https://www.aketchel.com Arthur Ketchel II - Fractional CTO, AI-Augmented Engineering & Strategic Consulting en-us Fri, 05 Jun 2026 16:54:52 GMT https://www.aketchel.com/comics/broken-windows-expand-exponentially https://www.aketchel.com/comics/broken-windows-expand-exponentially Fri, 05 Jun 2026 16:54:52 GMT Comic illustration for Broken Windows Theory: showing how AI-generated broken windows compound exponentially, with cultural transmission at machine speed. https://www.aketchel.com/blog/broken-windows-expand-exponentially https://www.aketchel.com/blog/broken-windows-expand-exponentially Fri, 05 Jun 2026 16:10:44 GMT Broken windows in the AI era don't just invite more broken windows — they train the AI to generate more of them. The decay is exponential, not linear. Here's why and how to fix it. https://www.aketchel.com/comics/parkinsons-law-is-weaponized-with-ai https://www.aketchel.com/comics/parkinsons-law-is-weaponized-with-ai Thu, 04 Jun 2026 11:22:03 GMT A comic illustrating how Parkinson's Law — work expands to fill the time available — becomes weaponized when AI amplifies the cycle, turning efficiency gains into expanded scope and complexity. https://www.aketchel.com/blog/parkinsons-law-is-weaponized-with-ai https://www.aketchel.com/blog/parkinsons-law-is-weaponized-with-ai Thu, 04 Jun 2026 04:00:08 GMT Parkinson's Law says work expands to fill available time. In the AI era, generation is faster but verification grows. The freed-up time fills with other work — and then some. https://www.aketchel.com/comics/distributed-fallacies-still-exist-with-ai https://www.aketchel.com/comics/distributed-fallacies-still-exist-with-ai Mon, 01 Jun 2026 16:21:55 GMT A comic illustrating the 8 fallacies of distributed computing, reimagined for multi-agent AI systems. From unreliable APIs to non-deterministic outputs, these are the lies your AI agents believe. https://www.aketchel.com/blog/distributed-fallacies-still-exist-with-ai https://www.aketchel.com/blog/distributed-fallacies-still-exist-with-ai Mon, 01 Jun 2026 14:38:29 GMT In 1994, Peter Deutsch listed seven assumptions that distributed systems engineers get wrong. Now we're building multi-agent AI systems, and we're making every single one of these mistakes again. https://www.aketchel.com/comics/putts-law-in-ai-era https://www.aketchel.com/comics/putts-law-in-ai-era Fri, 29 May 2026 18:25:47 GMT A comic illustration of Putt's Law in the AI era: the dangerous gap between those who manage AI systems without understanding them and those who understand AI without authority to manage it. When combined with the Dilbert Principle, this creates a death spiral of confidently wrong decisions approved by those who can't evaluate them. https://www.aketchel.com/blog/putts-law-ai-management-org-design https://www.aketchel.com/blog/putts-law-ai-management-org-design Thu, 28 May 2026 17:18:22 GMT Putts Law says technology is dominated by those who understand what they dont manage and those who manage what they dont understand. In the age of AI agents, this gap becomes a dangerous chasm — and the fix is organizational, not technical. https://www.aketchel.com/comics/sturgeons-law-with-agentic-code https://www.aketchel.com/comics/sturgeons-law-with-agentic-code Tue, 26 May 2026 04:12:04 GMT Comic illustrating Sturgeon's Law applied to AI-generated code — 90% is boilerplate, the 10% that matters is where risk lives. https://www.aketchel.com/blog/sturgeons-law-with-agentic-code https://www.aketchel.com/blog/sturgeons-law-with-agentic-code Tue, 26 May 2026 00:37:58 GMT Sturgeon said 90% of everything is crap. For AI-generated code, 90% is boilerplate — not crap, but not where the risk lives. The 10% that matters is where the AI is most likely to be wrong and the consequences are highest. Stop reviewing everything equally. https://www.aketchel.com/comics/ai-training-data-is-a-map-not-territory https://www.aketchel.com/comics/ai-training-data-is-a-map-not-territory Thu, 21 May 2026 20:15:45 GMT Law 50: The AI's training data is a map — compressed, biased, and incomplete. Your requirements are the territory. This comic illustrates why AI-generated code that looks right by training data standards can be wrong for your actual system, and how to navigate by territory instead of map. https://www.aketchel.com/blog/ai-training-data-is-a-map-not-territory https://www.aketchel.com/blog/ai-training-data-is-a-map-not-territory Thu, 21 May 2026 19:53:44 GMT Alfred Korzybski's famous dictum is the single most important principle for understanding AI-generated code. The AI's training data is a map — compressed, biased, incomplete. Your requirements are the territory — specific, current, unique. Here's why the difference matters and how to navigate by territory, not map. https://www.aketchel.com/comics/lindy-effect-with-ai-trends https://www.aketchel.com/comics/lindy-effect-with-ai-trends Wed, 20 May 2026 20:38:37 GMT A comic