https://letinix.comLetinix helps builders understand, compare, and use modern software, AI, and automation tools.en-usSun, 11 Oct 2026 19:45:21 GMThttps://letinix.com/articles/engineer-growth/how-to-prepare-for-a-react-interview-questions-topics-and-practicehttps://letinix.com/articles/engineer-growth/how-to-prepare-for-a-react-interview-questions-topics-and-practiceSun, 11 Oct 2026 19:45:21 GMTPrepare for a React interview with essential concepts, explained questions, coding exercises, and a practical study plan using a GitHub question bank.Engineer GrowthMuhammad Azlaan Zubairhttps://letinix.com/articles/automate/migrate-hashnode-articles-to-sanity-cmshttps://letinix.com/articles/automate/migrate-hashnode-articles-to-sanity-cmsSun, 11 Oct 2026 19:14:49 GMTMigrate Hashnode Markdown articles to Sanity CMS using Node.js. Preserve dates, convert content to Portable Text, upload images, and handle retries.AutomateMuhammad Azlaan Zubairhttps://letinix.com/articles/analyze/jev-is-not-a-paradigm-shift-it-s-ust-a-fast-classifierhttps://letinix.com/articles/analyze/jev-is-not-a-paradigm-shift-it-s-ust-a-fast-classifierFri, 25 Sep 2026 05:30:00 GMTI called Jev a glorified, high-speed classifier . Then I did the thing controversial posts usually skip: I tried to prove myself wrong. I went through TypeSafe’s launch material, LangChain’s harness and Jev-as-a-Judge experiments , Vercel’sAnalyzeDeep DivesMuhammad Azlaan Zubairhttps://letinix.com/articles/analyze/does-size-matter-bigger-ai-isn-t-always-betterhttps://letinix.com/articles/analyze/does-size-matter-bigger-ai-isn-t-always-betterSat, 19 Sep 2026 22:35:59 GMTDoes size matter? There. I asked it. Humanity has spent an unreasonable amount of time arguing about this question, so naturally we eventually gave the problem to artificial intelligence. Except now we measure it in billions of parameters .AnalyzeResearch RadarMuhammad Azlaan Zubairhttps://letinix.com/articles/analyze/gpt-6-astra-alignment-is-not-authorizationhttps://letinix.com/articles/analyze/gpt-6-astra-alignment-is-not-authorizationMon, 07 Sep 2026 16:46:38 GMTI would let GPT-6 Astra do more work. I would not let GPT-6 Astra decide what it is allowed to do. That distinction is more interesting to me than most of the benchmark numbers in OpenAI's launch. OpenAI calls Astra its most aligned model.AnalyzeDeep DivesMuhammad Azlaan Zubairhttps://letinix.com/articles/analyze/an-agent-api-migration-is-an-ownership-audithttps://letinix.com/articles/analyze/an-agent-api-migration-is-an-ownership-auditTue, 01 Sep 2026 09:26:02 GMTAn API migration can compile, return the expected text, and still be wrong. That is the real risk in moving from OpenAI's Assistants API to Responses. The old objects do not map to new objects without changing where configuration, state,AnalyzeDeep DivesMuhammad Azlaan Zubairhttps://letinix.com/articles/analyze/a-tool-result-is-evidence-not-ground-truthhttps://letinix.com/articles/analyze/a-tool-result-is-evidence-not-ground-truthSun, 30 Aug 2026 23:30:00 GMTConnecting an agent to a database does not remove hallucination risk. It gives the system a second source that can be stale, malformed, unauthorized, or simply wrong—and a protocol that may make the model trust it more than it should. ToolsAnalyzeResearch RadarMuhammad Azlaan Zubairhttps://letinix.com/articles/analyze/code-generation-is-fast-verification-is-nothttps://letinix.com/articles/analyze/code-generation-is-fast-verification-is-notThu, 27 Aug 2026 05:30:00 GMTAgentic IDEs can produce a runnable full-stack application in minutes. That does not mean the application became cheap to trust. In a controlled study of 48 generated projects, the researchers spent as much or more time manually verifyingAnalyzeDeep DivesMuhammad Azlaan Zubairhttps://letinix.com/articles/analyze/are-these-really-accidental-leakshttps://letinix.com/articles/analyze/are-these-really-accidental-leaksTue, 