as an engineering leader at Twitter and Pinterest scaling their data engineering, his...
Data Infrastructure for Consumer Platforms, Algorithmic Governance, and Responsible AI with Krishna Gade
The 131st episode of Datacast is my conversation with Krishna Gade, the founder and CEO of Fiddler AI, an AI Observability startup that helps AI-forward organizations build trusted AI solutions and connect model outcomes to business KPIs. Our wide-ranging conversation touches on his research on document clustering in grad school, his early career working on the Bing’s Search Engine at Microsoft, his time as an engineering leader at Twitter and Pinterest scaling their data engineering, his...
Episode 130: Towards Accessible Data Analysis with Emanuel Zgraggen
The 130th episode of Datacast is my conversation with Emanuel Zgraggen, the CEO and Co-Founder of Einblick - a visual computing platform that enables data teams to answer tougher, more meaningful questions by making advanced analytics and model building more streamlined and accessible.
Building Trust with Founders, VC Funding For The Cloud, and The Next Data Platform with Jason Risch — James Le
The 128th episode of Datacast is my conversation with Jason Risch, an investor on the enterprise team at Greylock — investing in security, AI/ML, data, infrastructure, and developer tools.
Episode 128: Building Trust with Founders, VC Funding For The Cloud, and The Next Platform For Data Apps with Jason Risch
The 128th episode of Datacast is my conversation with Jason Risch, an investor on the enterprise team at Greylock - investing in security, AI/ML, data, infrastructure, and developer tools.
Episode 127: Data Intelligence for Insurance Transformation with Heather Wentworth
The 127th episode of Datacast is my conversation with Heather Wentworth — the Chief Data Officer of Accelerant, an MGA-enabling insurtech focused on improving how risk is exchanged across the insurance ecosystem.
Episode 126: Vector Search Engine, Building An Open-Source Business, and Digital Technology Through The Lens of Language with Bob Van Luijt
The 126th episode of Datacast is my conversation with Bob Van Luijt, the CEO and co-founder of Weaviate, the business created around the open-source vector database Weaviate.
The Next Wave of Developer Platforms, Data Products, and Software Infrastructure with Sakib Dadi — James Le
The 125th episode of Datacast is my conversation with Sakib Dadi - a Vice President at Bessemer Venture Partners. He focuses primarily on early-stage investments in developer platforms, data products, and software infrastructure.
Episode 125: The Next Wave of Developer Platforms, Data Products, and Software Infrastructure with Sakib Dadi
The 125th episode of Datacast is my conversation with Sakib Dadi - a VP at Bessemer Venture Partners.
Episode 124: The Open-Source Cloud Playbook, The Modular Future of Data and AI Infrastructure, and Meta-Learning as a VC with Casber Wang
The 124th episode of Datacast is my conversation with - a Partner at Sapphire Ventures. He focuses primarily on security, enterprise infrastructure, and data analytics.
Episode 123: AI Monitoring, Product-Oriented Data Science, and the Israeli ML community with Itai Bar Sinai
The 123rd episode of Datacast is my conversation with Itai Bar Sinai - the Chief Product Officer and co-founder at Mona Labs, the leading AI monitoring intelligence company.
The 7 NLP Techniques That Will Change How You Communicate in the Future (Part I)
What is NLP? Natural Language Processing (NLP) is a field at the intersection of computer science, artificial intelligence, and linguistics. The goal is for computers to process or “understand” natural language in order to perform tasks like Language Translation and Question Answering. With the rise of voice interfaces and chatbots, NLP is one of the most important technologies of the information age a crucial part of artificial intelligence.
Data Visualization, Scaling Data Culture, and The Future of Data Transformation with Gabi Steele — James Le
Key Takeaways Here are the highlights from my conversation with Gabi: On Her Professional Interests I moved from London, England, to New York as a teenager. I studied industrial design at Brooklyn Technical High School for Science and Engineering. I took figure drawing classes at Parsons in high school, which sparked my interest in the creative arts. Despite not knowing what to study in university, I knew I wanted to focus on creative, open-ended subjects.
