Speed and software should go hand in hand, but with Moore's Law coming to an end and CPU clock speeds remaining steady over the last 20 years, software engineers need to turn to new techniques to improve the efficiency of their code. In this episode, Matt and Angelica and joined by Clement Jean to discuss one of these techniques: SIMD. They'll discuss what SIMD is, how it can be used from languages like Go, the various proposals to get better support for SIMD directly into Go and its compiler, and a bunch of other related topics.

Supporters get even more conversation and higher quality audio! Get access by signing up at https://fallthrough.fm/subscribe.

Thanks for tuning in and happy listening!

Notes:

• ARM Neon
• SVE
• SSE
• AVX, AVX-256, AVX-512
• Parsing Gigabytes of JSON Per Second
• Proposal #67520
• Proposal #73787
• Compiler Explorer
• Designing Data Intensive Applications, Chapter 4

Table of Contents:

• Prologue (03:36)
• Chapter 1: What is SIMD? (04:56)
• Chapter 2: SIMD in Go (13:20)
• Chapter 3: When should one use SIMD? (18:43)
• Chapter 4: Clement's journey to SIMD [Supporter Only] (21:19)
• Chapter 5: SIMD Intrinsics in Go (21:32)
• Chapter 6: SIMD in other languages (35:47)
• Chapter 7: Measuring SIMD performance (43:00)
• Chapter 8: Data Oriented Design [Supporter Only] (52:52)
• Chapter 9: SIMD-JSON (52:59)
• Chapter 10: Protocol Buffers, gRPC, and cURL [Supporter Only] (58:11)
• Epilogue (01:01:50)

Hosts

• Website
• Bluesky
• Threads
• X/Twitter
• LinkedIn
• Instagram

• (03:36) - Prologue

Building software is difficult and rarely can software engineers do it alone. In this episode, Kris is joined by Ian, Matthew, and Angelica to talk managing software projects and what three often used but rarely defined roles actually mean: project manager, product manager, and program manager. The panel also discusses how these roles interact with engineering managers, what they think these roles actually do, and whether product managers should actually exist.

For our supporters, this episode contains an extended discussion including a discussion on whether you should trust a new company you join, how to build trust, and what metrics we should measure for project success. Get access by signing up at https://fallthrough.fm/subscribe.

Thanks for tuning in and happy listening!

Table of Contents:

• Prologue (03:28)
• Chapter 1: Product? Project? Program? (05:18)
• Chapter 2: PMs and Software Engineers (13:56)
• Chapter 3: Splitting Roles (44:49)
• Chapter 4: Building Trust [Supporter Only] (53:34)
• Chapter 5: Metrics [Supporter Only] (54:13)
• Chapter 6: Should Product Managers Exist? (55:25)
• Appendix UNPOP: Unpopular Opinions (01:06:37)
• Epilogue (01:11:56)

Hosts

• Website
• Bluesky
• Threads
• X/Twitter
• LinkedIn
• Instagram

• (03:28) - Prologue

Error handling in Go is one of the language's most controversial topics. It ranks high in each Go Developer Survey and many different people have proposed many different solutions to the perceived problems. The Go Team's stance on error handling as a problem in need of a fix has changed recently, with a declaration that no proposals will be accepted for changes to Go's error handling syntax. In this episode, Kris is joined by Ian and Matt to discuss this news. They talk about whether error handling has ever actually been a problem, what the Go team actually is, and the benefits of deciding not to provided syntactic support for error handling.

For our supporters, this episode contains an extended discussion. Get access by signing up at https://fallthrough.fm/subscribe.

Thanks for tuning in and happy listening!

