Build-vs-buy decisions, internal tools, SaaS threats, product design, and the practical work of making software useful.
Ryan and Mark unpack the build-vs-buy software decision: when custom tools make sense, where off-the-shelf platforms win, how internal politics and maintenance costs creep in, and why the real answer depends on appetite, resources, and organizational reality.
Short on time? Jump straight into the parts of the conversation most likely to pull you in.
“Buying gets you most of the way there fast. Building is for the gap that actually matters.”
The core setup for the whole episode: software decisions are really tradeoff decisions.
Play on this site“Sometimes the tiny missing workflow is the thing that forces a custom solution.”
A concrete example of why “just use the standard tool” breaks down inside real operations.
Play on this site“A good automated interaction raises the question: when does it matter whether a human was involved?”
The For You moment where customer service becomes a test case for human-vs-AI expectations.
Play on this site“The internet blew up centralized media and created new kinds of creators. AI may do something similar.”
A historical parallel for thinking about disruption without only counting job losses.
Play on this site“Changelog work gets skipped when it starts with a blank page.”
Herald uses GitHub activity to create a first draft, so teams can edit release notes instead of reconstructing what shipped from scratch.
Play on this site“The changelog should not create double work.”
Draft once, publish in Herald, and keep GitHub releases aligned with the same source of truth.
Play on this siteEverything in this topic path.
Ryan and Mark unpack the build-vs-buy software decision: when custom tools make sense, where off-the-shelf platforms win, how internal politics and maintenance costs creep in, and why the real answer depends on appetite, resources, and organizational reality.
Ryan and Mark use Dead Internet Theory as a jumping-off point for AI agents, social networks, human-vs-bot interaction, AI-first product design, CMS workflows, APIs, prototypes, and what happens when the internet starts talking mostly to itself.
Mark and Ryan dig into Herald, Oodle’s developer-native changelog and release notes platform built for teams that ship through GitHub but hate writing product updates from scratch.
Mark and Ryan unpack the first real AI economics hangover: overpromised frontier models, token costs, executive hype, data-center bets, Google’s distribution advantage, and why the next phase of AI value probably lives in practical application layers.
Mark and Ryan debate the AI-era version of if you build it, they will come: building products is easier than ever, but distribution, trust, attention, and demand are still the hard part.
Mark and Ryan pick up the AI-in-education thread and push it further: if AI can write the paper, build the tool, summarize the research, and automate the busy work, what exactly are humans supposed to learn, practice, and protect?
When a school's strict no-AI policy expels two seniors weeks before graduation, Mark and Ryan dig into the real question: what are we actually testing — and where's the line between AI as a calculator and AI as plagiarism?
Ryan and Mark rant productively about meetings: remote work, async vs synchronous collaboration, open offices, unstructured thinking time, calendar theater, Level 10-style structure, and why being busy is not the same thing as doing useful work.
Ryan and Mark talk through the fast-moving AI landscape: agentic models, tools, skills, business use cases, security risks, human oversight, model choice, experimentation, and what teams should understand before they shove AI into everything.
