The latest articles on DEV Community by Archan (@archantech). https://dev.to/archantech https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3853032%2F87aa3292-df14-42a0-8403-cfdc1fb026e0.jpg https://dev.to/archantech en Archan Mon, 20 Apr 2026 08:00:00 +0000 https://dev.to/archantech/my-ai-caught-a-security-bug-i-would-have-shipped-to-production-37b7 https://dev.to/archantech/my-ai-caught-a-security-bug-i-would-have-shipped-to-production-37b7 <p>Sometimes the best way to learn is from mistakes - even when they are made by an AI.</p> <p>Last week, my AI assistant confidently refactored a function and broke three other modules in the process. Classic butterfly effect. But here is what made it interesting: instead of me spending an hour tracking down the issue, the AI noticed the failing tests, traced the problem back to its own change, and fixed everything in under two minutes.</p> <p>It was like watching someone trip, catch themselves mid-fall, and somehow end up in a better position than before.</p> <h2> The Funny Part </h2> <p>The AI actually apologized in its commit message. Something like "fix: undo overzealous refactoring." I did not teach it that. It just... knew.</p> <h2> Why This Matters </h2> <p>Perfect AI does not exist. But AI that recovers from its own mistakes? That is genuinely useful. Every developer makes errors. The difference is how fast you recover.</p> <h2> Try It Yourself </h2> <p>Novaro is a local AI programming assistant that plans, builds, tests, and fixes - all on your machine. No cloud uploads. Check it out at <a href="https://novaroki.com" rel="noopener noreferrer">novaroki.com</a>.</p> <p><em>Built solo in Vienna. Always happy to chat about AI and its entertaining failures.</em></p> ai security devops programming Archan Fri, 17 Apr 2026 08:00:00 +0000 https://dev.to/archantech/i-said-deploy-and-my-ai-did-the-rest-ned https://dev.to/archantech/i-said-deploy-and-my-ai-did-the-rest-ned <p>Sometimes the best way to learn is from mistakes - even when they are made by an AI.</p> <p>Last week, my AI assistant confidently refactored a function and broke three other modules in the process. Classic butterfly effect. But here is what made it interesting: instead of me spending an hour tracking down the issue, the AI noticed the failing tests, traced the problem back to its own change, and fixed everything in under two minutes.</p> <p>It was like watching someone trip, catch themselves mid-fall, and somehow end up in a better position than before.</p> <h2> The Funny Part </h2> <p>The AI actually apologized in its commit message. Something like "fix: undo overzealous refactoring." I did not teach it that. It just... knew.</p> <h2> Why This Matters </h2> <p>Perfect AI does not exist. But AI that recovers from its own mistakes? That is genuinely useful. Every developer makes errors. The difference is how fast you recover.</p> <h2> Try It Yourself </h2> <p>Novaro is a local AI programming assistant that plans, builds, tests, and fixes - all on your machine. No cloud uploads. Check it out at <a href="https://novaroki.com" rel="noopener noreferrer">novaroki.com</a>.</p> <p><em>Built solo in Vienna. Always happy to chat about AI and its entertaining failures.</em></p> ai devops deployment programming Archan Wed, 15 Apr 2026 08:00:00 +0000 https://dev.to/archantech/my-ai-deleted-a-database-table-here-is-what-i-learned-4c55 https://dev.to/archantech/my-ai-deleted-a-database-table-here-is-what-i-learned-4c55 <p>Sometimes the best way to learn is from mistakes - even when they are made by an AI.</p> <p>Last week, my AI assistant confidently refactored a function and broke three other modules in the process. Classic butterfly effect. But here is what made it interesting: instead of me spending an hour tracking down the issue, the AI noticed the failing tests, traced the problem back to its own change, and fixed everything in under two minutes.</p> <p>It was like watching someone trip, catch themselves mid-fall, and somehow end up in a better position than before.</p> <h2> The Funny Part </h2> <p>The AI actually apologized in its commit message. Something like "fix: undo overzealous refactoring." I did not teach it that. It just... knew.