Notes:

• [ On | No ] syntactic support for error handling
• #71203 proposal: spec: reduce error handling boilerplate using ?
• #71460 discussion: spec: reduce error handling boilerplate using ?
• Go Wiki: SliceTricks
• slices package
• maps package
• skriptble's Bluesky post on AI hype
• #19623 proposal: spec: change int to be arbitrary precision

Table of Contents:

• Prologue (02:32)
• Chapter 1: It's Fine. (04:03)
• Chapter 2: But What Are Errors? (09:19)
• Chapter 3: Who's Responsible For Solving Problems? (26:53)
• Chapter 4: We Should Panic (38:24)
• Chapter 5: Go Is A Low Level Language (41:32) [Supporter Only]
• Chapter 6: What Is The Go Team? (41:56)
• Chapter 7: What's Next for Go? (50:41)
• Chapter 8: Boundaries & Backpressure (59:08)
• Appendix UNPOP: Unpopular Proposals (59:39) [Supporter Only]
• Appendix WASM: The New FFI (59:43)
• Epilogue (01:05:36)

Hosts

• Website
• Bluesky
• Threads
• X/Twitter
• LinkedIn
• Instagram

• (02:32) - Prologue

AI is a divisive topic. Some people are AI boosters while others are AI doomers. And then there are those of us who sit in the middle, feeling both sides are not just wrong but that they aren't experiencing a shared reality with the rest of us. In this episode, Kris and Matt are joined by Steve Klabnik to talk about the state of AI and the discourse around it.

For our supporters, this episode contains extra chapters of discussion. Get access by signing up at https://fallthrough.fm/subscribe.

Thanks for tuning in and happy listening!

Notes:

• Carl Sagan Responds To Christian Grad Student Question About God
• Weapons of Math Destruction
• Google apologises for Photos app's racist blunder
• My AI Skeptic Friends Are All Nuts
• Playing to Win
• The Infinite Game
• Glenn Vanderburg of LivingSocial on why software development is an engineering discipline
• Analogs All The Way Down

Table of Contents:

• Prologue (01:35)
• Chapter 1: Chemistry Is Just Alchemy Rebranded (07:30)
• Chapter 2: Quantity Is A Quality All On Its Own (15:57)
• Chapter 3: Optimizing The Individual, Not The Collective (25:39)
• Chapter 4: What Is Fun? (30:09)
• Chapter 5: Balancing The Good And The Bad [Supporter Only] (38:18)
• Chapter 6: The Art of Constructive Criticism [Supporter Only] (38:28)
• Chapter 7: Embracing The Meta (38:44)
• Chapter 8: What's Popular Is Popular Because It's Popular (44:56)
• Chapter 9: What Is A Computer? (51:29)
• Chapter 10: What's Popular Is Popular Because It's Popular Part II (01:09:24)
• Epilogue (01:14:09)

Hosts

• Website
• Bluesky
• Threads
• X/Twitter
• LinkedIn
• Instagram

• (01:35) - Prologue

We all use open source software on a daily basis. Even though the software is free to consume doesn't mean it's free to produce. Over the years, there have been many attempts to support open source development financially, from companies that sell support to foundations and many things in between. On this week's episode, Angelica and Matt are joined by Filippo Valsorda and Daniel McCarney to discuss Geomys, an open source maintenance company. The goal of Geomys is to provide open source maintainers with stable income to support themselves financially while doing the very important work of keeping open source projects well maintained. The group discusses the why and how of Geomys, what the day to day is like, and open source maintenance as a whole.

For our supporters, this episode contains extra chapters of discussion. Get access by signing up at https://fallthrough.fm/subscribe.

Thanks for tuning in and happy listening!

Notes:

• Geomys

Table of Contents:

• Prologue (01:11)
• Chapter 1: Geomys, An Open Source Maintainership Company (04:27)
• Chapter 2: Day To Day At Geomys [Supporter Only] (12:29)
• Chapter 3: How Are Things Going? (12:29)
• Chapter 4: The Geomys Team [Supporter Only] (19:34)
• Chapter 5: Expansion Possibilities (19:34)
• Chapter 6: Interacting With The Community (30:42)
• Chapter 7: Building A Geomys Of Your Own [Supporter Only] (40:09)
• Chapter 8: Preventing Burnout (40:09)
• Chapter 9: Prioritizing Work (52:18)
• Appendix UNPOP: Unpopular Opinions (58:06)

Hosts

• Website
• Bluesky
• Threads
• X/Twitter
• LinkedIn
• Instagram

• (01:11) - Prologue

When it comes to building distributed systems, RPC and REST style interfaces aren't the only options. Events provide an alternative way to build a distributed system that can result in more robust systems that are easier to extend and deploy, while reducing the chances of a single bottlenecked service taking down the entire system. In this episode, your host Angelica is joined by Indu Alagarsamy, Ignacio Castillejos, and Chris Richardson to discuss what event-driven architecture is, what technologies can be used with it, and how to model and evolve events.