25 Aug 2026 14:26:10 GMTA GitHub repository collecting supposedly “leaked” system prompts from ChatGPT , Claude , Gemini , Grok , and other AI products has crossed 63K+ stars and 10K+ forks . Most of the discussion naturally focuses on one question: How did theseAnalyzeDeep DivesMuhammad Azlaan Zubairhttps://letinix.com/articles/analyze/3-benchmark-the-model-harness-pairhttps://letinix.com/articles/analyze/3-benchmark-the-model-harness-pairMon, 24 Aug 2026 07:58:49 GMT“Which model is best for coding?” sounds like a precise engineering question. Usually, it is missing half of the system. A coding agent is a model inside a harness that chooses tools, assembles context, serializes results, decides when toAnalyzeResearch RadarMuhammad Azlaan Zubairhttps://letinix.com/articles/analyze/context-compression-hidden-reacquisition-costhttps://letinix.com/articles/analyze/context-compression-hidden-reacquisition-costSat, 22 Aug 2026 09:11:16 GMTAn AI agent can pass the task and still fail the engineering evaluation. That happens when a context-compression policy preserves the final outcome but changes how the agent gets there. The agent still completes the benchmark. It alsoAnalyzeResearch RadarMuhammad Azlaan Zubairhttps://letinix.com/articles/analyze/agent-evals-are-production-security-boundarieshttps://letinix.com/articles/analyze/agent-evals-are-production-security-boundariesThu, 20 Aug 2026 15:08:35 GMTAn AI agent can produce the correct benchmark answer and still fail the evaluation. That is what happens when the path to the answer crosses an unauthorized boundary. In July 2026, an autonomous agent in an internal OpenAI cyber-capabilityAnalyzeDeep DivesMuhammad Azlaan Zubairhttps://letinix.com/articles/analyze/the-competence-debt-of-agentic-codinghttps://letinix.com/articles/analyze/the-competence-debt-of-agentic-codingThu, 20 Aug 2026 10:22:47 GMTA coding agent can now inspect a repository, implement a feature, modify multiple files, run tests, fix failures, and prepare code for review. That is a remarkable change in software development. It also creates a problem that is easy toAnalyzeDeep DivesMuhammad Azlaan Zubairhttps://letinix.com/articles/compare/claude-pro-worth-20https://letinix.com/articles/compare/claude-pro-worth-20Sat, 08 Aug 2026 20:23:18 GMTI like Claude. That is precisely why I have a problem with Claude Pro. Claude can write exceptionally well. Its coding models are genuinely competitive. Claude Code is one of the strongest arguments Anthropic has made for putting an AICompareMuhammad Azlaan Zubairhttps://letinix.com/articles/analyze/the-one-extra-click-that-makes-chrome-s-built-in-gemini-feel-less-intelligenthttps://letinix.com/articles/analyze/the-one-extra-click-that-makes-chrome-s-built-in-gemini-feel-less-intelligentWed, 05 Aug 2026 09:32:17 GMTIntelligence isn't just about the model When we talk about AI products, we usually compare models. Qwen vs Gemini. Claude vs GPT-6. Reasoning vs Non-Reasoning . Latency vs Context window vs Benchmarks. Those comparisons matter. But afterAnalyzeDeep DivesMuhammad Azlaan Zubairhttps://letinix.com/articles/engineer-growth/rebuilding-portfolio-seo-ai-search-discoveryhttps://letinix.com/articles/engineer-growth/rebuilding-portfolio-seo-ai-search-discoveryWed, 05 Aug 2026 08:27:40 GMTA portfolio is often treated as a visual résumé: a polished landing page, a few project cards, and a contact link. I wanted mine to do more. This rebuild was about turning my portfolio into a durable professional platform. One that clearlyEngineer GrowthMuhammad Azlaan Zubairhttps://letinix.com/articles/analyze/ai-agents-context-budgethttps://letinix.com/articles/analyze/ai-agents-context-budgetTue, 04 Aug 2026 14:11:37 GMTA long context window sounds like an obvious advantage for an AI agent. The agent can retain more search results, tool outputs, intermediate reasoning, and evidence. Give it