Episode 122: The Evolution of Data Visualization, Scaling Data Culture, and The Future of Data Transformation with Gabi Steele
The 122nd episode of Datacast is my conversation with Gabi Steele - the co-founder and co-CEO of Preql, a no-code data transformation tool that empowers business users to model and manage their own metrics.
High-Performance Processing Engine, Modern Data Streaming, and Minority in Tech with Alex Gallego — James Le
The backstory is that Shinji is a badass. She's a friend and a fan who just started her new company called Select Star. Rob is also an old-school buddy of mine who recently joined my team. It's really great to have such talented people on board at Concord. When I was at YieldMo, I was using a project called Apache Storm.
Episode 121: High-Performance Processing Engine, Modern Data Streaming, and Propelling Minority in Tech with Alex Gallego
The 121st episode of Datacast is my conversation with Alexander Gallego, the founder and CEO of Redpanda Data, a high-performance, Apache Kafka-compatible data streaming platform for mission-critical workloads. He has spent his career immersed in deeply technical environments and is passionate about finding and building solutions to the challenges of modern data streaming.
Next-Gen Experimentation, Statistics Engineering, and The Modern Growth Stack with Chetan Sharma — James Le
I grew up on the east coast in Connecticut and then went out west to attend Stanford for my undergraduate degree in electrical engineering. Like many undergraduates, I had a general interest in the field, but my focus became more refined over time. I knew I wanted to build things and found electronics and coding fascinating, so I knew I wanted to go in that direction.
Episode 120: Next-Generation Experimentation, Statistics Engineering, and The Modern Growth Stack with Chetan Sharma
The 120th episode of Datacast is my conversation with , the Founder & CEO of Eppo, a next-gen A/B experimentation platform that is designed to spur entrepreneurial culture.
Episode 119: Experimentation Culture, Immutable Data Warehouse, The Data Collaboration Problem, and The Rise of Data Contracts with Chad Sanderson
The 119th episode of Datacast is my conversation with , who was previously the Product Lead for Convoy’s Data Platform team. He has built everything from feature stores, experimentation platforms, metrics layers, streaming platforms, analytics tools, data discovery systems, and workflow development platforms.
The 5 Computer Vision Techniques That Will Change How You See The World
The task to define objects within images usually involves outputting bounding boxes and labels for individual objects. This differs from the classification / localization task by applying classification and localization to many objects instead of just a single dominant object. You only have 2 classes of object classification, which means object bounding boxes and non-object bounding boxes. For example, in car detection, you have to detect all cars in a given image with their bounding boxes.
Episode 118: Overcoming Hardships, Confident Learning, Dataset Improvement, and The Ph.D. Rapper with Curtis Northcutt
The 118th episode of Datacast is my conversation with Curtis Northcutt, an American computer scientist and entrepreneur focusing on AI to empower people.
Vector Databases, The Embeddings Revolution, and Working in China with Frank Liu — James Le
Stanford is an amazing place, and the four or five years I spent there were some of the most formative years of my life. I met different people, experienced different cultures, and took a variety of classes. However, Stanford can also be a bubble, as many of my fellow graduates would agree. Nevertheless, it was a unique and formative experience.
Episode 117: Vector Databases, The Embeddings Revolution, and Working in China with Frank Liu
The 117th episode of Datacast is my conversation with Frank Liu, the Director of Operations at Zilliz with nearly a decade of industry experience in machine learning and hardware engineering.
Distributed Databases, Open-Source Standards, and Streaming Data Lakehouse with Vinoth Chandar — James Le
Voldemort was started by Jay Kreps at LinkedIn, but he moved on to work on Kafka. When I joined the team, Voldemort was serving approximately 10,000 QPs. By the time I left in 2014, we had scaled the system to around 2.5 million or 3 million QPS. LinkedIn grew from approximately 80 million to over 500 million or 600 million members during that period, so it was a real hyper-growth phase for the company. My experience at LinkedIn really helped me develop my operational skills.
Episode 116: Distributed Databases, Open-Source Standards, and Streaming Data Lakehouse with Vinoth Chandar
The 116th episode of Datacast is my conversation with Vinoth Chandar - the creator and PMC chair of the Apache Hudi project and the founder of Onehouse (a cloud-native managed lakehouse to make data lakes easier, faster and cheaper).
