2026-09-02
How Should You Route Requests Across Multiple AI Models?
When an AI system uses several models, should requests be assigned by explicit rules, predictive routing, cascades, or fallback paths?
Read →9 min
Artificial Intelligence
AI where it changes the business
AI matters when it changes the economics or operations of a real business: the work people do, the decisions they make, the information they can use, or the time a process consumes. These essays begin with that consequence, then examine whether AI is the right intervention.
Browse by topic
2026(29)
2026-09-02
With AI, a Refactor I Wouldn't Bother With Took One Hour
PeşinTaksit was a neglected side project that still required four containers. With AI, I moved it to Cloudflare Workers in about an hour.
Read →3 min
2026-09-01
When Is One AI Model Enough?
When should an AI system begin with one model, and what improvement in quality, cost, speed, or risk would justify adding another?
Read →8 min
2026-09-01
From Asking Questions to Delegating Work: What the Codex Study Shows
OpenAI's Codex research shows how AI use is shifting from questions to delegated work—and what companies should measure before calling it value.
Read →11 min
2026-08-31
Who Captures the Productivity Gains from AI?
AI can make employees and service providers more productive without leaving them with the economic gain. An analysis through salaries, hourly billing, and competition.
Read →12 min
2026-08-31
Is Your Data Ready for AI? Start With the Decision It Must Support
Before choosing a model, test whether the information needed for a specific decision is usable, current, authoritative, and worth preparing.
Read →11 min
2026-08-31
Is Your AI System Actually Delivering Business Results?
A good model response, a completed task, and a measurable business result are three different outcomes. Here is how to evaluate each one.
Read →7 min
2026-08-30
How Much Authority Should You Give an AI System?
When should an AI system prepare a recommendation, and when should it be allowed to execute a real transaction? A practical boundary for authority, approval, and verification.
Read →6 min
2026-08-29
When Does Enterprise AI Actually Need RAG?
When should enterprise AI use RAG, direct context, a database query, or an API? A practical way to choose the right source for each question.
Read →6 min
2026-08-28
When Do You Actually Need an AI Agent?
Customer support, lead routing, REDAR, and incident diagnosis show the practical difference between software-defined workflows and AI agents.
Read →6 min
2026-08-28
Does Enterprise AI Really Need Fine-Tuning?
When is providing a document enough, when does RAG help, and when does fine-tuning actually make sense? The decision starts with information, rules, and patterns.
Read →7 min
2026-08-28
When Does Process Automation Actually Need AI?
Lead routing, financial disclosures, and bank integrations show where AI adds value—and where rules or conventional software are the better choice.
Read →5 min
2026-08-27
Start With the Business Problem, Not the AI Model
Why model selection should not be the first decision in an AI project, and what must be defined about the business outcome, action boundary, and evidence first.
Read →5 min
2026-08-27
Preserving API Quality in AI-Assisted Development
A short guide to preserving resource models, HTTP behaviour, error handling, authorization, security, and contract decisions in prompts, reviews, and delivery checks when building REST APIs with AI.
Read →8 min
2026-08-24
Managing Technology and Transforming the Business Are Not the Same Thing
Why keeping technology running and using it to transform the business should not be treated as the same responsibility.
Read →5 min
2026-08-24
Keeping Customers Happy Isn’t Enough. You Have to Follow Up.
A positive customer experience does not automatically turn into a review or recommendation. A small observation about follow-up, timing, and how AI changes the cost of building useful software.
Read →7 min
2026-08-23
Do AI Visibility Tools Really Work? What They Actually Measure
Do AI visibility tools really work? They track selected prompt outputs and trends, not real-user brand visibility or commercial impact on their own.
Read →9 min
2026-08-23
llms.txt Was Never the Point
A closer look at the AI Visibility hype around llms.txt, the information architecture experiment I built, and the difference between a technical hypothesis and a marketing promise.
Read →8 min
2026-08-22
Do You Know How Dependent Your Company Is on AI?
AI is not only accelerating work people already did. It is creating operational capacity companies could never afford to have — and that capacity needs a continuity plan before it becomes structural.
Read →11 min
2026-08-20
AI Wasn’t Always This Good. We Just Got Used to It.
AI is rapidly changing team capacity and software economics. That calls for a fair view of the past—and new expectations for today.
Read →5 min
2026-08-18
AI Made Code Cheap. Verification Is Still Expensive.
AI can produce an implementation in hours. If its author cannot explain the decisions behind it, the real engineering work has not disappeared — it has moved to the reviewer.
Read →9 min
2026-06-28
The Threshold Collapsed: What ProductLog Taught Me About Building in Public
In April 2026, Arvid Kahl revisited the argument that had helped define build in public for many founders. He had once seen a practical threshold: companies became more guarded when they reached roughly $20,000–$30,000 in monthly recurring revenue. In his view, agentic coding changed that calculation. The old threshold had, as he put it, “effectively collapsed to zero.”
Read →4 min
2026-06-28
Build in Public in the AI Era: What to Share and What to Keep Private
Building in public does not require choosing between silence and publishing an operating manual for your competitors. The useful question is narrower: for each thing you could share, will the feedback, trust, or connection it creates be worth the risk of making it public?
Read →5 min
2026-06-28
AI Visibility Dropped Before Search: The Google Data That Changed My Theory
Google's Generative AI report made me revise my live-fetch theory. The observed lead over search held, but the data does not prove a separate AI index.
Read →8 min
2026-06-27
The Era of the "Previous Vibe Coder" Begins: The Invisibility of Clean Code and the Technical Debt Bill of AI
An analysis of the false speed illusion of AI coding, DRY violations, the devaluation of invisible quality using Akerlof's Market for Lemons, and the upcoming crisis of the 'Previous Vibe Coder'.
Read →8 min
2026-06-25
AI Is Reducing Hiring Without Layoffs
Layoffs are counted. The role a team never opens is not. A CTO and founder's view of how AI can change the hiring decision before it appears in the data.
Read →9 min
2026-06-23
Comprehension Debt: The Bill Comes Due Alone
I ship with AI coding agents daily. The hidden risk is the gap between what the system can do and what I can still explain, debug, and safely change.
Read →9 min
2026-06-22
From WordPress to Nuxt: Building an AI-Powered Content Pipeline
How I moved from WordPress to a Nuxt and Cloudflare Pages setup, using AI to reduce publishing friction while retaining editorial and technical review.
Read →4 min
2026-06-22
The AI Visibility Illusion: What Bing's Citation Share Data Actually Reveals
AI Visibility tools model a selected set of prompts. Bing's first-party Citation Share data gave me a way to compare that model with what I was seeing across the sites I manage.
Read →9 min
2025(4)
2025-11-20
The End of Coding or a New Renaissance? The Invisible Crisis of AI
AI may reduce the routine work through which junior developers learned. The more consequential question is how organisations will preserve practice, feedback, and the path to senior capability.
Read →6 min
2025-11-14
Raising Children in the Age of Artificial Intelligence
On social media I usually share technical or work-related topics. But behind that image is also a dad with a 9-year-old son.
Read →10 min
2025-07-13
What We Learned Building a Healthcare Chatbot for International Patients
A year of running VaniBot taught us why healthcare chatbots depend on human handoff, internal APIs, multichannel integration, and continuous ownership.
Read →5 min
2025-06-05
Building camiler.org: A Programmatic SEO Experiment with Google Maps and OpenAI
How I built camiler.org with Diyanet data, Google Maps, OpenAI, Fastify, and Nuxt, and why the automated SEO experiment became a human review workflow.
Read →6 min