The freeCodeCamp Podcast

#208 The three paths AI could take from here - Shawn Wang SWYX interview

49 snips
Feb 20, 2026
Shawn Wang, software engineer, founder of the AI Engineer Conference, and host of the Latent Space podcast. He explores three research frontiers: memory/world models, multimodality, and embodied AI. He contrasts improving model weights with surrounding harness systems and discusses how developers should rethink learning and career strategies in an AI-driven world.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
INSIGHT

AI Agents Will Amplify Productivity Gaps

  • Teams commanding many coding agents will massively outstrip others, increasing inequality.
  • The opportunity cost of waiting to adopt AI tools has risen sharply.
INSIGHT

Reasoning Models Are Powerful But Costly

  • Reasoning models improved quality but are slow and costly to run at scale.
  • Further progress will come from memory, world models, and multimodality rather than just longer chains of thought.
INSIGHT

Three Frontiers Beyond LLMs

  • Three research frontiers are memory/continual learning, world modeling, and multimodality/embodiment.
  • Progress in these areas will enable persistent, grounded, and physically capable AI agents.
Get the Snipd Podcast app to discover more snips from this episode
Get the app