Super interesting perspective! Can't agree more as this could help us create more personalized interfaces that adapt on the go and ultimately serve the user better. I would be curious to know what's the ideal velocity of interfaces adaptating and what factors the velocity should be a function of!
This piece really made me think, building so well on your Designing Intelligent Products series. I'm completely fascinated by the idea of learning loops evolving and honestly wonder what the next iteraton after programmatic clicks will be for trully intelligent systems.
Super interesting perspective! Can't agree more as this could help us create more personalized interfaces that adapt on the go and ultimately serve the user better. I would be curious to know what's the ideal velocity of interfaces adaptating and what factors the velocity should be a function of!
This piece really made me think, building so well on your Designing Intelligent Products series. I'm completely fascinated by the idea of learning loops evolving and honestly wonder what the next iteraton after programmatic clicks will be for trully intelligent systems.
great insights. I can see how this is valuable for prompt based interactions, eg. "accept plan? yes, no, change".
Also maybe for tool calls: user reacted x or y or z on tool call dialog.
Can you help me understand how this works for actual interfaces, i mean situations in whcih LLMs create the actual UI components. Is this a thing yet?