Talking to Think

These two posts make a case for the same fundamental shift from different vantage points. One is about voice mode as a thinking partner — the ability to brainstorm out loud with an AI during a commute, free from social overhead and the weight of managing another person's experience. The other is about dictation as an input method — the way speaking bypasses the self-editing that happens when you type, producing richer, more specific prompts that models respond to dramatically better.

Together they paint a picture of what a voice-first AI workflow actually looks like: you're not just transcribing — you're thinking differently. The spoken word unlocks a messiness and specificity that typing filters out, and the AI is better for it. The tools are ready. The social norms just haven't caught up yet.