This episode's extended content includes two chapters: one about domain drive design and another about AI/ML. Get access by signing up at https://fallthrough.fm/subscribe.

Thanks for tuning in and happy listening!

Notes:

• Async API 3.0.0
• Eventuate
• NServiceBus
• MassTransit
• Domain-Driven Design: Tackling Complexity in the Heart of Software
• Language in Context - Eric Evans - DDD Europe 2019
• DDD and LLMs - Eric Evans - Explore DDD 2024
• Vibe coding: the good, the almost, and the @#$%**
• She Is in Love With ChatGPT
• This Therapist Helped Clients Feel Better. It Was A.I.

Table of Contents:

• Prologue (01:11)
• Chapter 1: What Is Event Driven Architecture? (05:26)
• Chapter 2: Event Driven Architecture Technologies (12:49)
• Chapter 3: Evolving Event Models (29:36)
• Chapter 4: Domain Driven Design [Supporter Only] (49:52)
• Chapter 5: AI/ML [Supporter Only] (49:52)
• Epilogue (52:53)

Hosts

• Website
• Bluesky
• Threads
• X/Twitter
• LinkedIn
• Instagram

• (01:11) - Prologue

From deploying WebAssembly all over the world, to using SQLite to keep track of distributed data, to blowing glass, this episode is quite the adventure! Your co-hosting power duo of Matt and Angelica are joined by Danielle Lancashire, a Principal Engineer at Fermyon, to navigate this discussion filled with insights and advice.

This episode's extended content includes two chapters: one about the groups thoughts on artificial intelligence and another on the local tech scenes across the world. Get access by signing up at https://fallthrough.fm/subscribe.

Thanks for tuning in and happy listening!

Notes:

• Fermyon

Table of Contents:

• Prologue (01:11)
• Chapter 1: WebAssembly (02:58)
• Chapter 2: SQLite (15:20)
• Chapter 3: How To Solve Big Problems (21:35)
• Chapter 4: Artifical Intelligence [Supporter Only] (42:55)
• Chapter 5: Local Technology Communities [Supporter Only] (42:55)
• Chapter 6: Life With Work (44:18)
• Chapter 7: Parting Advice (56:07)
• Appendix UNPOP: Unpopular Opinions (58:40)
• Epilogue (01:11:13)

Hosts

• Website
• Bluesky
• Threads
• X/Twitter
• LinkedIn
• Instagram

• (01:11) - Prologue

From implementing minesweeper on a TI calculator to becoming a prolific open source maintainer, Evan Phoenix has had quite the journey. In this week's episode, Matt & Angelica sit down with Evan to discuss his journey, from his early days writing basic on his TI calculator, to writing an alternative implementation of Ruby, being GitHub user #7, becoming a Principal Engineer at Hashicorp, starting his own company (with his wife!), and so much more. This episode is packed with life lessons and advice helpful for everyone, from the most junior of engineers to the most seasoned VPs.

A few of the chapters of this week's episode are supporter only content, including long conversations about Evan's involvement with Ruby and a deep discussion of his interesting habit of being an early adopter. Get access by signing up at https://fallthrough.fm/subscribe.

Thanks for tuning in and happy listening!

Notes:

• Miren
• Rubinius

Table of Contents:

• Prologue (01:11)
• Chapter 1: The Beginning (02:34)
• Chapter 2: Building In Open Source [Supporter Exclusive] (28:26)
• Chapter 3: The Path To Principal (29:27)
• Chapter 4: Post Hashicorp (54:21)
• Chapter 5: Miren (01:00:31)
• Chapter 6: Being Early [Supporter Exclusive] (01:05:37)
• Chapter 7: East vs West [Supporter Exclusive] (01:05:37)
• Chapter 8: Remote vs In Office [Supporter Exclusive] (01:05:37)
• Chapter 9: Life With Work (01:07:09)
• Chapter 10: The Open Source Funding Problem (01:25:36)
• Epilogue (01:34:16)

