Latent Space: The AI Engineer Podcast

Latest Episodes:

Information Theory for Language Models: Jack Morris Jul 02, 2025

Show notes

Our last AI PhD grad student feature was Shunyu Yao, who happened to focus on Language Agents for his thesis and immediately went to work on them for OpenAI. Our pick this year is Jack Morris, who bucks the “hot” trends by -not- working on agents, benchmarks, or VS Code forks, but is rather known for his work on the information theoretic understanding of LLMs, starting from embedding models and latent space representations (always close to our heart).

Jack is an unusual combination of doing underrated research but somehow still being to explain them well to a mass audience, so we felt this was a good opportunity to do a different kind of episode going through the greatest hits of a high profile AI PhD, and relate them to questions from AI Engineering.

Papers and References made

* AI grad school:

* A new type of information theory:

* Embeddings

* Text Embeddings Reveal (Almost) As Much As Text: https://arxiv.org/abs/2310.06816

* Contextual document embeddings https://arxiv.org/abs/2410.02525

Harnessing the Universal Geometry of Embeddings: https://arxiv.org/abs/2505.12540

* Language models

* GPT-style language models memorize 3.6 bits per param:

* Approximating Language Model Training Data from Weights: https://arxiv.org/abs/2506.15553

* LLM Inversion

* “There Are No New Ideas In AI.... Only New Datasets”

* misc reference: https://junyanz.github.io/CycleGAN/

—

for others hiring AI PhDs, Jack also wanted to shout out his coauthor

Zach Nussbaum, his coauthor on Nomic Embed: Training a Reproducible Long Context Text Embedder.

Full Video Episode

Timestamps

00:00 Introduction to Jack Morris01:18 Career in AI03:29 The Shift to AI Companies03:57 The Impact of ChatGPT04:26 The Role of Academia in AI05:49 The Emergence of Reasoning Models07:07 Challenges in Academia: GPUs and HPC Training11:04 The Value of GPU Knowledge14:24 Introduction to Jack's Research15:28 Information Theory17:10 Understanding Deep Learning Systems19:00 The "Bit" in Deep Learning20:25 Wikipedia and Information Storage23:50 Text Embeddings and Information Compression27:08 The Research Journey of Embedding Inversion31:22 Harnessing the Universal Geometry of Embeddings34:54 Implications of Embedding Inversion36:02 Limitations of Embedding Inversion38:08 The Capacity of Language Models40:23 The Cognitive Core and Model Efficiency50:40 The Future of AI and Model Scaling52:47 Approximating Language Model Training Data from Weights01:06:50 The "No New Ideas, Only New Datasets" Thesis

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.latent.space/subscribe

Scaling Test Time Compute to Multi-Agent Civilizations — Noam Brown, OpenAI Jun 19, 2025

Show notes

Solving Poker and Diplomacy, Debating RL+Reasoning with Ilya, what’s *wrong* with the System 1/2 analogy, and where Test-Time Compute hits a wall

Full Video Episode

Timestamps

00:00 Intro – Diplomacy, Cicero & World Championship 02:00 Reverse Centaur: How AI Improved Noam’s Human Play 05:00 Turing Test Failures in Chat: Hallucinations & Steerability 07:30 Reasoning Models & Fast vs. Slow Thinking Paradigm 11:00 System 1 vs. System 2 in Visual Tasks (GeoGuessr, Tic-Tac-Toe) 14:00 The Deep Research Existence Proof for Unverifiable Domains 17:30 Harnesses, Tool Use, and Fragility in AI Agents 21:00 The Case Against Over-Reliance on Scaffolds and Routers 24:00 Reinforcement Fine-Tuning and Long-Term Model Adaptability 28:00 Ilya’s Bet on Reasoning and the O-Series Breakthrough 34:00 Noam’s Dev Stack: Codex, Windsurf & AGI Moments 38:00 Building Better AI Developers: Memory, Reuse, and PR Reviews 41:00 Multi-Agent Intelligence and the “AI Civilization” Hypothesis 44:30 Implicit World Models and Theory of Mind Through Scaling 48:00 Why Self-Play Breaks Down Beyond Go and Chess 54:00 Designing Better Benchmarks for Fuzzy Tasks 57:30 The Real Limits of Test-Time Compute: Cost vs. Time 1:00:30 Data Efficiency Gaps Between Humans and LLMs 1:03:00 Training Pipeline: Pretraining, Midtraining, Posttraining 1:05:00 Games as Research Proving Grounds: Poker, MTG, Stratego 1:10:00 Closing Thoughts – Five-Year View and Open Research Directions

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.latent.space/subscribe

The Utility of Interpretability — Emmanuel Amiesen Jun 06, 2025

Show notes

Emmanuel Amiesen is lead author of “Circuit Tracing: Revealing Computational Graphs in Language Models” (https://transformer-circuits.pub/2025/attribution-graphs/methods.html ), which is part of a duo of MechInterp papers that Anthropic published in March (alongside https://transformer-circuits.pub/2025/attribution-graphs/biology.html ).

