# Swizec Teller A geek with a hat This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## What B2B SaaS taught me about having kids DevFeed: [What B2B SaaS taught me about having kids]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-09-05T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Software as a service](), [Software]() Tags: [b2b](), [saas](), [software]() ### AI overview The author compares lessons from running a B2B SaaS business with the realities of early parenthood. The article emphasizes maintaining sufficient resources, filtering advice for personal circumstances, prioritizing manageable improvements, and accepting that good enough can be better than perfection. ### Source excerpt I've been a dad for 3 months now so obviously I have it all figured out and there's nothing new to learn. My theories are rock solid and definitely apply to everyone. ## Focus and followthrough are the moat DevFeed: [Focus and followthrough are the moat]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-08-27T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Development](), [SRE]() Tags: [development](), [focus](), [production](), [projects](), [sre](), [team]() ### AI overview An opinion piece about how a small startup team uses focus, sprint planning, prioritization, and regular stakeholder communication to finish work despite having more requests than capacity. ### Source excerpt Because starting is easy and finishing is hard. ## Stop burning tokens on code review DevFeed: [Stop burning tokens on code review]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-08-15T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Code review](), [Claude Code](), [Code](), [GitHub Actions](), [Architecture & Design](), [cursor]() Tags: [agents](), [claude-code](), [code](), [code-review](), [cursor](), [engineering](), [github-actions](), [security]() ### AI overview The author argues that custom linters can provide fast, inexpensive guardrails and feedback loops for both human and agent-written code. They describe mixed experiences with Cursor's BugBot and Claude code review skills, especially the high noise and review burden produced when those skills run in GitHub Actions. ### Source excerpt I've been experimenting with different approaches to build a system where humans and agents can ship fast safely. Think I've found something that works - custom linters. ## AI-assisted design-system rewrites are fast to generate but slow to refine DevFeed: [AI-assisted design-system rewrites are fast to generate but slow to refine]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-08-08T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Artificial Intelligence](), [Large Language Model](), [Claude](), [Design system](), [stacked pull requests](), [Bootstrap](), [Tailwind CSS]() Tags: [ai](), [claude](), [design-system](), [github](), [llms](), [stacked-pull-requests](), [tailwind]() ### AI overview The author describes using AI and Claude to migrate a design system from Bootstrap to Tailwind, clean up HTML and JavaScript, and rewrite functional code. The article argues that LLMs handle mechanical translation well but struggle with conceptual design decisions and fine UI details, making cleanup substantially more time-consuming than code generation. ### Source excerpt Spaceship first, cleanup second. Start with the ending then work your way back to something useful. ## Agents change the math on big bang migrations DevFeed: [Agents change the math on big bang migrations]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-08-01T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [migration](), [Tailwind CSS](), [Bootstrap](), [Design system](), [Tech Debt](), [ui](), [React]() Tags: [agents](), [bootstrap](), [design-system](), [migrations](), [react](), [tailwind](), [tech-debt]() ### AI overview An opinion article argues that AI agents may change how teams approach large software migrations. It describes a one-step migration of hundreds of pages from Bootstrap to Tailwind in a 22,000-line pull request, which the author says fixed 13 small bugs without materially disrupting the company. ### Source excerpt I just did a team's worth of work in 5 days. Honestly I am shook to the core of my 20 YoE ## How to be useful as a software architect DevFeed: [How to be useful as a software architect]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-07-25T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [engineering-culture](), [Code review](), [Automation](), [Software]() Tags: [architecture](), [automation](), [code-review](), [culture](), [software](), [system-design]() ### AI overview The article argues that useful software architects improve the future of a codebase by building systems that let teams move quickly and safely. It emphasizes culture, feedback loops, automation, code review, system design review, testing, and shepherding work into production without becoming a bottleneck. ### Source excerpt Software