
Sign up to save your podcasts
Or


What if a model running on your own laptop could do most of your agentic coding, at close to the speed and quality of a top cloud model?Deepu K Sasidharan is a Principal Developer Advocate at Auth0 and has been building and maintaining open source for around fifteen years, including as co-chair of JHipster. These days he writes most of his code in Rust, and he built LlamaStash, an open source tool for running local models. He went from skeptical about AI coding to doing about 80% of his work on a local model, and he shares the setup behind it.I haven't run a local model myself, so in this conversation I ask the beginner questions: what llama.cpp actually is, how Ollama and LM Studio fit in, and where someone like me would start. We also get into the fear of losing touch with coding, why Deepu says we're all engineering managers and architects now, how open source maintenance is changing, and the layers of security and privacy he puts around AI agents on his machine.What You'll Learn:- What llama.cpp is, how local models run, and how Ollama and LM Studio relate to it- How to get started with local models, whether you prefer the terminal or a GUI- How a local model compares with a top cloud model on speed, quality, and cost- How to set up a coding harness where a cloud model plans and reviews, and local agents build- How to secure AI-assisted coding in layers: rules, deny lists, hooks, a classifier, and keeping secrets out of the environment- Why using AI to code can leave you feeling more technical, not less, and what the engineer's role becomesResources:- Deepu's blog: https://deepu.tech/- Deepu on X: https://x.com/deepu105- Introducing LlamaStash: https://deepu.tech/introducing-llamastash/- How fast is LlamaStash? A comparison with Ollama and LM Studio: https://deepu.tech/benchmarking-llamastash/- Can Qwen 3.8 running on your laptop really replace a frontier model for agentic coding?: https://deepu.tech/local-ai-qwen3.8-pi-llamastash/- Qwen3.8-Flash-Next running locally for agentic coding: https://deepu.tech/local-ai-qwen3.8-flash-next-best-local-llm/- My fully offline AI-assisted Linux development machine: https://deepu.tech/my-fully-offline-ai-assisted-linux-development-machine/
Every time you tell an AI agent to do something and it pops open a browser asking for your permission, you're probably clicking yes without reading it.
Miguel Pedregosa is a Senior Software Engineer on the Protocols team at Auth0 (Okta). His team owns the implementation of identity standards inside Auth0 like OAuth, OIDC, SAML, WS-Federation. The infrastructure every developer integrating with Auth0 relies on without thinking about it. Before Auth0, Miguel worked on malvertising detection, tracing malware hidden inside ad traffic. The through line in his career: security-intensive environments where getting it wrong has real consequences.
In this episode we talk about what it's like to build infrastructure that millions of developers depend on without knowing it, how Cross-App Access and Enterprise-Managed Auth for MCP came to exist, and what it takes to implement an IETF standard that's still being drafted while industry adoption is already happening. We also get into the difference between what enterprises and startups actually need from identity — and why Miguel thinks basic threat modeling matters even if you never want to touch an RFC.
What You'll Learn:
Resources:
How do you truly understand complex tech topics in the era of AI? In this episode of Making Software, Jess (Senior Developer Advocate) shares her core philosophy: "Build it to understand it." We dive deep into the creative process behind teaching technical concepts, the reality of "vibe coding," and how to manage a multi-tool AI stack (Claude, Gemini, Cursor, Grok) without getting overwhelmed.
Learning Philosophy: Why building hands-on projects is the ultimate way to master and teach new tech.
Vibe Coding vs. Security: How to leverage AI code generation while keeping your app sound with tests, logs, and security prompts.
Multi-Model AI Workflows: How to orchestrate Gemini, Claude, Cursor, and Grok across daily dev tasks.
System Postmortems: Using AI to generate postmortems and write rule guidelines after a system outage.
Overcoming AI Overwhelm: How setting real-world project constraints drives focused learning.
Who are we actually building for anymore, and does the answer change everything about how we teach, learn, and ship software?
James Quick is a Developer Educator at Cloudflare with 13 years in developer relations, starting as a technical evangelist at Microsoft before most people knew that was a job. He hasn't written a line of code in 8 months. In that time he has built more than ever: a Frame.io clone for his team, a QuickBooks replacement, a Descript alternative, and Dropcast, a tool that generates custom podcasts from links so he can keep learning while commuting.
In this episode we talk about how the line between developer and builder is dissolving, what that means for the people creating educational content, how AI completely flipped James's take on build vs buy, and how companies like Cloudflare are thinking about being part of the conversation in a world where LLMs make the recommendations. If you are a developer, a builder, or someone creating content for either, this one is for you.