illustrating the Lindy Effect in software development: why building on time-tested patterns like SQL and HTTP beats chasing the latest AI tooling trend. https://www.aketchel.com/blog/lindy-effect-with-ai-trends https://www.aketchel.com/blog/lindy-effect-with-ai-trends Wed, 20 May 2026 20:09:00 GMT The Lindy Effect says the future life expectancy of a technology is proportional to its current age. For AI-augmented development, this means: don't build your architecture around AI tools that launched last year. Build around patterns that have survived 20 years, then use AI to implement on top of those patterns. https://www.aketchel.com/comics/inversion-ai-prompts-for-better-code https://www.aketchel.com/comics/inversion-ai-prompts-for-better-code Tue, 19 May 2026 01:07:45 GMT A comic illustration of the inversion principle applied to AI code generation: instead of asking how to write good prompts, ask what conditions make AI produce terrible code and eliminate those. https://www.aketchel.com/blog/inversion-ai-prompts-for-better-code https://www.aketchel.com/blog/inversion-ai-prompts-for-better-code Mon, 18 May 2026 18:18:43 GMT Charlie Munger's favorite thinking tool, inversion, is the most powerful prompt design strategy for AI code generation that nobody's using. Stop chasing "good code" and start eliminating the conditions that produce bad code. https://www.aketchel.com/comics/dilbert-principle-in-the-ai-era https://www.aketchel.com/comics/dilbert-principle-in-the-ai-era Fri, 15 May 2026 12:36:36 GMT Law 41: In agentic AI systems, the agents with the least understanding of the domain are given the most autonomous decision-making power, mirroring how the Dilbert Principle promotes the least competent people to management. https://www.aketchel.com/comics/occams-razor-in-agentic-ai https://www.aketchel.com/comics/occams-razor-in-agentic-ai Fri, 15 May 2026 12:30:21 GMT Law 29: When your AI agent's output seems inexplicable, the simplest explanation is usually correct. It's doing exactly what you told it to do, not what you meant. https://www.aketchel.com/blog/dilbert-principle-in-the-ai-era https://www.aketchel.com/blog/dilbert-principle-in-the-ai-era Thu, 14 May 2026 19:28:18 GMT The Dilbert Principle says the least competent people get promoted to management. For AI agents, the least competent agent gets the most autonomy — because it produces the most output. Volume is not competence. But volume is what gets rewarded. https://www.aketchel.com/blog/occams-razor-in-agentic-ai https://www.aketchel.com/blog/occams-razor-in-agentic-ai Thu, 14 May 2026 15:35:45 GMT When AI output is wrong, the simplest explanation is a vague prompt, not model failure. 90% of wrong AI output comes from prompts — but teams spend 90% of their debugging time on the AI. Here's how to fix that. https://www.aketchel.com/comics/sunk-cost-fallacy-with-agentic-ai https://www.aketchel.com/comics/sunk-cost-fallacy-with-agentic-ai Tue, 12 May 2026 14:06:36 GMT Law 28: The sunk cost fallacy takes a new form with AI-generated code. The AI's time is free, so stop defending bad AI code just because you've invested in understanding it. https://www.aketchel.com/blog/sunk-cost-fallacy-with-agentic-ai https://www.aketchel.com/blog/sunk-cost-fallacy-with-agentic-ai Tue, 12 May 2026 13:49:31 GMT The sunk cost fallacy takes a new form with AI-generated code. The AI's generation cost is nominal, but the human review, integration, and debugging time is real — and already spent. Don't keep bad AI code because you've invested in understanding it. Throw it away and regenerate. https://www.aketchel.com/comics/lehmans-laws-of-software-development https://www.aketchel.com/comics/lehmans-laws-of-software-development Mon, 11 May 2026 14:38:42 GMT Comic illustration for Law 42: Lehman's Laws. AI-generated code that isn't actively understood declines faster than code that isn't actively maintained. Understanding debt is combinatorial, not linear. https://www.aketchel.com/blog/lehmans-laws-of-software-development https://www.aketchel.com/blog/lehmans-laws-of-software-development Mon, 11 May 2026 14:05:54 GMT Lehman's Laws say systems must evolve or die. AI introduces a darker corollary: AI-generated code that isn't actively understood declines faster than code that isn't actively maintained. Understanding debt is combinatorial, not linear. https://www.aketchel.com/comics/ai-technical-debt https://www.aketchel.com/comics/ai-technical-debt Fri, 08 May 2026 19:17:48 GMT Comic illustration for AI Technical Debt. Explores how AI-generated code creates new forms of technical debt, Spec Debt, Understanding Debt, Validation Debt, Prompt Debt and why it compounds faster than traditional tech debt. https://www.aketchel.com/blog/ai-technical-debt https://www.aketchel.com/blog/ai-technical-debt Fri, 08 May 2026 18:33:39 GMT Technical debt isn't a metaphor. In the AI era, it's a loan with compounding