I'm bullish on voice-only computing. Every month since AI has become usable, I feel increasing pain at having to use screens at all. With LLMs it now feels like a huge waste of time to do things like "click" and move between "tabs." When I look at the improvement rate of agents especially, it now seems like a waste of time to do almost any kind of manual task on my computer. The gold standard life very soon may be reading paper and thinking in a sun-filled room, possibly with some kind of team if you need one, and essentially just making judgments and speaking words to handle every single thing on a todo list. Everything else somehow feels intolerably archaic already. I sort of hate to admit this but the OpenAI device might be the thing.