Stenobird

Episode

Mike Samuel - Temper

Podcast
devtools.fm: Developer Tools, Open Source, Software Development
Published
Dec 15, 2025
Duration seconds
3146
Processing state
processed
Canonical source
https://podcasters.spotify.com/pod/show/devtoolsfm/episodes/Mike-Samuel---Temper-e3cbua4
Audio
https://anchor.fm/s/dd6922b4/podcast/play/112637700/https%3A%2F%2Fd3ctxlq1ktw2nl.cloudfront.net%2Fstaging%2F2025-11-15%2F414389101-44100-2-ab229d30265ae.mp3
JSON
/v1/public/podcasts/devtools-fm/episodes/mike-samuel-temper
Markdown
/podcast/devtools-fm/mike-samuel-temper.md

Actions

  • POST https://stenobird.com/v1/public/podcasts/devtools-fm/episodes/mike-samuel-temper/transcription-requests
    Idempotently request low-priority transcript generation for this episode.
  • GET https://stenobird.com/podcast/devtools-fm/mike-samuel-temper.md
    Read the agent-friendly Markdown representation of this episode resource.

Summary

Mike Samuel, the engineer behind the original Google Calendar, discusses his new venture, Temper, a programming language designed to translate logic across different ecosystems. The conversation explores the evolution of web development, from early AJAX-heavy applications to the future of cross-platform language interoperability.

Topics

  • Programming Languages
  • Compiler Design
  • Software Architecture
  • Transpilation
  • Web Development History
  • AI in Software Engineering
  • Cross-platform Development
  • Google Calendar

Highlights

  • Main idea: Temper acts as a high-level language that transpiles into target languages like JavaScript, Python, and Java to maintain idiomatic interfaces
  • Practical takeaway: Using a single source of truth for logic can ensure that simulations, mobile clients, and backend services all operate on identical definitions
  • Failure mode: Relying solely on AI for code generation can lead to a massive maintenance burden if the underlying specifications are not synchronized with the output
  • Technical insight: The shift from memory-oriented programming to value-oriented programming requires a fundamental change in how we approach language design
  • Future vision: Combining traditional compiler methods with AI-driven development can create a more reliable way to produce software across multiple platforms

Chapters

  1. 1:00 The Google Calendar Era: Mike reflects on his background in programming languages and the early days of building large-scale JavaScript applications at Google.
  2. 4:55 The Evolution of the Web Ecosystem: A look at how the rise of AJAX and the lack of module systems in early JavaScript shaped modern web frameworks.
  3. 9:00 Security Engineering Lessons: Discussing the challenges of protecting user data and the impact of security engineering on language design.
  4. 16:55 The Quest for Cross-Platform Instruction Sets: Exploring the difficulties of creating runtime-independent solutions and the potential of WASM-like approaches.
  5. 24:40 Introducing Temper: An overview of Temper's design philosophy: a language built to translate logic into existing, popular programming languages.
  6. 40:50 AI and the Future of Software Development: Analyzing the risks of AI-generated code and how compiler-based translation can serve as a reliable multiplier for AI tools.
  7. 48:15 Unified Logic Across the Network: The benefits of using shared definitions to synchronize logic between mobile devices, backends, and data science simulations.