architects have a weird role: Think about the code _tomorrow_. Today's code is what it is, how do we make it good tomorrow? Without slowing down, losing business, or falling off a cliff. ## swizec.com is now on TimberJS, a new React framework DevFeed: [swizec.com is now on TimberJS, a new React framework]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-07-21T00:00:00Z Content type: article Language: en Sources: [Swizec Teller]() Topics: [React](), [Framework](), [gatsby](), [Web Development]() Tags: [gatsby](), [javascript](), [react](), [render](), [server]() ### AI overview The author describes moving swizec.com from Gatsby to TimberJS, a new React framework. They report that deployment time fell from 45 minutes to 4 minutes and describe server-rendered navigation, interactive elements, and server actions. ### Source excerpt swizec.com is 20 years old and has gone through many rewrites. Here's the latest ## Why Project Management Is a Critical Skill for Experienced Engineers DevFeed: [Why Project Management Is a Critical Skill for Experienced Engineers]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-07-11T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Development](), [Software](), [Users]() Tags: [development](), [project-management](), [software](), [users]() ### AI overview The article argues that experienced engineers often struggle when given ownership of a project defined by goals rather than preassigned tasks. It presents breaking complex problems into actionable work, clarifying the definition of done, and understanding user impact as critical engineering skills. ### Source excerpt In the last few months every engineer I hired stumbled in exactly the same way: Project management. They learned quickly, but that first big project has become a rite of passage. ## Theory of constraints, AI, and code review DevFeed: [Theory of constraints, AI, and code review]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-07-06T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Artificial Intelligence](), [Code review](), [engineering-culture](), [Pull Request](), [Code]() Tags: [ai](), [code](), [code-review](), [productivity](), [pull-requests](), [review]() ### AI overview The article argues that producing code faster with AI does not necessarily increase company-wide delivery because overall throughput is constrained by slower parts of the development system, especially code review and release processes. ### Source excerpt Feedback from an expert is how you learn taste. Your code doesn't matter until it's shipped. ## Using vibes to rewrite my blog DevFeed: [Using vibes to rewrite my blog]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-06-27T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [gatsby](), [React](), [MDX](), [Syntax Highlighting](), [Claude](), [wasm](), [codex]() Tags: [article](), [blog](), [claude](), [codex](), [gatsby](), [mdx](), [react](), [syntax-highlighting]() ### AI overview The author describes rewriting the swizec.com blog after Gatsby became difficult to maintain and slow to deploy. They experiment with TimberJS, a React Server Components-based framework, to render MDX with file-based routing, metadata, embeds, image generation, and server-side syntax highlighting. The rewrite reduced reported deployment time from more than 45 minutes to about 3 minutes, while the author emphasizes that effective vibe coding still requires technical knowledge. ### Source excerpt The death of engineering has been greatly exaggerated. You still need to know what you're doing. ## You get what you hire DevFeed: [You get what you hire]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-06-23T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [systems](), [Front end](), [Back end](), [Python](), [React](), [dashboards]() Tags: [backend](), [dashboard](), [distributed-systems](), [frontend](), [hiring](), [htmx](), [python](), [react]() ### AI overview The article argues that engineers' backgrounds and preferred tools strongly influence the systems they build. It compares two working logistics dashboards: a backend-heavy Python system with form submissions and page reloads, and a frontend-oriented React system with self-contained components, multiple pages, CRUD operations, and a themable design system. ### Source excerpt Hiring engineers is like deciding what kind of system you'll have. This seems obvious but surprised me to see first-hand. ## Code is the easy part, or how we refactored half the business to fix a janky script DevFeed: [Code is the easy part, or how we refactored half the business to fix a janky script]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-06-17T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Code](), [Python](), [Finance]() Tags: [automated](), [billing](), [code](), [pretty](), [python](), [stress]() ### AI overview A startup engineering war story about refactoring a Python-based billing and invoicing system that had accumulated undocumented