To help you get the most out of AI, we're building a new benchmark to measure how helpful LLMs are for Android development. The benchmark will be composed of real-world problems sourced from public GitHub Android repositories, allowing us to measure a model’s ability to navigate complex codebases, understand dependencies, and solve the kind of problems you encounter every day. Read more → goo.gle/48PjjpX

@AndroidDev Android, this is a brilliant initiative! It's about time someone focused on practical LLM applications, no?

@AndroidDev You gotta give Gemini a direct feed to latest Android patterns. All the LLMs give out of date info (they recommend navigation 2 for example, or classic xml views instead of jetpack.)

@AndroidDev One day or another @Xiaomi , or @Huawei will get tired of all this shitty NDKs, SDKs, cmdlines, NDKs, JDKs to make a stupid 600 lines code and then they will decide make their own stuff and every body will buy it. On this day nobody will ever remember existed ur shitty OS