Salience is about redefining how humans and computers collaborate. For decades, computing has forced people to adapt to machines—learning rigid interfaces, breaking ideas into commands, and translating rich human intent into low‑bandwidth inputs. As AI agents become more capable, this mismatch becomes the limiting factor: not what computers can do, but how well they can understand us.
Salience flips that relationship. Instead of requiring humans to speak the language of machines, it enables computers to understand humans more naturally—through context, behavior, and cognitive signals. By increasing the bandwidth between people and AI, Salience allows agents to move from passive tools to active collaborators that operate at the level of goals, not commands.
While our first focus is accessibility—unlocking independence and productivity for people with motor impairments—this approach generalizes to everyone. As AI becomes embedded in daily work and creative processes, higher‑bandwidth human‑AI communication will be essential. Salience is laying the foundation for an operating system where intent, not interaction mechanics, is the primary interface.