We recorded the initial conversation a month ago, but then held off publishing until the open source tooling for the graph generation discussed in this work was released last week: https://www.anthropic.com/research/open-source-circuit-tracing

This is a 2 part episode - an intro covering the open source release, then a deeper dive into the paper — with guest host Vibhu Sapra (https://x.com/vibhuuuus ) and Mochi the MechInterp Pomsky (https://x.com/mochipomsky ). Thanks to Vibhu for making this episode happen!

While the original blogpost contained some fantastic guided visualizations (which we discuss at the end of this pod!), with the notebook and Neuronpedia visualization (https://www.neuronpedia.org/gemma-2-2b/graph ) released this week, you can now explore on your own with Neuronpedia, as we show you in the video version of this pod.

Full Video Episode

Timestamps

00:00 Intro & Guest Introductions01:00 Anthropic's Circuit Tracing Release06:11 Exploring Circuit Tracing Tools & Demos13:01 Model Behaviors and User Experiments17:02 Behind the Research: Team and Community24:19 Main Episode Start: Mech Interp Backgrounds25:56 Getting Into Mech Interp Research31:52 History and Foundations of Mech Interp37:05 Core Concepts: Superposition & Features39:54 Applications & Interventions in Models45:59 Challenges & Open Questions in Interpretability57:15 Understanding Model Mechanisms: Circuits & Reasoning01:04:24 Model Planning, Reasoning, and Attribution Graphs01:30:52 Faithfulness, Deception, and Parallel Circuits01:40:16 Publishing Risks, Open Research, and Visualization01:49:33 Barriers, Vision, and Call to Action

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.latent.space/subscribe

[AIEWF Preview] Containing Agent Chaos — Solomon Hykes Jun 03, 2025

Show notes

Solomon most famously created Docker and now runs Dagger… which has something special to share with you on Thursday.

Catch Dagger at:

- Tuesday: Dagger’s workshop https://www.ai.engineer/schedule#ship-agents-that-ship-a-hands-on-workshop-for-swe-agent-builders

- Wednesday: Dagger’s talk: https://www.ai.engineer/schedule#how-to-trust-an-agent-with-software-delivery

- Thursday: Solomon’s Keynote https://www.ai.engineer/schedule#containing-agent-chaos

Full Video Episode

Timestamps

00:00 Introduction & Guest Background00:29 What is Dagger? Post-Development Automation01:08 Dagger’s Community & Platform Engineers02:32 AI Agents and Developer Workflows03:40 Environment Isolation & The Power of Containers06:28 The Need for Standards in Agent Environments07:25 Design Constraints & Challenges for Dev Environments11:26 Limitations of Current Tools & Agent-Native UX14:11 Modularity, Customization, and the Lego Analogy16:24 Convergence of CICD and Agentic Systems17:41 Ephemeral Apps, Resource Constraints, and Local Execution21:01 Adoption, Ecosystem, and the Role of Open Source23:30 Dagger’s Modular Approach & Integration Philosophy25:38 Looking Ahead: Workshops, Keynotes, and the Future of Agentic Infrastructure

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.latent.space/subscribe

[AIEWF Preview] Gemini in 2025 and Realtime Voice AI Jun 02, 2025

Show notes

As part of our AI Engineer World’s Fair preview, we’re releasing a special cross podcast recorded with Sam Charrington of TWiML AI at last week’s Google I/O!

TUESDAY: Shrestha and Kwindla’s workshop: https://www.ai.engineer/schedule#milliseconds-to-magic-real-time-workflows-using-the-gemini-live-api-and-pipecat

TUESDAY: Kwindla’s workshop: https://www.ai.engineer/schedule#building-voice-agents-with-gemini-and-pipecat

WEDNESDAY: Shrestha and Kwindla’s talk: https://www.ai.engineer/schedule#milliseconds-to-magic-real-time-workflows-using-the-gemini-live-api-and-pipecat

WEDNESDAY: Kwindla’s keynote: https://www.ai.engineer/schedule#-voice-keynote-your-realtime-ai-is-ngmi

THURSDAY: Logan’s keynote: https://www.ai.engineer/schedule#a-year-of-gemini-progress-what-comes-next

Catch all the speakers at AIE (both workshops and talks):

Logan Kilpatrick: https://www.latent.space/p/chatgpt-gpt4-hype-and-building-llm

Shrestha Basu Mallick: https://www.linkedin.com/in/shresthabm/

Kwindla Hultman Kramer: https://www.linkedin.com/in/kwkramer

Full Video Episode

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.latent.space/subscribe

[AIEWF Preview] CloudChef: Your Robot Chef - Michellin-Star food at $12/hr (w/ Kitchen tour!) May 31, 2025

Show notes

One of the new tracks at next week’s AI Engineer conference in SF is a new focus on LLMs + Robotics, ft. household names like Waymo and Physical Intelligence. However there are many other companies applying LLMs and VLMs in the real world!