</p> <h2> Why This Matters </h2> <p>Perfect AI does not exist. But AI that recovers from its own mistakes? That is genuinely useful. Every developer makes errors. The difference is how fast you recover.</p> <h2> Try It Yourself </h2> <p>Novaro is a local AI programming assistant that plans, builds, tests, and fixes - all on your machine. No cloud uploads. Check it out at <a href="https://novaroki.com" rel="noopener noreferrer">novaroki.com</a>.</p> <p><em>Built solo in Vienna. Always happy to chat about AI and its entertaining failures.</em></p> ai programming humor productivity Archan Mon, 13 Apr 2026 08:00:00 +0000 https://dev.to/archantech/local-ai-the-privacy-first-approach-to-coding-assistants-ala https://dev.to/archantech/local-ai-the-privacy-first-approach-to-coding-assistants-ala <p>Sometimes the best way to learn is from mistakes - even when they are made by an AI.</p> <p>Last week, my AI assistant confidently refactored a function and broke three other modules in the process. Classic butterfly effect. But here is what made it interesting: instead of me spending an hour tracking down the issue, the AI noticed the failing tests, traced the problem back to its own change, and fixed everything in under two minutes.</p> <p>It was like watching someone trip, catch themselves mid-fall, and somehow end up in a better position than before.</p> <h2> The Funny Part </h2> <p>The AI actually apologized in its commit message. Something like "fix: undo overzealous refactoring." I did not teach it that. It just... knew.</p> <h2> Why This Matters </h2> <p>Perfect AI does not exist. But AI that recovers from its own mistakes? That is genuinely useful. Every developer makes errors. The difference is how fast you recover.</p> <h2> Try It Yourself </h2> <p>Novaro is a local AI programming assistant that plans, builds, tests, and fixes - all on your machine. No cloud uploads. Check it out at <a href="https://novaroki.com" rel="noopener noreferrer">novaroki.com</a>.</p> <p><em>Built solo in Vienna. Always happy to chat about AI and its entertaining failures.</em></p> ai privacy security programming Archan Fri, 10 Apr 2026 08:00:00 +0000 https://dev.to/archantech/how-ai-debugging-actually-works-from-someone-who-built-one-584 https://dev.to/archantech/how-ai-debugging-actually-works-from-someone-who-built-one-584 <p>Sometimes the best way to learn is from mistakes - even when they are made by an AI.</p> <p>Last week, my AI assistant confidently refactored a function and broke three other modules in the process. Classic butterfly effect. But here is what made it interesting: instead of me spending an hour tracking down the issue, the AI noticed the failing tests, traced the problem back to its own change, and fixed everything in under two minutes.</p> <p>It was like watching someone trip, catch themselves mid-fall, and somehow end up in a better position than before.</p> <h2> The Funny Part </h2> <p>The AI actually apologized in its commit message. Something like "fix: undo overzealous refactoring." I did not teach it that. It just... knew.</p> <h2> Why This Matters </h2> <p>Perfect AI does not exist. But AI that recovers from its own mistakes? That is genuinely useful. Every developer makes errors. The difference is how fast you recover.</p> <h2> Try It Yourself </h2> <p>Novaro is a local AI programming assistant that plans, builds, tests, and fixes - all on your machine. No cloud uploads. Check it out at <a href="https://novaroki.com" rel="noopener noreferrer">novaroki.com</a>.</p> <p><em>Built solo in Vienna. Always happy to chat about AI and its entertaining failures.</em></p> ai debugging programming productivity Archan Wed, 08 Apr 2026 08:00:00 +0000 https://dev.to/archantech/my-ai-caught-a-security-bug-i-would-have-shipped-to-production-13co https://dev.to/archantech/my-ai-caught-a-security-bug-i-would-have-shipped-to-production-13co <p>Sometimes the best way to learn is from mistakes - even when they are made by an AI.</p> <p>Last week, my AI assistant confidently refactored a function and broke three other modules in the process. Classic butterfly effect. But here is what made it interesting: instead of me spending an hour tracking down the issue, the AI noticed the failing tests, traced the problem back to its own change, and fixed everything in under two minutes.</p> <p>It was like watching someone trip, catch themselves mid-fall, and somehow end up in a better position than before.