https://ram.oxrolabs.com/blogPosts by Ramswaroop Patel on AI, security and shipping. https://ram.oxrolabs.com/blog/life-latelyhttps://ram.oxrolabs.com/blog/life-latelySat, 03 Oct 2026 00:00:00 GMTOne year. From thinking ₹1L was a huge number to seeing ₹4.2L in my bank account, signing company documents, meeting lawyers and CAs, staying up till 3 AM, going to college at 6:40, and slowly becoming someone different. https://ram.oxrolabs.com/blog/why-the-verified-tickhttps://ram.oxrolabs.com/blog/why-the-verified-tickFri, 02 Oct 2026 00:00:00 GMTI gave myself a blue tick. It took 4 lines of SVG and one Claude prompt. Here's the full, completely serious engineering report. https://ram.oxrolabs.com/blog/add-a-post-in-60-secondshttps://ram.oxrolabs.com/blog/add-a-post-in-60-secondsThu, 01 Oct 2026 00:00:00 GMTNo CMS. No database. No tears. Drop an HTML file into /public/posts, run the dev server, done. A guide to the laziest blog engine ever built. https://ram.oxrolabs.com/blog/ai-agents-without-the-hypehttps://ram.oxrolabs.com/blog/ai-agents-without-the-hypeMon, 28 Sep 2026 00:00:00 GMTAn agent is a language model, a handful of tools and a loop. That is it. Here is how the loop works, where it breaks, and when a boring workflow beats it. https://ram.oxrolabs.com/blog/mcp-the-usb-c-of-ai-toolshttps://ram.oxrolabs.com/blog/mcp-the-usb-c-of-ai-toolsThu, 24 Sep 2026 00:00:00 GMTThe Model Context Protocol turns every app-times-tool integration into one standard plug. What it is, why it matters, and the security habits it demands. https://ram.oxrolabs.com/blog/prompting-is-dead-long-live-contexthttps://ram.oxrolabs.com/blog/prompting-is-dead-long-live-contextSat, 19 Sep 2026 00:00:00 GMTClever wording matters less than what you put in the context window. A practical way to think about instructions, examples, retrieved facts, tools and budgets. https://ram.oxrolabs.com/blog/prompt-injection-101https://ram.oxrolabs.com/blog/prompt-injection-101Sat, 12 Sep 2026 00:00:00 GMTWhy an AI that reads untrusted text can be tricked by that text, and the boring, effective defences: least privilege, isolation, and humans in the loop. https://ram.oxrolabs.com/blog/rag-vs-long-context-vs-fine-tuninghttps://ram.oxrolabs.com/blog/rag-vs-long-context-vs-fine-tuningSat, 05 Sep 2026 00:00:00 GMTThree ways to make a model know your stuff, and a plain decision table for choosing between them without a flame war.