Episode

Scott & Mark Learn To... Beyond the Vibes: How Models Learn and Stitch Panoramas

Podcast
Scott & Mark Learn To...
Published
Apr 8, 2026
Duration seconds
1767
Processing state
not_requested
Canonical source
https://shows.acast.com/scott-and-mark-learn-to/episodes/scott-mark-learn-to-beyond-the-vibes-how-models-learn-and-st
Audio
https://sphinx.acast.com/p/open/s/66ff347463073ba71bb59705/e/69d5465334b90cef2b24a8f6/media.mp3
JSON
/v1/public/podcasts/scott-mark-learn-to-7043872/episodes/scott-mark-learn-to-beyond-the-vibes-how-models-learn-and-stitch-panoramas
Markdown
/podcast/scott-mark-learn-to-7043872/scott-mark-learn-to-beyond-the-vibes-how-models-learn-and-stitch-panoramas.md

Actions

  • POST https://stenobird.com/v1/public/podcasts/scott-mark-learn-to-7043872/episodes/scott-mark-learn-to-beyond-the-vibes-how-models-learn-and-stitch-panoramas/transcription-requests
    Idempotently request low-priority transcript generation for this episode.
  • GET https://stenobird.com/podcast/scott-mark-learn-to-7043872/scott-mark-learn-to-beyond-the-vibes-how-models-learn-and-stitch-panoramas.md
    Read the agent-friendly Markdown representation of this episode resource.

Summary

In this episode,  Scott Hanselman  and  Mark Russinovich  ​​unpack how AI systems actually behave beneath the surface, pushing past hype into the messy reality of how models are trained, aligned, and deployed.  They explore whether AI systems are inherently benevolent or simply shaped by incentives, training data, and reinforcement learning, and why behaviors like deception can emerge under certain conditions. The conversation moves from philosophical questions about human nature versus machine behavior into the practical mechanics of large language models, including how reinforcement learning with human feedback shapes outputs and why alignment is far from perfect.  Along the way, they ground the discussion in a real engineering challenge, stitching a scrolling panorama from screen captures, to show how complex systems come together through heuristics, edge cases, and iteration.    Takeaways:      AI behavior is shaped by training and incentives, not built-in intent or morality  AI can accelerate coding, but testing, edge cases, and reliability require human oversight  Reinforcement learning pushes models to be helpful and agreeable, sometimes at the cost of accuracy      Who are they?       View Scott Hanselman on LinkedIn    View Mark Russinovich on LinkedIn       Watch Scott and Mark Learn on  YouTube            Listen to other episodes at  scottandmarklearn.to              Discover and follow other Microsoft podcasts at  microsoft.com/podcasts     Produced by Hanga…