The Limits Nobody Graphed

There's a slide on a board deck that says "AI-native platform." Three floors down, the AI call sits in the middle of a request handler with a string constant for a prompt, and nobody knows what one request costs.Both are true, same company, same week.In this episode, Lee goes further than his article "You're at Level 2 and You Think You're at Level 4." The article gave you a way to grade a system. This episode is about who does the grading, why the answer comes back two levels high, and which single piece of work moves you up.In this episode:Bolt-on versus AI-native, and why it's the cloud's lift-and-shift all over againThe four things that come apart in a bolt-on system: wrongness, cost, context, and changeThe five levels of AI maturity, each with a test you can actually runWhy self-assessments land two levels high, and the five questions to ask insteadGrade one system, not the companyGoing from Level 2 to Level 3: fifty cases, one threshold, one cost numberWhat to say when the evidence says Level 2 and the board deck says otherwiseLinks:Read the original articleThe AI-Native Architecture seriesFrom Cloud-Native to AI-Native: the four propertiesGet the SAI newsletterAtchison Academy courses

In March 2024, Lee wrote about a GitClear report that had analyzed 153 million lines of changed code, and said the early numbers on AI-assisted code were worrying.GitClear kept counting. The newest report covers 623 million changes through the first half of 2026.So this episode is a check on his own words: what he got right, what he got wrong, and the thing he missed completely. The metric he led with in 2024 turned out to be the least alarming one in the report. The one he underplayed got worse than he expected.In this episode:The churn number that didn't double, and what projections doDuplication up 81%, copy/paste from 9% to 16%, cross-file calls down 35%Refactoring collapsed: 21% of changed lines in 2022, under 4% nowWhy "does AI write good code?" is the wrong questionThe price change underneath it all: writing got nearly free, understanding didn'tTwo fair objections, and what he'd write today instead of "use it in moderation"Links:Read the original 2024 articleGitClear: Coding on Copilot, the 2024 reportGitClear: The Maintainability Gap, the new reportGet the SAI newsletterAtchison Academy courses

GitHub published five incidents for August 2026. Three of them are the same failure.On the 17th, seven and a half hours of degraded service: 29,000 organizations, 4.8 million requests failed or slow, 56% of front-door requests failing at the peak. Three weeks, three incidents, one blind spot — capacity ran out in places nobody was graphing.Your autoscaler scales the thing you are observing. Saturation happens in the thing you don't.In this episode:The August 17th outage, and the two beside it on the 6th and the 26thSidecar concurrency, network flow tables, database primary headroom: limits with no dashboardWhy these limits are dangerous: someone else's layer, hard rather than soft, invisibleRetry amplification, and how one data center's capacity problem became an authentication outage everywhereProtective throttles set so high they engaged after the damage was doneTwo lists to build this week: your real hard limits, and your retry codeLinks:GitHub Availability Report: August 2026Get the SAI newsletterAtchison Academy courses

It's 3am and a payment service is failing. The engineer who answers has never touched it, and the next hour goes to finding someone who can fix it.In this episode, Lee goes further than his article "When Everything Is Critical, Nothing Is." The article named the two missing decisions. This episode is about how to make them: which services matter most, and which one team owns each of them.In this episode:The four service tiers, in plain termsThree questions that keep everything from landing in Tier 1Why every tier is a budget decision, and why you should cap Tier 1The dependency trap: a Tier 1 service is only as reliable as the lower-tier service it can't live withoutWhat "owner" means: paged, planned, permittedInherited services, and an exercise to run this weekLinks:Read the original articleThe ownership-gap diagnosticSTOSAGet the SAI newsletterAtchison Academy courses