business rules, manual workarounds, and operational stress. The team automated the end-to-end invoicing process; a test run took two days instead of a week, although the work was not yet complete. ### Source excerpt This is a classic startup war story: Someone had built a leaky rowboat. It worked! Then we started flying the rowboat as a Cessna. What we needed was an aircraft carrier. ## Good software gets out of the way DevFeed: [Good software gets out of the way]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-06-04T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Software](), [coding]() Tags: [metrics](), [product-engineering](), [product-sense](), [software](), [users]() ### AI overview The article argues that good software should be nearly invisible because it helps users complete their work. It recommends focusing on user outcomes, observing workflows, understanding the domain, defining success early, and prioritizing product sense in engineering. ### Source excerpt Explore the art of invisible software and user-centric workflow design with Swizec and Kent C Dodds. Uncover what separates a good engineer from a great one in this insightful dialogue. ## User outcomes, workflow design, and biotech software DevFeed: [User outcomes, workflow design, and biotech software]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-06-03T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Software](), [Users](), [Learning]() Tags: [developer](), [product-engineering](), [software](), [users](), [workflow]() ### AI overview An interview about product engineering for software used by real businesses and non-developer users, covering unfamiliar domains, workflow observation, friction, feature placement, and outcome-focused success metrics. ### Source excerpt Kent C. Dodds talks with Swizec Teller about product engineering for software that serves real businesses and non-developer users. ## What Entry-Level Software Engineering Looks Like in the AI Age DevFeed: [What Entry-Level Software Engineering Looks Like in the AI Age]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-05-28T00:00:00Z Content type: article Language: en Sources: [Swizec Teller]() Topics: [Software Engineering](), [Artificial Intelligence](), [proof of work](), [systems]() Tags: [ai](), [engineering](), [interviews](), [proof-of-work](), [software-engineering](), [systems](), [talks]() ### AI overview Clara Shih, Chip Huyen, and Swizec Teller discuss how AI-assisted coding is affecting entry-level software engineering, including interviews, proof-of-work projects, systems thinking, and side projects. ### Source excerpt Clara Shih talks with Chip Huyen and Swizec Teller about what entry-level software engineering looks like in the AI age. ## Relentlessly focus on the outcomes DevFeed: [Relentlessly focus on the outcomes]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-05-26T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Users](), [Artificial Intelligence](), [Code]() Tags: [ai](), [code](), [users]() ### AI overview This commentary argues that teams should measure success by the outcomes they create for users rather than by activity such as pull requests, documents, tasks, and meetings. It says AI can increase code and output while making it easier to lose focus and leave work unfinished. ### Source excerpt AI makes it easy to focus inwards and feel impressed with all the churn and motion and output you're producing. But your users don't care. ## Engineering levels aren't about skill anymore DevFeed: [Engineering levels aren't about skill anymore]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-05-19T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Job](), [systems](), [Code]() Tags: [code](), [engineering](), [hiring](), [interviews](), [system-design]() ### AI overview This commentary argues that engineering levels have shifted from a straightforward progression of coding skill to role expectations that vary by level. It describes responsibilities from entry-level engineering through staff-level work and says hiring decisions increasingly assess candidates through interviews, system design, and behavioral evaluation, with coding treated as a baseline. ### Source excerpt Traditional engineering levels were about skill progression. I think this has changed. Everyone now is a mini staff engineer. ## Using bots to implement fixes from pull request review comments DevFeed: [Using bots to implement fixes from pull request review comments]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-05-13T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Code review](), [Pull Request](), [Artificial Intelligence](), [engineering-culture]() Tags: [ai](), [api-design](), [code-review](), [code-reviews](), [knowledge-sharing](), [pull-requests]() ### AI overview The article argues that bots can implement fixes requested in pull request review comments, allowing developers to review the changes and merge them when appropriate. It presents