Topics
Follow a thread through the archive.
19 hubs group the conversations by the problem they help you reason about. Pick one and hear how practitioners at NVIDIA, Google DeepMind, AMD and dozens more approached it.
- 19 topic hubs
- 80 episodes
- 81 explainers
New to the archive?
All 19 topic hubs
Agents 6 hubs
- AI Agents Building, orchestrating, and shipping agentic systems. 41 episodes
- RAG & Retrieval Grounding models in your data. 19 episodes
- Agent Memory How agents remember, retrieve, and reason over context. 12 episodes
- Context Management Engineering what the model sees at inference time. 7 episodes
- MCP (Model Context Protocol) The interconnect between models and external systems. 7 episodes
- Multi-Agent Systems Coordinating many agents into one working system. 4 episodes
Building and shipping 7 hubs
- AI Evaluation & Reliability Measuring, testing, and trusting AI systems. 39 episodes
- Enterprise AI Deploying AI that ships and shows ROI. 36 episodes
- Open Source AI Open models, open weights, and the ecosystem around them. 21 episodes
- AI Coding AI that writes and reviews software. 18 episodes
- AI Observability Seeing inside AI systems once they hit production. 16 episodes
- AI Security Attacking AI systems, and defending them. 6 episodes
- AI Engineering The discipline of building reliable AI products. 5 episodes
Infrastructure and models 6 hubs
- AI Hardware The chips and accelerators that make AI run. 18 episodes
- Multimodal AI Models that see, hear, and read at once. 8 episodes
- AI Infrastructure The compute, inference, and platforms underneath AI. 8 episodes
- AI Energy & Data Centers The power, cooling, and footprint behind the AI boom. 7 episodes
- Model Training Training, fine-tuning, and adapting models for real-world work. 3 episodes
- Model Architecture How models are built under the hood. 1 episode
The newsletter
New episodes and Conor’s essays, in your inbox
New episodes and Conor’s essays on what they add up to. 1,200+ subscribers. Up to once a week, no spam, unsubscribe anytime.