Today, we discuss the significant impact of AI on software development, particularly the use of AI assistants like GitHub Copilot. Our guest, Burke Holland, a principal developer advocate at Microsoft, shares insights on how developers are increasingly integrating AI into their workflows. We explore the current spectrum of AI usage, from code completion to AI agents that can write code autonomously. Burke emphasizes the need for developers to adapt to these changes while maintaining a critical perspective on AI-generated code. This conversation sheds light on the evolving role of software developers in an AI-enhanced landscape.In this insightful episode, Burke Holland, a principal developer advocate at Microsoft for the Visual Studio Code team, discusses the transformative role of artificial intelligence (AI) in software development. The conversation starts with the observation that developers have shifted from skepticism about AI's capabilities in coding to a more supportive stance, with many now using AI tools like GitHub Copilot to assist in their work. Burke explains how these AI tools can provide significant assistance through 'completions'—intelligent suggestions that understand the context of what the developer is working on, effectively acting as a cognitive partner in the coding process.The discussion further explores the concept of AI 'agents,' which represent a cutting-edge approach to automating aspects of coding. Burke acknowledges the potential of these agents to undertake more complex tasks but also expresses concern about the implications for developers' roles. He emphasizes the importance of human oversight, suggesting that while AI can enhance efficiency, developers must remain engaged with the code they produce to ensure quality and comprehensibility. The episode highlights the balance between leveraging AI's strengths and maintaining the critical thinking skills that are essential in programming.As the episode wraps up, Burke encourages developers to adopt an exploratory mindset toward AI tools. He notes that the landscape of software development is rapidly evolving, and by integrating AI thoughtfully, developers can enhance their productivity while still preserving their essential skills and understanding of coding. The conversation underscores a future where AI and human collaboration can lead to better software outcomes, emphasizing the need for a nuanced approach to AI integration.Takeaways:In the podcast, we discussed how AI is changing the software development landscape significantly.Burke Holland highlighted that the use of AI in coding has seen a dramatic increase recently.We talked about the importance of understanding the code produced by AI to maintain quality.Completions in AI tools like GitHub Copilot can enhance a developer's workflow remarkably.Burke shared insights on the difference between completions and agents in AI coding practices.We concluded that while AI can write code, developers still play a crucial role in ensuring its accuracy.Links referenced in this episode:softwarearchitectureinsights.comCompanies mentioned in this episode:MicrosoftVS CodeGitHub CopilotJetBrainsSpeckit

Today's discussion centers on the vulnerabilities associated with AI systems and the increasing threats they face. Our guest, Preston Wood, the Chief Security and Strategy Officer at Databox, highlights the lack of transparency in AI technologies as a significant factor that makes them more susceptible to attacks. We explore how this obfuscation creates challenges in understanding and defending against potential threats. As AI continues to advance, we also consider the evolving nature of phishing attacks and the importance of robust data management strategies to mitigate risks. This episode aims to provide insights for software architects and leaders on navigating the complexities of AI integration while ensuring security and reliability.The podcast episode features an insightful discussion about the growing vulnerabilities associated with AI systems. The guest, Preston Wood, the Chief Security and Strategy Officer at Databox, addresses the surge in AI-related attacks, emphasizing the need for greater transparency and understanding of AI operations. He explains that the ambiguous nature of AI systems makes them appealing targets for attackers, who can exploit the lack of visibility into how these systems function. Throughout the conversation, Preston highlights the importance of ensuring that AI-generated data is clean and comprehensible to mitigate risks. He compares today's AI landscape to early phishing attacks, which have evolved into sophisticated threats due to advancements in AI technology. This episode serves as a crucial resource for software architects and technology leaders, offering them guidance on how to navigate the complexities of securing AI systems and understanding the implications of AI on data management and security practices.Takeaways:The podcast discusses the growing vulnerabilities associated with AI-based systems due to their lack of transparency.Preston Wood emphasizes the importance of clean and understandable data for AI performance and security.Organizations are advised to improve their data architecture to ensure AI projects are successful and not hindered by poor data quality.The conversation highlights the evolving nature of phishing attacks, which are now more sophisticated due to AI advancements.Effective security requires a layered approach that combines model training and guardrails for AI systems.Listeners are encouraged to consider how well their organizations are integrating AI into their existing technology frameworks.