interest, and your AI agent is making the minimum payments for you. AI has introduced new, insidious forms of debt including Spec Debt, Understanding Debt, Validation Debt, Prompt Debt and more. https://www.aketchel.com/comics/dunning-kruger-ai-confidence-amplifier https://www.aketchel.com/comics/dunning-kruger-ai-confidence-amplifier Thu, 07 May 2026 09:15:00 GMT A comic exploring the intersection of the Dunning-Kruger effect and generative AI. In the age of agentic AI, low competence combined with AI tools leads to dangerous overconfidence, while true experts use AI as a tool for leverage, not a replacement for judgment. https://www.aketchel.com/blog/dunning-kruger-ai-confidence-amplifier https://www.aketchel.com/blog/dunning-kruger-ai-confidence-amplifier Thu, 07 May 2026 08:35:06 GMT The junior who used AI to build microservices in 10 minutes is 3x more confident than the senior who spent 3 hours reviewing the output. Guess who the organization thinks is more productive? https://www.aketchel.com/blog/pareto-principle-ai-prompts-and-bug-distribution https://www.aketchel.com/blog/pareto-principle-ai-prompts-and-bug-distribution Tue, 05 May 2026 19:44:00 GMT The Pareto Principle is more extreme in the age of AI. 80% of your AI-generated bugs come from just 20% of your prompts. Learn how to identify the critical 20% and bridge the gap between vague instructions and secure, functional code. https://www.aketchel.com/comics/pareto-principle-ai-prompts-and-bug-distribution https://www.aketchel.com/comics/pareto-principle-ai-prompts-and-bug-distribution Tue, 05 May 2026 16:56:08 GMT A comic exploring how the Pareto Principle is more extreme in the age of AI. 80% of your AI-generated bugs come from just 20% of your prompts. Learn how to identify the critical 20% and bridge the gap between vague instructions and secure, functional code. https://www.aketchel.com/comics/brooks-law-in-the-age-of-agentic-ai https://www.aketchel.com/comics/brooks-law-in-the-age-of-agentic-ai Tue, 05 May 2026 01:15:55 GMT A comic illustrating Brooks' Law - adding manpower to a late software project makes it later - and how agentic AI changes the dynamics of team scaling and communication overhead. https://www.aketchel.com/comics/pesticide-paradox-ai-tests-catch-ai-bugs https://www.aketchel.com/comics/pesticide-paradox-ai-tests-catch-ai-bugs Tue, 05 May 2026 00:56:08 GMT A comic exploring how the Pesticide Paradox compounds dangerously in AI-generated code - when the same AI writes both the code and the tests, they share the same blind spots. https://www.aketchel.com/blog/pesticide-paradox-ai-generated-code-and-testing-traps https://www.aketchel.com/blog/pesticide-paradox-ai-generated-code-and-testing-traps Mon, 04 May 2026 14:40:00 GMT The Pesticide Paradox warns that AI-generated tests often share the same blind spots as AI-generated code. Break the circular confidence trap by using multi-model validation, property-based testing, and human-led adversarial review. https://www.aketchel.com/blog/brooks-law-agentic-ai https://www.aketchel.com/blog/brooks-law-agentic-ai Mon, 27 Apr 2026 02:23:26 GMT Brooks’ Law is not obsolete; it has been reframed. While agentic AI reduces human coordination costs, it intensifies the need for architectural clarity. In the age of AI, scale without structure is just faster failure. https://www.aketchel.com/blog/law-of-specialization-automation https://www.aketchel.com/blog/law-of-specialization-automation Tue, 16 Feb 2016 01:47:26 GMT In my post entitled **Focus your Energy and Launch this Blog** I referred to a question that was once asked of me. The question **“If you had to boil down your professional skill sets and/or professional value into a single word… what word would you choose?”** is one that we evaluated in light of **specialization**. Specialization allows you to focus your energy and identify what sets you apart from everyone else. This singular focus will allow you to perfect your skills in your area of choice and really begin to master your specific trade. It is often difficult for someone to choose such a specialty because we as humans have many interests and skills, but choosing a single specialty will allow you the focus necessary to become the expert in that area. It is often said, especially in the age of technology and highly available information, that with enough time you can become an expert in anything. It is my belief that this is true and I am a firm believer in lifelong learning; however, it is also said that it takes roughly 10,000 hours of practice on a particular skill to become a master. Given that there are 8,760 hours in a year, it would take you a full year plus 52 days to master a particular skill. If you are getting the recommended amount of sleep (does anyone do that?) of 8 hours a day, then you would spend 2920 hours of that time sleeping, which will require another 122 days to compensate. Therefore, you