Distributed Databases, Open-Source Standards, and Streaming Data Lakehouse with Vinoth Chandar on Apple Podcasts
Vinoth Chandar is the creator and PMC chair of the Apache Hudi project, a seasoned distributed systems/database engineer, and a dedicated entrepreneur. He has deep experience with databases, distributed systems, and data systems at the planet scale, strengthened through his work at Oracle, Linkedin, Uber, and Confluent. During his time at Uber, he created Hudi, which pioneered transactional data lakes as we know them today, to solve unique speed and scale needs for Uber’s massive data platform.
Product-Led Sales, Community-Led Category Creation, and Unlocking Revenue Data with Alexa Grabell — James Le
The 115th episode of Datacast is my conversation with Alexa Grabell, the co-founder and CEO of Pocus, a Revenue Data platform that is purpose-built for GTM teams to analyze, visualize, and action data about their prospects and customers without needing engineers.
Episode 115: Product-Led Sales, Community-Led Category Creation, and Unlocking Revenue Data with Alexa Grabell
The 115th episode of Datacast is my conversation with Alexa Grabell, the co-founder and CEO of Pocus, a Revenue Data platform that is purpose-built for GTM teams to analyze, visualize, and action data about their prospects and customers without needing engineers.
Product-Led Sales, Community-Led Category Creation, and Unlocking Revenue Data with Alexa Grabell on Apple Podcasts
Alexa Grabell is the co-founder and CEO of Pocus, a Revenue Data platform that is purpose-built for GTM teams to analyze, visualize, and action data about their prospects and customers without needing engineers. Alexa’s passion for Product-Led Sales started when she led sales strategy & operations at Dataminr, where she built internal solutions to equip sales teams with data. She studied engineering at Vanderbilt University and received her MBA from Stanford University. Top Podcasts In Technology
Building Data Products and Unlocking Data Insights with Carlos Aguilar — James Le
The 114th episode of Datacast is my conversation with Carlos Aguilar, the CEO and Founder of Glean, a data visualization startup based in New York City.
Episode 114: Building Data Products and Unlocking Data Insights with Carlos Aguilar
The 114th episode of Datacast is my conversation with Carlos Aguilar, the CEO and Founder of Glean, a data visualization startup based in New York City.
Episode 113: Data Applications, Real-Time Analytics, and Cloud Product Management with Shruti Bhat
The 113th episode of Datacast is my conversation with Shruti Bhat, the CPO and SVP of Marketing at Rockset, a real-time analytics database for building data-intensive applications at scale.
Data Applications, Real-Time Analytics, and Cloud Product Management with Shruti Bhat on Apple Podcasts
Shruti Bhat leads product management and marketing at Rockset. Before Rockset, she led Product Management for Oracle Cloud, focusing on AI, IoT, and Blockchain. Previously, she was the VP of Marketing at Ravello Systems, where she drove the start-up's rapid growth from pre-launch to hundreds of customers and a successful acquisition. Before that, she was responsible for launching VMware's vSAN and led engineering teams at HP and IBM.
Episode 112: Distributed Systems Research, The Philosophy of Computational Complexity, and Modern Streaming Database with Arjun Narayan
The 112th episode of Datacast is my conversation with Arjun Narayan, the CEO and Co-Founder of Materialize, a streaming database for real-time applications and analytics, built on top of a next generation stream processor – Timely Dataflow.
Distributed Systems Research, The Philosophy of Computational Complexity, and Modern Streaming Database with Arjun Narayan on Apple Podcasts
Arjun Narayan is the co-founder and CEO of Materialize. Materialize is a streaming database for real-time applications and analytics, built on top of a next-generation stream processor – Timely Dataflow. He was previously an engineer at Cockroach Labs and held a Ph.D. in Computer Science from the University of Pennsylvania.
Episode 111: Astrophysics, Visualization Recommendation, and Scalable Data Science with Doris Lee
The 111th episode of Datacast is my conversation with Doris Lee, the CEO and Co-Founder of Ponder, a startup delivering scalable, enterprise-ready Pandas that improve the productivity of data teams.
Show More
loading
Actions
Is this you?
As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.Get in touch with James
Contact James, search articles and posts on X, monitor coverage, and track replies from one place.
Learn more about Muck Rack