AI is no longer a novelty; it is now a significant part of our lives. In this podcast, we discuss the importance of using AI thoughtfully in software architecture. As architects, our role is not to chase the latest AI trends but to determine how AI can genuinely enhance our applications. We emphasize that AI should be viewed as a tool, not just a product strategy. We also provide guidance on evaluating AI requests, ensuring that the focus remains on achieving meaningful business outcomes while managing the complexities that come with AI implementation.Software Architecture Insights provides a comprehensive overview of the evolving role of AI in software architecture as we enter 2025. The discussion begins with the acknowledgement that AI is no longer a novel concept, but a standard aspect of modern technology. Generative AI models and multimodal systems are now common, seamlessly integrating text, images, and speech. Despite the substantial growth in AI capabilities, the hype surrounding AI continues to escalate, with companies feeling pressure to incorporate AI into their products. This episode emphasizes that software architects must not succumb to this hype. Instead, their focus should be on ensuring that AI is used appropriately and effectively within their applications. The speaker stresses that AI should be viewed as a tool to enhance products rather than a mere trend to follow. This means that architects need to translate management’s desire for AI into actionable, realistic strategies that align with business goals and product visions.The conversation delves into the responsibilities of a software architect, who must evaluate whether AI solutions genuinely address identified problems. This includes assessing the available data, understanding the types of AI needed, and determining the long-term operational costs associated with implementing AI features. The speaker underscores that not every problem requires an AI solution; sometimes, a simpler approach may suffice. For instance, traditional algorithms might be more efficient and effective than complex AI systems for certain tasks. As architects navigate these challenges, they are advised to engage in thoughtful discussions with leadership to align expectations and ensure that AI implementations are sustainable.In conclusion, the episode provides practical guidelines for evaluating AI requests in a software architecture context. Architects are encouraged to clarify business goals, assess data readiness, explore simpler alternatives, and plan for compliance and sustainability. By doing so, they can ensure that AI is integrated into products in a meaningful way, ultimately bridging the gap between innovation and practicality. The key takeaway is that AI should enhance the product and align with business objectives rather than merely serving as a checkbox to satisfy market demands.Takeaways:In 2025, AI is a significant part of everyday software development and design.Software architects must not only embrace AI but also understand its real implications and costs.AI should enhance products and services, not be an unthoughtful addition driven by hype.The role of the architect includes translating management's AI requests into meaningful technical strategies.Links referenced in this episode:softwarearchitectureinsights.comCompanies mentioned in this episode:OpenAI

Today, we discuss the advancements in optical storage technology and its implications for data archival. Our guest, Greg Kittilson, VP of Engineering at Folio Photonics, shares insights on how their innovative approach to optical storage offers multiple layers of data storage, significantly enhancing capacity. Unlike traditional methods, this technology can provide long-lasting and environmentally friendly solutions for data archiving. We explore the importance of understanding data accessibility and the cost implications of different storage solutions. Join us as we delve into the intersection of technology, sustainability, and software architecture.Software Architecture Insights presents a compelling dialogue with Greg Kittleson, the VP of Engineering at Folio Photonics. We delve into the evolution of data archiving and the pivotal role of optical storage technology in modern software architecture. Unlike traditional methods, Folio Photonics is advancing optical storage to offer sustainable and durable solutions for data archiving. Greg explains how their innovative approach overcomes the limitations of conventional optical storage, such as Blu-ray technology, by enabling significantly more data layers on a single disk through advanced polymer co-extrusion techniques. This advancement allows for larger storage capacities while maintaining lower costs, making it a viable option for enterprise-scale archival systems.As the conversation unfolds, we explore the implications of data storage in the context of environmental sustainability. Greg emphasizes the importance of reducing energy consumption in data centers and how optical storage can alleviate some of the environmental impacts associated with traditional storage solutions. We discuss the staggering amount of data generated annually, estimated at around 200 zettabytes, and how much of this data remains archived rather than actively used. This leads us to examine the different classes of data—active, near-archive, and cold archive—and how software architects can make informed decisions about data storage strategies based on access speed and cost considerations.The episode further highlights the potential for optical storage to contribute to a greener future. With minimal power requirements and fewer environmental controls needed compared to hard drives and tapes, optical storage emerges as a strong contender in the quest for sustainable data solutions. We conclude with insights into how software architects can leverage these advancements in optical storage technology, ensuring that they not only meet current data demands but also anticipate future needs in an increasingly data-driven world.Takeaways:The podcast discusses the importance of data archiving and how optical storage can improve sustainability in data management.Greg explains how Folio Photonics technology allows for multiple layers of data storage, enhancing capacity significantly.Accessibility of data is crucial for software architects, as different types of data require different storage solutions and access speeds.The conversation highlights the environmental impact of data centers and how optical storage reduces power consumption compared to hard drives.Folio's optical storage technology offers a cost-effective and reliable solution for active and cold data storage needs.The podcast emphasizes the need for software architects to understand data lifecycle and compliance requirements in their design choices.Links referenced in this episode:softwarearchitectureinsights.comfoliophotonics.comCompanies mentioned in this episode:Folio PhotonicsSonyPickle