CloudChef, the first industrial-scale kitchen robotics company with one-shot demonstration learning and an incredibly simple business model, will be serving tasty treats all day with Zippy (https://www.cloudchef.co/zippy ) their AI Chef platform.

This is a lightning pod with CEO Nikhil Abraham to preview what Zippy is capable of!

https://www.cloudchef.co/platform

See a real chef comparison:

See it in the AI Engineer Expo at SF next week: https://ai.engineer

Full Video Episode

Timestamps

00:00 Welcome and Introductions00:58 What is Cloud Chef?01:36 How the Robots Work: Culinary Intelligence05:57 Commercial Applications and Early Success07:02 The Software-First Approach10:09 Business Model and Pricing13:10 Demonstration Learning: Training the Robots16:03 Call to Action and Engineering Opportunities18:45 Final Thoughts and Technical Details

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.latent.space/subscribe

The AI Coding Factory May 29, 2025

Show notes

We are joined by Eno Reyes and Matan Grinberg, the co-founders of Factory.ai. They are building droids for autonomous software engineering, handling everything from code generation to incident response for production outages. After raising a $15M Series A from Sequoia, they just released their product in GA!

https://factory.ai/

https://x.com/latentspacepod

Full Video Episode

Timestamps

00:00 Introductions 00:35 Meeting at Langchain Hackathon 04:02 Building Factory despite early model limitations 06:56 What is Factory AI? 08:55 Delegation vs Collaboration in AI Development Tools 10:06 Naming Origins of 'Factory' and 'Droids' 12:17 Defining Droids: Agent vs Workflow 14:34 Live Demo17:37 Enterprise Context and Tool Integration in Droids 20:26 Prompting, Clarification, and Agent Communication 22:28 Project Understanding and Proactive Context Gathering 24:10 Why SWE-Bench Is Dead 28:47 Model Fine-tuning and Generalization Challenges 31:07 Why Factory is Browser-Based, Not IDE-Based 33:51 Test-Driven Development and Agent Verification 36:17 Retrieval vs Large Context Windows for Cost Efficiency 38:02 Enterprise Metrics: Code Churn and ROI 40:48 Executing Large Refactors and Migrations with Droids 45:25 Model Speed, Parallelism, and Delegation Bottlenecks 50:11 Observability Challenges and Semantic Telemetry 53:44 Hiring55:19 Factory's design and branding approach 58:34 Closing Thoughts and Future of AI-Native Development

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.latent.space/subscribe

[AIEWF Preview] Multi-Turn RL for Multi-Hour Agents — with Will Brown, Prime Intellect May 23, 2025

Show notes

In an otherwise heavy week packed with Microsoft Build, Google I/O, and OpenAI io, the worst kept secret in biglab land was the launch of Claude 4, particularly the triumphant return of Opus, which many had been clamoring for. We will leave the specific Claude 4 recap to AINews, however we think that both Gemini’s progress on Deep Think this week and Claude 4 represent the next frontier of progress on inference time compute/reasoning (at last until GPT5 ships this summer).

Will Brown’s talk at AIE NYC and open source work on verifiers have made him one of the most prominent voices able to publicly discuss (aka without the vaguepoasting LoRA they put on you when you join a biglab) the current state of the art in reasoning models and where current SOTA research directions lead. We discussed his latest paper on Reinforcing Multi-Turn Reasoning in LLM Agents via Turn-Level Credit Assignment and he has previewed his AIEWF talk on Agentic RL for those with the temerity to power thru bad meetup audio.

Full Video Episode

Timestamps

00:00 Introduction to the Podcast and Guests01:00 Discussion on Claude 4 and AI Models03:07 Extended Thinking and Tool Use in AI06:47 Technical Highlights and Model Trustworthiness10:31 Thinking Budgets and Their Implications13:38 Controversy Surrounding Opus and AI Ethics18:49 Reflections on AI Tools and Their Limitations21:58 The Chaos of Predictive Systems22:56 Marketing and Safety in AI Models24:30 Evaluating AI Companies and Their Strategies25:53 The Role of Academia in AI Evaluations27:43 Teaching Taste in Research28:41 Making Educated Bets in AI Research30:12 Recent Developments in Multi-Turn Tool Use32:50 Incentivizing Tool Use in AI Models34:45 The Future of Reward Models in AI39:10 Exploring Flexible Reward Systems

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.latent.space/subscribe

⚡️The Rise and Fall of the Vector DB Category May 01, 2025

Show notes

Note from your hosts: we were off this week for ICLR and RSA! This week we’re bringing you one of the top episodes from our lightning podcast series, the shorter format, Youtube-only side podcast we do for breaking news and faster turnaround. Please support our work on YouTube! https://www.youtube.com/playlist?list=PLWEAb1SXhjlc5qgVK4NgehdCzMYCwZtiB

The explosion of embedding-based applications created a new challenge: efficiently storing, indexing, and searching these high-dimensional vectors at scale. This gap gave rise to the vector database category, with companies like Pinecone leading the charge in 2022-2023 by defining specialized infrastructure for vector operations.