will realistically require at least 1.5 to 2 years to master any specific skill, if you spend every waking moment perfecting that skill. In looking at these numbers, it becomes clear why it would be extremely difficult for your average person to truly master a range of skills in their lifetime, since it will be extremely difficult for them to dedicated that much time to refining multiple skills, especially if there is any balance of life in the equation to add additional realism to the timeline. After much consideration of these facts, I decided on a word that I believe is a good depiction of what I have identified as a particular niche of interest for me and have worked to achieve so far in my career. I derived my specialty through significant effort and study and defining this specialization is not something that has come easily to me, as I expect that it won’t for others, and required that I consider not only the range of projects throughout my professional career, but also in studying the nature of business and which organizations and products have been successful in the market. I have spent the time to break those projects, products and ultimately the businesses that support them down to look at the fundamental aspects of each one and only in doing so was I able to identify the root of the underlying need and use this to define my specialization. Although these projects have spanned technologies, organizations and industries, the concepts defining their success were similar. It is this fact that excites me so much to launch this blog and offer up the insights that I have been able to gather in my career in hopes that it facilitates some level of communication and sparks thoughts and innovations in others. It is an extremely powerful idea to consider the most generic of concepts and evaluate your vision through such a lens. My interests and passions have always straddled both areas of business and technology and it has been through this lens that I was able to view the underlying factor linking my two interests to my ultimate specialization. For me, this word is **“Automation”**. The dictionary definition of **“Automation”** describes the word as **“the use of various control systems in a system of manufacturing or other production process with minimal or reduced human intervention.”**. It is clear that achieving the desired levels of automation have occurred using various methods in various industries and various platforms throughout history. This is not a new concept; however, I believe that it is truly the underlying principle that when properly applied has led to a number of significant successes throughout history. **“Automation”** as a core principle serves to enhance productivity and efficiency with the major benefits being the ability to save labor and improve quality, accuracy and precision. These benefits ultimately lead to the business goals of achieving return on investment and competitive advantages; however, it is important to understand how automation can bring value and how to apply Automation across the business coupled with the use of technology to really make an impact. Throughout this blog we will explore how to leverage automation to enhance your business and how the use of technology can facilitate these areas. It is arguable that this principle is what has fostered the rapid implementation of technology across enterprises throughout the world; however, it is important to note that organizations still continue to struggle with being able to efficiently use this technology in which they have invested millions and that is exactly the challenge we seek to resolve. The need to continuously upgrade to garner the added benefits of technological advancements enhances this challenge while taking focus away from the businesses core challenge of meeting the continued changing demands of their business. The solution for this challenge requires a return to the fundamental mindset of **“Automation”**. Although throwing technology itself at the job may solve certain challenges, as discussed previously, it will also introduce others. In order to successfully implement the concept of **“Automation”**, the enterprise will need to buy into the value as part of the organizational culture and the concept will require that a choice be made to optimize each component within the organization to reduce the amount of effort required to achieve optimal output with minimal required input. It is in this area that I have spent my career working with organizations to implement this concept and an area that I expect will continue to become more important in the years to come. The successful implementation of these techniques will significantly enhance an organizations competitive advantage and help build a moat in their market as the organizations begin to spend more time focusing on the core challenges instead of having to become experts in