AI is now essential for software architects. While it may seem that AI has recently become important, it has been around since the 1950s. In this episode, we discuss how AI affects modern software architecture and why understanding AI is crucial for our careers. We explore the new challenges and opportunities AI brings, such as the need for structured data and the importance of ethical considerations. As we navigate this complex landscape, it is clear that knowledge of AI will shape the future of software design and architecture.Takeaways:AI has been an integral part of software since the 1950s, not just a recent trend.As software architects, we need to understand AI's applications and implications in our work.AI creates new architectural challenges and opportunities that we must navigate effectively.Data management is critical for AI systems, requiring careful structuring and ethical sourcing.AI introduces unpredictability in software behavior, requiring us to plan for uncertainty.Ethics in AI is now a significant concern for architects, affecting our design decisions.Links referenced in this episode:softwarearchitectureinsights.com

Cloud infrastructure provisioning can be a challenging task, often requiring a delicate balance between speed and accuracy. In our discussion today, we explore how AI can streamline this process. Our guest, Marcin Vizensky, co-founder of SpaceLift, introduces us to their innovative solution called Spacelift Intent, which simplifies infrastructure management by removing the complexities of traditional tools like Terraform. Marcin explains how this platform allows users to express their needs directly, enabling quicker provisioning while maintaining necessary controls and policies. Join us as we delve into the intricacies of AI-driven cloud provisioning and its potential to make infrastructure management more accessible for developers and data scientists alike.The discussion centers around the challenges of cloud infrastructure provisioning and how AI can provide solutions. Marcin Vizensky, co-founder of SpaceLift, outlines two extremes in the current landscape: the rapid but unrepeatable manual provisioning of cloud resources through console clicks, and the slow, complex processes involving tools like Terraform that require extensive knowledge and setup. He emphasizes that many users, such as developers and data scientists, do not need to become cloud experts; they simply want to provision resources effectively. Marcin introduces Spacelift Intent, a tool designed to simplify this process by allowing users to express their needs in natural language, which AI translates into API calls. This approach shortens the lengthy deployment cycle typically associated with traditional tools, making it easier for users to manage infrastructure without deep technical expertise.Takeaways:Cloud infrastructure provisioning is challenging due to the extreme approaches we have adopted.AI can facilitate the process of managing cloud infrastructure by streamlining resource provisioning.SpaceLift Intent provides a middle ground between fast but chaotic and slow but formal cloud setups.Using AI in infrastructure management can allow for quick prototyping without extensive learning curves.The approach of SpaceLift Intent helps users transition from initial setups to formal infrastructure management.Historical data from previous configurations is preserved, allowing for easy migration and state management.Links referenced in this episode:spacelift.comopentofu.orgsoftwarearchitectureinsights.comCompanies mentioned in this episode:SpaceLiftOpen TOFUterraformpulumiCloud formationAWSGCPMicrosoft