Hosts

• Website
• Bluesky
• Threads
• X/Twitter
• LinkedIn
• Instagram

• (01:11) - Prologue

Little languages are powerful. From text processing with AWK to calculations with APL, domain specific languages help us do things more concisely and with more precision than with other languages. Anthony Starks knows this well, as he's designed several little languages, including decksh, the focus of this episode, which allows users to create powerful presentations without the hassle of pushing around pixels in Powerpoint. He joins Angelica and Matt to discuss the background of why he created these languages, how they work, and so much more. And we have some fantastic unpopular opinions to round out the episode.

This episode's supporter content is truly fantastic. It includes an extended conversation where the group discusses data visualization in depth, including how Anthony was able to recreate graphics from W. E. B. Du Bois that are challenging or impossible to recreate with regular data visualization tools. Get access by signing up at https://fallthrough.fm/subscribe.

Thanks for tuning in and happy listening!

Notes:

• entr
• https://speakerdeck.com/ajstarks/interactive-image-manipulation-with-popi
• https://speakerdeck.com/ajstarks/giftsh-a-dsl-and-shell-for-image-editing
• decksh object reference
• decksh
• giftsh
• giftsh code
• 2025 Du Bois Challenge
• Du Bois challenge overviews and recaps:
  • https://speakerdeck.com/ajstarks/du-bois-challenge-2024-recap
  • https://speakerdeck.com/ajstarks/challenge2023
  • https://speakerdeck.com/ajstarks/curating-the-number-duboischallenge2022
• little languages
• Du Bois Data Portraits
• Black Lives 1900
• Fyne Canvas:
  • https://speakerdeck.com/ajstarks/fc-a-high-level-canvas-api-for-the-fyne-toolkit
  • https://github.com/ajstarks/fc
• Giocanvas:
  • https://speakerdeck.com/ajstarks/giocanvas-a-high-level-canvas-api-for-gio
  • https://github.com/ajstarks/giocanvas
• ebiten canvas:
  • https://github.com/ajstarks/ebcanvas
• "Like"

Table of Contents:

• Prologue (01:29)
• Chapter 1: Why Make DSLs? (04:49)
• Chapter 2: decksh (09:57)
• Chapter 3: Visualizing Data [Supporter Only] (23:34)
• Chapter 4: Documentation (24:22)
• Chapter 5: Metaprogramming (37:33)
• Chpater 6: Good Tools (45:13)
• Appendix UNPOP: Unpopular Opinions (54:23)
• Epilogue (01:08:21)

Hosts

• Website
• Bluesky
• Threads
• X/Twitter
• LinkedIn
• Instagram

• (01:29) - Prologue

Maintaining software is challenging, but monetizing it while properly valuing your time complicate matters even more. Doing all of this for open source projects is a challenge that few have figured out. In this week's episode Matt and Angelica are joined by Carlos Becker to discuss maintaining and monetizing open source projects, what it's like to work out an open source company, and how to value your own time. Plus, we've got some unpopular opinions at the end of the episode.

Supporters get access to an extended conversation including a chapter on valuing time, the tools Carlos uses, and the groups thoughts on AI tooling! Get access by signing up at https://fallthrough.fm/subscribe.

Thanks for tuning in and happy listening!

Notes:

• Carlos's Website

Table of Contents:

• Prologue (01:11)
• Chapter 1: How Carlos Got Started (03:12)
• Chapter 2: Open Source Product Management (12:29)
• Chapter 3: Go Releaser's Architecture (15:47)
• Chapter 4: Monetization (23:23)
• Chapter 5: The Value Of Time [Extended Episode Content] (27:05)
• Chapter 6: Life at Charm (27:37)
• Chapter 7: Monetizing Open Source Projects (32:38)
• Chapter 8: Valuing Your Own Time (39:45)
• Chapter 9: The Tools Carlos Loves [Extended Episode Content] (46:57)
• Appendix UNPOP: Unpopular Opinions (46:57)
• Epilogue (56:06)

Hosts

• Website
• Bluesky
• Threads
• X/Twitter
• LinkedIn
• Instagram

• (01:11) - Prologue