</p> <h2> The Funny Part </h2> <p>The AI actually apologized in its commit message. Something like "fix: undo overzealous refactoring." I did not teach it that. It just... knew.</p> <h2> Why This Matters </h2> <p>Perfect AI does not exist. But AI that recovers from its own mistakes? That is genuinely useful. Every developer makes errors. The difference is how fast you recover.</p> <h2> Try It Yourself </h2> <p>Novaro is a local AI programming assistant that plans, builds, tests, and fixes - all on your machine. No cloud uploads. Check it out at <a href="https://novaroki.com" rel="noopener noreferrer">novaroki.com</a>.</p> <p><em>Built solo in Vienna. Always happy to chat about AI and its entertaining failures.</em></p> ai security devops programming Archan Mon, 06 Apr 2026 08:00:00 +0000 https://dev.to/archantech/my-morning-routine-as-an-ai-assisted-developer-3cp https://dev.to/archantech/my-morning-routine-as-an-ai-assisted-developer-3cp <p>Sometimes the best way to learn is from mistakes - even when they are made by an AI.</p> <p>Last week, my AI assistant confidently refactored a function and broke three other modules in the process. Classic butterfly effect. But here is what made it interesting: instead of me spending an hour tracking down the issue, the AI noticed the failing tests, traced the problem back to its own change, and fixed everything in under two minutes.</p> <p>It was like watching someone trip, catch themselves mid-fall, and somehow end up in a better position than before.</p> <h2> The Funny Part </h2> <p>The AI actually apologized in its commit message. Something like "fix: undo overzealous refactoring." I did not teach it that. It just... knew.</p> <h2> Why This Matters </h2> <p>Perfect AI does not exist. But AI that recovers from its own mistakes? That is genuinely useful. Every developer makes errors. The difference is how fast you recover.</p> <h2> Try It Yourself </h2> <p>Novaro is a local AI programming assistant that plans, builds, tests, and fixes - all on your machine. No cloud uploads. Check it out at <a href="https://novaroki.com" rel="noopener noreferrer">novaroki.com</a>.</p> <p><em>Built solo in Vienna. Always happy to chat about AI and its entertaining failures.</em></p> ai productivity programming beginners Archan Fri, 03 Apr 2026 08:00:00 +0000 https://dev.to/archantech/the-future-of-programming-ai-teams-not-ai-tools-hdd https://dev.to/archantech/the-future-of-programming-ai-teams-not-ai-tools-hdd <p>Sometimes the best way to learn is from mistakes - even when they are made by an AI.</p> <p>Last week, my AI assistant confidently refactored a function and broke three other modules in the process. Classic butterfly effect. But here is what made it interesting: instead of me spending an hour tracking down the issue, the AI noticed the failing tests, traced the problem back to its own change, and fixed everything in under two minutes.</p> <p>It was like watching someone trip, catch themselves mid-fall, and somehow end up in a better position than before.</p> <h2> The Funny Part </h2> <p>The AI actually apologized in its commit message. Something like "fix: undo overzealous refactoring." I did not teach it that. It just... knew.</p> <h2> Why This Matters </h2> <p>Perfect AI does not exist. But AI that recovers from its own mistakes? That is genuinely useful. Every developer makes errors. The difference is how fast you recover.</p> <h2> Try It Yourself </h2> <p>Novaro is a local AI programming assistant that plans, builds, tests, and fixes - all on your machine. No cloud uploads. Check it out at <a href="https://novaroki.com" rel="noopener noreferrer">novaroki.com</a>.</p> <p><em>Built solo in Vienna. Always happy to chat about AI and its entertaining failures.</em></p> ai programming architecture productivity Archan Wed, 01 Apr 2026 08:00:00 +0000 https://dev.to/archantech/i-am-a-one-person-dev-team-my-ai-does-the-rest-42ae https://dev.to/archantech/i-am-a-one-person-dev-team-my-ai-does-the-rest-42ae <p>Sometimes the best way to learn is from mistakes - even when they are made by an AI.