Keynote - Android Exploits & Insights
summary
Argues that large language models are effectively trained on community content—Stack Overflow, conference transcripts and books—and shows a concrete experiment where 15 tips vs Gemini/ChatGPT produces surprising agreement with legacy advice; notes three dozen ISBNs of older books show up as LLM sources. Describes building small company prototypes using Compose Multiplatform demos and KMP, highlights that many tests only catch known failures so invest in on‑call preparedness, observability and fast revert paths, and reports the demos work enough to ship.
AI-written summary · about
chapters (11)
- 0:00Android exploits & insights
- 2:27Mobile development journey
- 9:08AI in mobile development
- 13:20Why sharing insights matters
- 21:13AI in decision-making
- 25:48Feedback & growth in engineering
- 31:05Android development shifts
- 37:53KMP and Compose opportunities
- 44:09Influence Android's future
- 47:38Share novel ideas publicly
- 49:43Closing remarks
related talks
34mKeynote: Comedy Talk - Better Android Development (BAD)
33mKeynote: What is the future of the community in an always-evolving world?
57mKeynote: Using LLMs and generative AI to fix software vulnerabilities
44mAI Foundations: Mitigate Risks and Boost SOC Efficiency
41mKeynote: Now in Modern Android
51mDeveloper Keynote
44mSecuring LLM-Powered Apps: Overcoming Security and Privacy Challenges
45mThe Keynote!