AI is Moving Too Fast. Here's How I'm Keeping Up.
My personal system for staying on top of the AI explosion as an Android developer, without spending all day on Twitter.
The Problem
Every week there's a new model, a new framework, a new paper that supposedly changes everything. If you try to follow all of it, you burn out. If you ignore it, you fall behind.
As an Android developer, I don't need to know everything. I need to know what's relevant to what I build.
My Filter
I ask one question about anything new:
"Can this run on a phone, or integrate with an Android app in a meaningful way?"
If yes, I dig in. If no, I file it away and move on.
This has been incredibly useful. It cuts 80% of the noise immediately.
What's Actually Relevant to Android Developers Right Now
- On-device models: Gemma 2B, Phi-3 Mini, Mistral 7B quantized. These fit on modern phones.
- MediaPipe: Google's ML framework for mobile, covering face detection, gesture recognition, and LLM inference.
- Gemini API: Multimodal, fast, and now has an Android SDK.
- AI Agents: ADK and AWS Bedrock Agents are useful for building backend AI that your Android app talks to.
- llama.cpp: The runtime that makes on-device LLMs practical.
My Learning System
- Build something every 2 weeks. Even a tiny PoC. Nothing teaches like shipping.
- Write about it. Writing forces you to actually understand something instead of just skimming it.
- Follow 5 people max on Twitter/X. Quality over quantity. Follow the people who are building, not the ones who are just commentating.
- Read papers only if you're actually building with the technique. Otherwise it is just noise.
The Mindset Shift
Stop trying to understand everything. Start trying to build one thing well, then the next thing well.
The developers who will stand out in the AI era are not the ones who know the most buzzwords. They are the ones who actually shipped something.
This site is my attempt to be one of those people.