A 16-year-old raises the core friction with AI: tools still demand you explain yourself every time, instead of learning from your phone's data

A teenager posted on Hacker News questioning why current AI tools require manual prompts and context every single time, rather than learning from existing data on your device (calendar, reminders, health records) to anticipate what you need — sparking a discussion about whether 'AI that knows' is a real problem users face or a technical dead-end.

The core trade-off emerged: AI could learn your patterns by accessing everything on your phone to infer intent, but commenters warned this creates the same algorithmic trap as social media — self-reinforcing bubbles that narrow what you see rather than expand what you can do.

For you: this conversation signals growing frustration that AI assistants still feel like dumb tools requiring constant hand-holding, not smart agents. It highlights why companies investing in 'proactive AI' (systems that act before you ask) face a hard choice between privacy and usefulness — blanket device access feels creepy to most people, but limited access leaves AI guessing.