[21st April 2025] Interesting Things I Learnt This Week

1. How Major Indian Apps Are Bypassing Google’s Policy to Spy on Your App Drawer? - Report reveals that many major Android apps on the Google Play Store are allegedly bypassing Google's privacy policies to access a list of other apps installed on users' devices. Apps like Swiggy, Zepto, and HDFC are among those seemingly "peeping" into users' app drawers. While some argue this access might be needed for core functionality (like fraud detection in banking apps), the report suggests it's being used for fingerprinting and user profiling. Analysis of AndroidManifest.xml files shows apps use methods like querying by package name or intent signature to extract this data. Concerns are raised about apps like KreditBee, which appear to be profiling users based on the presence of specific apps, including those related to religion and astrology. The report questions Google's enforcement of its policies and the lack of user consent or awareness regarding this data collection.

My Take: Apps have been doing this for ages. Games and loan apps are equally complicit. This isn't just a local issue; it happens worldwide. It's disheartening that Google doesn't take action to stop these kinds of things, even if they are aware. Considering their recent layoffs in the Android team (among others), the situation is likely to worsen. I think we might see a resurgence of companies like McAfee and Norton who could potentially address this. However, I have very little hope.

 

2.  Iris - IRIS and open source app offers a streamlined experience for offline AI model interaction through a redesigned user interface. Key features include direct integration with Hugging Face for model downloads, efficient model management, speech-to-text and text-to-speech capabilities, and hyperparameter tuning for customized model behavior. The app also allows for model benchmarking to measure performance and features improved conversational flow with automatic chat templates. Recent updates have enhanced inference stability, long-running session reliability, and fixed UI issues. IRIS empowers users to manage and utilize advanced AI models with powerful tools, all without an internet connection. 

My Take: This is a lovely app. It boasts a beautiful design and, depending on your mobile's capabilities, works locally for many models. This truly represents the future of mobile LLMs. I envision numerous different actions, such as agentic tasks, being powered by apps like these. Apple Intelligence and Gemini, despite their power, will become obsolete due to network requirements. My hope is that developers design these apps with easy extensibility in mind. Having said that, most of today's hardware doesn't meet the requirements for truly powerful LLMs. However, I am optimistic that as hardware improves and software becomes more efficient, they will enable this future.

3. The NSE confusions - Explains the confusion with Nobody, Somebody and Everybody. The work begins with finding the right somebodies, while ignoring the imaginary everyone.

My Take: Seth Godin possesses a rare ability to distill profound ideas into remarkably few words. His statement, "The work begins with finding the right somebodies, while ignoring the imaginary everyone," is particularly insightful. Identifying the "right somebodies" can be challenging, as we often fall into the trap of trying to cater to everyone. Furthermore, the act of saying no, and persuading others to agree to that, stands as one of the most difficult aspects of being a founder or for that matter a product manager.