every area required to keep operating. I’m interested in my readers thoughts and ideas around this concept and I welcome you to join in the discussion and post comments and questions. I have a number of specific topics that I wish to discuss as we consider the more generic topic of Automation, but I hope to get feedback from others as well as you think through how these concepts can affect your business and your personal life. As previously mentioned, evaluating the challenges you face through the lens of specialization and simplification will not only help in the exercise of identifying an area of personal focus, but in being able to really step back and evaluate any of the challenges you may face and may just lead you to the answer you are seeking. Although this is only the first step in a longer journey, it is important that you begin by evaluating and considering your areas of specialization in hopes that it will also lead you to focus and identify areas outside of your area of mastery. For these areas, if they are truly necessary, I’m happy to discuss how automation can help get the job done such that you may focus on what is important. https://www.aketchel.com/blog/power-of-specialization https://www.aketchel.com/blog/power-of-specialization Fri, 12 Feb 2016 01:47:26 GMT I once had a friend ask me… “If you had to boil down your professional skill sets into a single word… what word would you choose?”. Obviously choosing a single word to identify the value that you bring to your profession is a difficult exercise; however, the exercise identifies a set of valuable advice at the core of the question. Our lives become more and more complex each day as society continues to advance. The complexities of our lives continue to allow us to experience more that life has to offer and we have the unique opportunity, today more than ever before, to specifically align our passions and interests with our talents to define a profession that allows us to provide value to society. At least, that has always been the nature of society and this perspective has been part of the traits that have lead to success. Instead of each person focusing on their own needs and trying to achieve everything required to survive by themselves, society established the practice of specialization of labor by organizing into roles such as farmers, bakers, blacksmiths, craftsmen, scribes, etc. that allowed each individual member to bring unique value to society as a whole while garnering benefit from other members to meet each individual’s needs. This specialization lead first to the barter system and later to the system of commerce and capitalism that we now enjoy. The challenge in modern society is in choosing. Specifically, the choice of how to focus your time when a number of competing priorities and opportunities compete to make the best use of our energy. Additionally, once you have chosen a specialization to focus your energy, you must then determine how to monetize that opportunity to allow you to achieve survival and continue your efforts. This often becomes a significant barrier and although the use of modern technology has provided us efficiency in many ways and has led to more freedom than ever before to allow a person the ability to explore multiple opportunities to find their ultimate passion, this freedom is the cause of significant inefficiency due to the paradox of choice. Although we still follow the practices of separation of labor, which helps to define specific roles, it is important that we as members of society make a conscious effort to more specifically refine how we will focus our energy during any given time period and make that use of energy have a positive impact. Of course, there will always be barriers of entry and challenges to overcome to achieve particular goals; however, that is exactly what we seek in this effort of continued growth, as overcoming those challenges are how we continue to excel. The ability to overcome an obstacle and share the knowledge of how we achieved the result is our contribution to society. As long as each member of society continues to make positive contributions, society as a whole will continue to thrive. My goal with the launch of this blog is to do just that in hopes that this blog will serve as a medium to share and grow together. The topics you will find here are specific areas of interest of mine and will obviously follow my focus of energy. I will follow-up with additional posts to explore my background and hopefully provide insight into my specific areas of interest in hopes that this blog will provide value to others. Of course, I cannot guarantee that the content on this blog will be anything more than just my thoughts, opinions and interests, but I hope that you will join me in this journey and share your own journey along the way. I am open to comments and feedback and look forward to the journey ahead.