The category saw explosive growth following ChatGPT’s launch in late 2022, as developers rushed to build AI applications using Retrieval-Augmented Generation (RAG). This surge was partly driven by a widespread misconception that embedding-based similarity search was the only viable method for retrieving context for LLMs!!!

The resulting “vector database gold rush” saw massive investment and attention directed toward vector search infrastructure, even though traditional information retrieval techniques remained equally valuable for many RAG applications.

Full Video Episode

Timestamps

00:00 Introduction to Trondheim and Background03:03 The Rise and Fall of Vector Databases06:08 Convergence of Search Technologies09:04 Embeddings and Their Importance12:03 Building Effective Search Systems15:00 RAG Applications and Recommendations17:55 The Role of Knowledge Graphs20:49 Future of Embedding Models and Innovations

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.latent.space/subscribe

⚡️GPT 4.1: The New OpenAI Workhorse Apr 15, 2025

Show notes

We’ll keep this brief because we’re on a tight turnaround: GPT 4.1, previously known as the Quasar and Optimus models, is now live as the natural update for 4o/4o-mini (and the research preview of GPT 4.5). Though it is a general purpose model family, the headline features are:

Coding abilities (o1-level SWEBench and SWELancer, but ok Aider)

Instruction Following (with a very notable prompting guide)

Long Context up to 1m tokens (with new MRCR and Graphwalk benchmarks)

Vision (simply o1 level)

Cheaper Pricing (cheaper than 4o, greatly improved prompt caching savings)

We caught up with returning guest Michelle Pokrass and Josh McGrath to get more detail on each!

Full Video Episode

Timestamps

Part 100:00:00 Introduction and Guest Welcome00:00:57 GPT 4.1 Launch Overview00:01:54 Developer Feedback and Model Names00:02:53 Model Naming and Starry Themes00:03:49 Confusion Over GPT 4.1 vs 4.500:04:47 Distillation and Model Improvements00:05:45 Omnimodel Architecture and Future Plans00:06:43 Core Capabilities of GPT 4.100:07:40 Training Techniques and Long Context00:08:37 Challenges in Long Context Reasoning00:09:34 Context Utilization in ModelsPart 200:10:31 Graph Walks and Model Evaluation00:11:31 Real Life Applications of Graph Tasks00:12:30 Multi-Hop Reasoning Benchmarks00:13:30 Agentic Workflows and Backtracking00:14:28 Graph Traversals for Agent Planning00:15:24 Context Usage in API and Memory Systems00:16:21 Model Performance in Long Context Tasks00:17:17 Instruction Following and Real World Data00:18:12 Challenges in Grading Instructions00:19:09 Instruction Following Techniques00:20:09 Prompting Techniques and Model Responses00:21:05 Agentic Workflows and Model PersistencePart 300:22:01 Balancing Persistence and User Control00:22:56 Evaluations on Model Edits and Persistence00:23:55 XML vs JSON in Prompting00:24:50 Instruction Placement in Context00:25:49 Optimizing for Prompt Caching00:26:49 Chain of Thought and Reasoning Models00:27:46 Choosing the Right Model for Your Task00:28:46 Coding Capabilities of GPT 4.100:29:41 Model Performance in Coding Tasks00:30:39 Understanding Coding Model Differences00:31:36 Using Smaller Models for Coding00:32:33 Future of Coding in OpenAIPart 400:33:28 Internal Use and Success Stories00:34:26 Vision and Multi-Modal Capabilities00:35:25 Screen vs Embodied Vision00:36:22 Vision Benchmarks and Model Improvements00:37:19 Model Deprecation and GPU Usage00:38:13 Fine-Tuning and Preference Steering00:39:12 Upcoming Reasoning Models00:40:10 Creative Writing and Model Humor00:41:07 Feedback and Developer Community00:42:03 Pricing and Blended Model Costs00:44:02 Conclusion and Wrap-Up

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.latent.space/subscribe

Our TOPPODCAST Picks

Follow Us

Stay Connected

Latent Space: The AI Engineer Podcast

Related Podcasts

Links

Stay Connected