Adventures in DevOps

Share Episode

We sit down with Jaikumar Ganesh, Head of Engineering at AnyScale, to explore the intricacies of heterogeneous compute. He unpacks the growing CPU/GPU divide, detailing how ML pipelines require precise orchestration — using CPUs for data reading and writing while leveraging expensive, massive-die GPUs for chunking and embedding.

Warren brings the insight that, with AI agents rapidly changing how software is created, building is now a requirement of the business-focused team. And our guest shares how sales and marketing departments are increasingly using tools like Cursor and Claude to develop their own workflow automations. We discuss the challenges that this shift begs: what is engineering really doing for us?

JK emphasizes that the core responsibility of the engineering organization is reliability. While anyone can generate code, running stable production software requires the deep "battle scars", robust observability, and meticulous release processes that only a dedicated engineering team can provide.

That results in needing to find the right talent. But, finding the talent to maintain this critical infrastructure isn't easy, which is why JK advocates for highly creative hiring strategies. He shares incredible success stories of bypassing traditional recruiting by running hiring ads in foreign-language movies at local movie theaters and setting up booths at social food festivals to find uniquely qualified candidates.

🎯 Picks:
Warren - Archer's Don't Fire Volleys
JK - Book: The Explorer's Gene

Share Episode

In this adventure, we sit down with Dan Wahlin, Principal of DevRel for JavaScript, AI, and Cloud at Microsoft, to explore the complexities of modern infrastructure. We examine how cloud platforms like Azure function as "building blocks". Which of course, can quickly become overwhelming without the right instruction manuals. To bridge this gap, one potential solution we discuss is the emerging reliance on AI "skills"—specialized markdown files. They can give coding agents the exact knowledge needed to deploy poorly documented complex open-source projects to container apps without requiring deep infrastructure expertise.

And we are saying the silent part outloud, as we review how handing the keys over to autonomous agents introduces terrifying new attack vectors. It's the security nightmare of prompt injections and the careless execution of unvetted AI skills. Which is a blast from the past, and we reminisce how current downloading of random agent instructions to running untrusted executables from early internet sites. While tools like OpenClaw purport to offer incredible automation, such as allowing agents to scour the internet and execute code without human oversight, it's already led us to disastrous leaks of API keys. We emphasize the critical necessity of validating skills through trusted repositories where even having agents perform security reviews on the code before execution is not enough.

Finally, we tackle the philosophical debate around AI productivity and why Dorota's LLMs raise the floor and not the ceiling is so spot on. The standout pick requires mentioning, a fascinating 1983 paper titled "Ironies of Automation" by Lisanne Bainbridge. This paper perfectly predicts our current dilemma: automating systems often leaves the most complex, difficult tasks to human operators, proving that as automation scales, the need for rigorous human monitoring actually increases, destroying the very value that was attempting to be captured by the original innovation.

💡 Notable Links:
Agent Skill Marketplace
AI Fatigue is real
Episode: Does Productivity even exist?
🎯 Picks:
Warren - Paper: Ironies of Automation (& AI)
Dan - Tool: SkillShare

Share Episode

How much do you really know about the protocol that everything is built upon? This week, we go behind the scenes with Simone Carletti, a 13-year industry veteran and CTO at DNSimple, to explore the hidden complexities of DNS. We attempt to uncover why exactly DNS is often the last place developers check during an outage, drawing fascinating parallels between modern web framework abstractions and network-level opaqueness.

Simone shares why his team relies on bare-metal machines instead of cloud providers to run their Erlang-based authoritative name servers, highlighting the critical need to control BGP routing. We trade incredible war stories, from Facebook locking themselves out of their own data centers due to a BGP error, to a massive 2014 DDoS attack that left DNSimple unable to access their own log aggregation service. The conversation also tackles the reality of implementing new standards like SVCB and HTTPS records, and why widespread DNSSEC adoption might require an industry-wide mandate.

And of course we have the picks, but I'm not spoiling this weeks, just yet...

💡 Notable Links:
Episode: IPv6
SVCB + HTTPS DNS Resource Records RFC 9460
Avian Carrier RFC 1149
🎯 Picks:
Warren - Book: One Second After
Simone - Recommended diving locations in Italy and Wreck diving projects

Share Episode

We are joined by Daan Boerlage, CTO at Mavexa as we tackle the long-awaited arrival of IPv6 in cloud infrastructure. Here, we highlight how migrating to an IPv6-native setup eliminates public/private subnet complexity and expensive NAT gateways natively. As well as entirely sidestepping the nightmare of IP collisions during VPC peering.

Beyond the financial savings of ditching IPv4 charges, we explore the technical superiority of IPv6. Daan breaks down just how mind-bogglingly large the address space is, and focuses on how it solves serverless IP exhaustion while systematically debunking the pervasive myth that NAT is a security feature. We also discuss how IPv6's end-to-end connectivity, paving the way for next-generation protocols like QUIC, HTTP/3, and WebTransport.

The episode rounds out with a cathartic venting session about legacy architecture, detailing a grueling nine-year migration away from a central shared database that ironically culminated in a move to Salesforce. Almost by design, Daan recommends his pick, praising its intuitive use of signals and fine-grained reactivity over React. And Warren's pick explores storing data in the internet itself by leveraging the dwell time of ICMP ping packets.

💡 Notable Links:
FOSDEM talk on the internet of threads
Hilbert Map of IPv6 address space
🎯 Picks:
Warren - Harder Drive: what we didn't want or need
Daan - SolidJS

Share Episode

In this episode, we examine how the software industry is fundamentally changing. We're joined by our expert guest, Matt Edmunds, a long-time UX director, principal designer, and Principal UX Consultant at Tiny Pixls. The episode kicks, analyzing how early AI implementation in Applicant Tracking Systems (ATS) created rigid hiring processes that actively filter out the varied candidates who actually bring necessary diversity to engineering teams.

Of course we get to the world of "vibe coding", and revisit the poor LLM usage highlighted in the DORA 2025 report, exploring how professionals without traditional software engineering backgrounds are leveraging models to generate functional code.

Matt details his hands-on experience using the latest models of Claude Opus and Gemini Pro, successfully building low-level C virtual audio driver in 30 minutes drive by personal needs. We discuss the inherent challenges of large context windows, and coin the term "guess-driven development". To combat these hallucinations, Matt shares his strategy of using question-based prompting and anchoring the AI with comprehensive test files and documented schemas, which the models treat as an undeniable source of truth.

Beyond the code, we look at the broader economic and physical limitations of the current AI boom, noting that AI providers are operating at massive financial losses while awaiting hardware efficiency improvements.

💡 Notable Links:
Oatmeal on hating AI Art
Episode: DORA 2025 Report
🎯 Picks:
Warren - Book: Start With Why
Matt - Book: Creativity, Inc.