Swyx breaks down pre-training vs post-training and what's holding back agents
iamrobotbear · x · 2026-08-31
In the Navigators podcast, Swyx breaks down the state of the AI industry and bottlenecks. Topics include: the three acts of the AI boom; the difference between pre-training, mid-training, and post-training; the divergence between Agent Labs (Cursor, Sierra) and Model Labs; why 80% of US GDP could be automated by agents but adoption is lagging (due to chip and memory shortages); and the industry split on diffusion models.
More from coding & agent
- $400 Bounty: Build the Slack Bot That Catches YC Acceptances Before YC Announces — ThePeterMick · 2026-08-31
- With Codex and Claude Code This Good, Is Hand-Coding Agents Still Worth It? — Red_Pudding_pie · 2026-08-31
- Built a RAG app that cites course timestamps and quotes — Al_Grigor · 2026-08-31
- Using Codex with Blender beats hiring a 3D modeler for speed — natesiggard · 2026-08-31
- Exploring Object Detection Prompt Optimization for LLMs in DSPy — lateinteraction · 2026-08-31
- Report: OpenAI bought tens of thousands of Mac minis for agent training — ZeroStateReflex · 2026-08-31