this as useful for small fixes and notes that code owners may need to re-verify the result. ### Source excerpt You know what's better than a code review comment? A fix. ## Why AI Still Requires Human Delegation and Ownership DevFeed: [Why AI Still Requires Human Delegation and Ownership]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-05-09T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Artificial Intelligence](), [Job](), [context](), [round robin]() Tags: [agents](), [ai](), [delegation](), [job](), [ownership](), [tasks]() ### AI overview The article argues that AI can provide more hands to do work but does not eliminate the human responsibility of managing context, tracking tasks, delegating, and taking ownership of problems and domains. ### Source excerpt On AI, mental load, and delegation ## Using AI at work on This Month in React DevFeed: [Using AI at work on This Month in React]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-05-07T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Artificial Intelligence](), [ai-coding](), [coding](), [Software Engineering](), [Code review](), [Testing]() Tags: [ai](), [ai-at-work](), [ai-coding](), [code-review](), [coding](), [testing](), [workflows]() ### AI overview An interview with Mark Erikson and Carl Vitullo about using AI coding tools at work, including agent workflows, hands-on involvement, background agents, code review, testing, and the continuing role of software engineering judgment. ### Source excerpt Mark Erikson and Carl Vitullo talk with Swizec Teller about using AI coding tools at work. ## Building an Agent with RAG, Tool Calling, Memory, and Feedback DevFeed: [Building an Agent with RAG, Tool Calling, Memory, and Feedback]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-05-05T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Retrieval-Augmented Generation](), [Large Language Model](), [Tool]() Tags: [agent](), [memory](), [rag]() ### AI overview The author outlines four components for an agent used in internal tooling and processes: retrieval-augmented generation (RAG), tool calling, memory, and a feedback loop. The article describes using tools to retrieve external data and returning results to an LLM's context, while noting that long prompts may require memory-management techniques. ### Source excerpt I've been digging into agents for internal tooling and processes. Looking for how to scale without drowning in manual work. Here's what I found ## Using evals and user data to measure AI product improvements DevFeed: [Using evals and user data to measure AI product improvements]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-05-01T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Artificial Intelligence](), [dataset](), [test]() Tags: [ai](), [dataset](), [evals](), [test]() ### AI overview The article argues that AI developers should use evaluations to measure whether changes improve a system. It recommends building test datasets from user behavior, testing models, prompts, and tools, avoiding overfitting, incorporating feedback from real-world use, and tracking human intervention in failure cases. ### Source excerpt When candidates put AI on their resume, the key thing I try to find out is whether they used evals. How did you measure making improvements? ## AI now writes 97% of my code. Here's what I learned DevFeed: [AI now writes 97% of my code. Here's what I learned]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-04-25T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Artificial Intelligence](), [cursor](), [Vibe coding](), [prompt](), [Code]() Tags: [ai](), [code](), [coding](), [cursor](), [library](), [prompt](), [react]() ### AI overview The author says Cursor now generates about 97% of their production code and describes how background agents changed their workflow. They emphasize providing context, desired outcomes, implementation guidance, and codebase details through Slack, then iterating on the results. ### Source excerpt Forget the bullshit tips and tricks, just talk to it. ## Why Reading Research Papers Can Accelerate Learning in Agentic Development DevFeed: [Why Reading Research Papers Can Accelerate Learning in Agentic Development]() Original publisher: [Read original article]() Author: hi@swizec.com (Swizec Teller) Published: 2026-04-20T00:00:00Z Content type: opinion Language: en Sources: [Swizec Teller]() Topics: [Agentic development](), [Development]() Tags: [agentic-development](), [agents](), [paper](), [reading](), [research]() ### AI overview The author argues that research papers are an efficient source of distilled knowledge for learning technical subjects. They describe using papers to build a mental framework for agentic development, while noting that expert conversations and books are better suited to very recent information or broad historical context. ### Source excerpt Read more papers. You can learn the latest and greatest in your field in one chill afternoon. [Next page]()