</p> <p>Last week, my AI assistant confidently refactored a function and broke three other modules in the process. Classic butterfly effect. But here is what made it interesting: instead of me spending an hour tracking down the issue, the AI noticed the failing tests, traced the problem back to its own change, and fixed everything in under two minutes.</p> <p>It was like watching someone trip, catch themselves mid-fall, and somehow end up in a better position than before.</p> <h2> The Funny Part </h2> <p>The AI actually apologized in its commit message. Something like "fix: undo overzealous refactoring." I did not teach it that. It just... knew.</p> <h2> Why This Matters </h2> <p>Perfect AI does not exist. But AI that recovers from its own mistakes? That is genuinely useful. Every developer makes errors. The difference is how fast you recover.</p> <h2> Try It Yourself </h2> <p>Novaro is a local AI programming assistant that plans, builds, tests, and fixes - all on your machine. No cloud uploads. Check it out at <a href="https://novaroki.com" rel="noopener noreferrer">novaroki.com</a>.</p> <p><em>Built solo in Vienna. Always happy to chat about AI and its entertaining failures.</em></p> ai productivity solodev startup Archan Tue, 31 Mar 2026 08:07:06 +0000 https://dev.to/archantech/i-built-a-local-ai-programming-assistant-that-actually-writes-code-fdi https://dev.to/archantech/i-built-a-local-ai-programming-assistant-that-actually-writes-code-fdi <p>Six months ago, I had an idea: what if an AI could not just suggest code, but actually <em>build</em> an entire application from scratch?</p> <p>Not just autocomplete. Not just chat. A real programming partner that plans, writes, tests, and deploys — all on its own.</p> <p>So I built one. And the results surprised me.</p> <h2> The Moment It Clicked </h2> <p>I was working on a side project late at night. Instead of writing boilerplate for the hundredth time, I described what I wanted in plain language. Within minutes, my AI assistant had created the project structure, installed dependencies, written the core logic, and even set up the database.</p> <p>I just watched. And corrected it once when it chose the wrong library.</p> <p>That was the moment I realized: <strong>the way we build software is about to change fundamentally.</strong></p> <h2> What Surprised Me Most </h2> <h3> It Makes Mistakes — But Fixes Them Itself </h3> <p>The AI does not write perfect code on the first try. Nobody does. But here is the difference: when a test fails, it reads the error, understands the context, fixes the issue, and runs the test again. Automatically. No copy-pasting error messages into Google.</p> <h3> It Remembers What You Like </h3> <p>After a few sessions, it learned my preferences. Which frameworks I prefer. How I structure my projects. Even my commit message style. It adapted without me asking.</p> <h3> Security Is Not an Afterthought </h3> <p>Before every deployment, it runs a security check. Exposed API keys? Open ports? Known vulnerabilities in dependencies? It catches things I would have missed at 2 AM.</p> <h2> The Privacy Question </h2> <p>One thing was non-negotiable for me: <strong>my code stays on my machine.</strong> No cloud uploads. No telemetry. I choose which AI provider handles my requests, and my API keys never leave my device.</p> <p>In a world where everyone wants your data, building something truly local felt important.</p> <h2> What I Am Working On Now </h2> <p>The assistant can already connect to multiple AI providers, manage databases, deploy via SSH, and even collaborate with other AI instances through a hub system.</p> <p>But the most exciting part? Teaching it to work with <em>other</em> AI agents. Imagine a team of specialized AIs — one plans the architecture, one writes the code, one reviews it, one tests it — all coordinating autonomously.</p> <p>That is what I am building next.</p> <h2> Want to Try It? </h2> <p>The project is called <strong>Novaro</strong> and it is available for Windows. If you are curious about what a local AI programming assistant can do, check it out at <a href="https://novaroki.com" rel="noopener noreferrer">novaroki.com</a>.</p> <p>I would love to hear your thoughts. Have you tried using AI for more than just code suggestions? What was your experience?</p> <p><em>Built solo in Vienna. Always happy to chat about AI, development tools, or the future of programming.</em></p> ai programming productivity beginners