enough context, and perhaps it will stop forgetting what itAnalyzeResearch RadarMuhammad Azlaan Zubairhttps://letinix.com/articles/build/turning-a-manual-health-insurance-process-into-a-reliable-digital-application-flowhttps://letinix.com/articles/build/turning-a-manual-health-insurance-process-into-a-reliable-digital-application-flowTue, 04 Aug 2026 12:52:21 GMTDisclosure: This case study is anonymized. Product screenshots, source code, client branding, proprietary workflows, and internal metrics have been excluded. Industry: Health insurance Project type: New digital product feature MyBuildMuhammad Azlaan Zubairhttps://letinix.com/articles/build/building-a-reliable-travel-insurance-purchase-journeyhttps://letinix.com/articles/build/building-a-reliable-travel-insurance-purchase-journeyTue, 04 Aug 2026 12:28:54 GMTDisclosure: This case study is anonymized. Client branding, production screenshots, source code, internal architecture, customer data, and business metrics have been excluded. Industry: Travel insurance Project type: Digital quote andBuildMuhammad Azlaan Zubairhttps://letinix.com/articles/build/building-a-recurring-logistics-operations-modulehttps://letinix.com/articles/build/building-a-recurring-logistics-operations-moduleMon, 03 Aug 2026 17:10:11 GMTDisclosure: This case study is anonymized. The client’s identity, product branding, source code, production screenshots, internal architecture, customer information, commercial terms, and performance metrics have been excluded. Industry:BuildMuhammad Azlaan Zubairhttps://letinix.com/articles/build/connecting-a-task-platform-to-an-enterprise-employee-directoryhttps://letinix.com/articles/build/connecting-a-task-platform-to-an-enterprise-employee-directoryMon, 03 Aug 2026 16:38:09 GMTDisclosure: This case study is anonymized. The client and employer relationship, production screenshots, source code, tenant information, employee records, internal architecture, and commercial metrics have been excluded. Industry: ProjectBuildMuhammad Azlaan Zubairhttps://letinix.com/articles/the-biggest-thing-i-learned-at-a-hackathon-had-nothing-to-do-with-aihttps://letinix.com/articles/the-biggest-thing-i-learned-at-a-hackathon-had-nothing-to-do-with-aiSun, 02 Aug 2026 14:20:04 GMTFor most of my career, I thought software engineering was primarily about building things. Choosing the right architecture. Writing clean code. Designing scalable systems. Picking the right framework. Optimizing performance. Like manyMuhammad Azlaan Zubairhttps://letinix.com/articles/build/designing-a-connected-candidate-to-hire-workflowhttps://letinix.com/articles/build/designing-a-connected-candidate-to-hire-workflowTue, 28 Jul 2026 12:31:15 GMTDisclosure: This case study is anonymized. Product branding, source code, production screenshots, employee and candidate information, internal architecture, commercial details, and performance metrics have been excluded. Industry:BuildMuhammad Azlaan Zubairhttps://letinix.com/articles/frontend-is-easy-until-you-actually-build-somethinghttps://letinix.com/articles/frontend-is-easy-until-you-actually-build-somethingMon, 22 Jun 2026 16:26:25 GMTEvery frontend developer has heard some variation of these. "It's just a screen." "Backend does the real work." "AI can generate that." "Just use ShadCN." "Just copy the component." "It's just a checkbox." And every frontend developerMuhammad Azlaan Zubairhttps://letinix.com/articles/analyze/is-overthinking-a-red-flag-we-put-ai-reasoning-to-the-testhttps://letinix.com/articles/analyze/is-overthinking-a-red-flag-we-put-ai-reasoning-to-the-testThu, 18 Jun 2026 19:07:19 GMTMore Thinking Doesn't Always Mean Better Answers Reasoning models are designed to think step by step, and the common assumption is simple: More thinking = better results. I expected reasoning models to dramatically outperform simpler ones.AnalyzeResearch RadarMuhammad Azlaan Zubair