#ux

2 notes

#design

Here’s a quick reference list of questions to ask when designing a product, when you want to get to the core of your product, trim the fat, sort out the signal from the noise, etc.

VISCERAL IMPACT & DESTINY

  1. “What would make someone’s eyes light up when they first see this?”
  2. “How does this change people’s relationship with [the domain]?”
  3. “What limitation are people accepting as ‘just how it is’ that we could eliminate?”
  4. “If this succeeds wildly, what industry becomes obsolete?”

EXPERIENCE PERFECTIONISM

  1. “What moment in using this feels even slightly awkward or inelegant?”
  2. “Where are we making the user do something the computer should know?”
  3. “What detail would seem obsessive to include but users would love once they discover it?”
  4. “Which part of this experience isn’t bringing joy yet?”

PRODUCT TRUTH

  1. “What’s the ‘one thing’ this product is about that everything else serves?”
  2. “Which feature are we keeping because we’re afraid to remove it?”
  3. “What would this look like if we had to remove half the features?”
  4. “Where are we compromising on excellence to accommodate edge cases?”

REVOLUTIONARY POTENTIAL

  1. “How could this be 10x better, not just 2x better?”
  2. “What crazy idea for this product would be dismissed as impossible but would be magical if it worked?”
  3. “What would make the mainstream mock this at first but later seem obvious?”
  4. “How does this contribute to making technology more human?”
Permalink →

#ai There are two fundamental, user-friendly UIs for AI - the Chatbot and the Contextual Suggestion. Both simulate human interaction, making them intuitive and easy to use, especially with the advent of GPT models.

🤖 Chatbot

Imagine sending a message to the AI, and it responds back - just like a conversation on SMS, WhatsApp, or Facebook Messenger. This is the Chatbot UI, a familiar and straightforward interface used worldwide.

💡 Contextual Suggestion

The Contextual Suggestion UI is all about the AI enhancing your input. As you type, the AI suggests completions, ranging from a single word to an entire sentence. You can select from these suggestions, provide more input, or even delegate tasks to the AI.

A popular example? GitHub Copilot.

🤝 Can They Be Combined?

Absolutely! This fusion has been successfully achieved by a small team of developers at Cursor. For more insights, check out this post.

Permalink →