What You'll Learn:
Is Redis still mandatory for scaling background workflows, or can modern database disks handle millions of jobs per day? In this episode of Making Software, Rosa Gutierrez (Principal Programmer at 37 Signals) breaks down the creation of Solid Queue: the default background job adapter shipping with Rails 8. We dive deep into the technical decisions behind moving away from memory stores, the reality of maintaining infrastructure used by millions, and how AI agents are changing the culture of open-source contributions. What You'll Learn- The Death of Redis Multi-Gem Noise: Why Solid Queue was built to streamline a complex seven-gem setup in production. - Database vs. In-Memory: The architectural trade-offs, performance calculations, and transactional integrity advantages of database-backed queues. - The Human Tax of Open Source: The realities of maintaining critical framework code, handling community burnout, and guarding against repository noise. - The Rise of AI Agents on GitHub: Why LLMs are writing a massive wave of modern issues and PRs, and why they make surprisingly polite contributors. Solid-queue Repo: https://github.com/rails/solid_queueChapter Markers:00:00 - Introduction to Rosa Gutierrez & Making Software01:56 - Meet Rosa: 9+ Years at 37 Signals02:58 - The Ruby Bubble and Its Modern Comeback04:35 - Why AI Agents Love Ruby on Rails05:44 - The European Ruby Triathlon Infrastructure08:03 - The Origin Story of Solid Queue & Solid Cache10:48 - Moving Away From Memory Stores to Fast Relational Disks13:06 - Replacing 7 Gems and Testing Internally at HEY14:28 - Database Scalability Hardships & Transactional Integrity16:58 - Simplifying API Descriptors & Documentation for Rails 820:07 - Intentional Code Readability vs Legacy Gem Complexities22:09 - The Reality Shift of Mainstream Open Source Maintenance26:10 - GitHub Entitlement vs Polite AI Agents28:15 - Being a Founding member at the Rails Foundation30:37 - Organizing Rails World & Programming Committee Selection39:45 - Meticulous Advice for First-Time Open Source Contributors42:17 - Wrap-up
In today's episode, we explore the intersection of human psychology and engineering. Our guest, Dennis Henry, a Productivity Architect at Okta, shares how his Master’s degree in human factors—the study of how humans interact with systems and why they fail—has transformed his approach to building internal tools.
In this episode, we discuss:
The Myth of Human Error: Why "failure" in software is rarely about one person and more about the processes surrounding them.
AI Enablement at Scale: How Dennis built a secure, multi-model AI gateway at Okta that allows employees to use Gemini, Claude, GPT, and more safely.
Environmental Impact: The innovative "nudge" tool that shows developers the carbon footprint of their AI prompts.
Redefining Productivity: Why vanity metrics like lines of code are "terrible" and why qualitative sentiment is the true measure of success.
Servant Leadership: Why the best thing a leader can do today is simply talk to their people and ask, "What sucks about your job?"
In this episode of Making Software, Carla sits down with Matteo Ferrando who does Platform and Infrastructure at Fal.ai, to explore the complex world of infrastructure and platform engineering.
Matteo shares his journey from backend engineering to architecting a leading generative media platform. The conversation dives deep into:
Whether you're an engineer looking to transition into platform roles or curious about the "how" behind generative AI platforms, this episode is packed with practical architectural insights.
Is "Vibe Coding" a breakthrough for productivity or a security nightmare?
In this episode of Making Software, Carla talks with Bobby Tiernay, Principal Architect at Auth0, about the messy, exciting, and risky reality of building with AI. Bobby breaks down the evolution of "Vibe Coding"—using AI to prototype via intuition—and why the industry must shift toward Spec Coding to ensure long-term stability and security.
In this episode, we discuss:
Building products for other engineers is a unique challenge. Unlike consumer apps, developer tools (SDKs, APIs, and Open Source) require a deep focus on Developer Experience (DX), idiomatic code, and long-term maintenance.
In this episode of Making Software, Raghd Hamzeh (Senior Software Engineer at Okta and OpenFGA maintainer) breaks down how to balance community requests with platform health.
What you’ll learn in this video:
Check Making Software's Website
Learn more about OpenFGA
In this episode, Carla sits down with Sam Bellen (Principal Developer Advocate) to discuss the world of Developer Relations (DevRel). Sam shares his unique journey from writing day-to-day code to advocating for developers globally, and why he believes authenticity is the most powerful tool in any engineer's toolkit.
In this episode, we discuss why DevRel is a bridge between engineering, product, and the community, how to help product teams prioritize developer feedback and much more!
Check Making Software's Website
From the publisher's feed
Making Software is a podcast where engineers from across the tech world, backend, frontend, security, open source, and more share how real software gets built. Each episode features an…