Dinesh Babu (@dinesh_r7) 's Twitter Profile
Dinesh Babu

@dinesh_r7

Android Developer | iOS User | Tech Enthusiast | Exploring the art of photography and drawing | Sharing insights on productivity, tech, and creativity

ID: 406809785

calendar_today07-11-2011 06:53:04

56 Tweet

50 Followers

212 Following

Dinesh Babu (@dinesh_r7) 's Twitter Profile Photo

Share your thoughts. Given a choice, which one would you choose and why? KMP with Compose UI for both Android and iOS? or KMP with native UI (Compose for Android and SwiftUI for iOS)?

Dinesh Babu (@dinesh_r7) 's Twitter Profile Photo

Most developers treat AI agents like autocomplete. Real power emerges through agentic workflows: reasoning loops, tool integration, and contextual memory. When agents break down problems and learn from context, they shift from helpful to transformative. #AIAgent

Dinesh Babu (@dinesh_r7) 's Twitter Profile Photo

Another win for device security and trust. We are going somewhere, thats a nice progress overall. #GooglePixel androidcentral.com/phones/google-…

Dinesh Babu (@dinesh_r7) 's Twitter Profile Photo

This could be the perfect official documentation for Liquid Glass for KMP. It might help someone in #BuildInPublic. medium.com/mateedevs/liqu…

Dinesh Babu (@dinesh_r7) 's Twitter Profile Photo

Why don't all companies fund/develop a single IDE to be the greatest agentic IDE and allow users to sign in to any of their providers?

Dinesh Babu (@dinesh_r7) 's Twitter Profile Photo

Agents transform KMP development by generating shared logic while you focus on architecture. Provide clear platform context for reliable, faster delivery without compromising quality. Yes! You can build a Prod ready high Quality code with LLM and agents. I'm so obsessed with

Dinesh Babu (@dinesh_r7) 's Twitter Profile Photo

Building AI features? Keep API keys in your backend, not client code. Route model requests through your server to control access, prevent credential exposure, and maintain flexibility. This isn't overhead, it's essential architecture.

Dinesh Babu (@dinesh_r7) 's Twitter Profile Photo

Spent the week integrating MCP into a work project. The difference is night and day. No more writing custom adapters for each LLM. No more maintaining separate context handlers. One protocol, swap models at runtime. Clean architecture principle in action: depend on

Dinesh Babu (@dinesh_r7) 's Twitter Profile Photo

2025: The year we stopped saying "Generative AI" and started building "Agents". Watched the shift from one-off prompts to multi-agent workflows with specialized roles. Wonder what 2026 brings when everyone assumes agents > chatbots.

Dinesh Babu (@dinesh_r7) 's Twitter Profile Photo

Mobile app talking directly to 5+ microservices? That's a maintenance nightmare. API Gateway solves this: single entry point for auth, rate limiting, and request composition. Backend changes don't break mobile clients. Clean architecture principle: depend on abstractions, not

Dinesh Babu (@dinesh_r7) 's Twitter Profile Photo

I have been using Claude Opus 4.5 for the past week. It is far better for coding than Sonnet or GPT Codex models. Its a wide margin than the best model I’ve ever used. Anthropic Please keep Opus accessible and cheaper forever.