The class looked at where UX is heading in the context of AI operating systems and adaptive environments. The idea that devices configure themselves around what you're doing rather than requiring you to configure them is something that maps almost directly onto the Future Living brief. Intent over commands is the phrase that stuck — the interface shouldn't require the user to know what to ask for, it should infer it.
The smart home in its current app-based form was discussed as a friction problem. Having to open a specific app to control a specific device defeats a lot of the purpose. The value of ambient computing is that it removes the layer between intention and outcome. But that comes with real trust implications — automation is good until it isn't, and the moment it fails badly enough, users abandon it entirely.
For Pane OS, this is the core tension to design around. How do you build something that feels genuinely ambient and intelligent without making the user feel like they've lost control? The summon flow I'm thinking about is partly an answer to that — the system is passive until called on, which preserves the sense of agency even as the underlying technology is doing a lot of work
The discussion of adaptive tech — systems that change to meet user needs rather than requiring users to adapt to them — opened up an interesting design question. How do you design for something that adapts? Do you define the variables and let the system fill them, or does the adaptation happen in a space that hasn't been designed for at all?
For a smart home OS this is very live. The system needs to learn patterns without being creepy about it, respond to context without being presumptuous, and fail gracefully when its inferences are wrong. Those are hard design problems and they're mostly unsolved in current products
There was a conversation about the difference between a designer prompting an AI tool and a non-designer doing the same thing. My instinct is that it comes down to refinement and direction. A novice gets an output and accepts it. A designer knows where to take it next — what's wrong about it, what direction to push it, what to strip out. The AI doesn't prompt itself to innovate, it repeats and interpolates. The creative direction has to come from somewhere and that somewhere is still the designer.
Kyle talked about the importance of side projects, which is something I feel strongly about but find genuinely hard to act on as a student who also works. The design community online cannot stress them enough — people are getting hired off the back of them constantly. It's something I want to make more space for, and hopefully by the end of this module I'll have something to show for it.
<aside> 💡
This was probably the most directly relevant class to the Future Living project. Intent over commands as a design principle is essentially the brief I've set myself with Pane OS — a system that responds to presence and context rather than requiring explicit instruction. The trust discussion was equally useful. Automation that works invisibly is fine until it doesn't, and that single failure can undo a lot of goodwill. Designing the failure states and the moments of user override feels just as important as designing the ambient functionality. The prompting conversation also landed well. The difference between a designer and a non-designer using AI tools isn't access, it's direction — knowing where to take the output is the skill, and that's still entirely human.
</aside>