Korva is a scroll of the internet’s best ideas.
One product, huge range
Instagram reached billions of people with a product that had a couple of ways to use it. ChatGPT covers a far wider spread: people brainstorming their next haircut, people helping to automate their farm in Japan, and a range from users who pay nothing every day to people paying hundreds of dollars.
Capability overhang
Most people get only a sliver of the value the product could give them, an idea called capability overhang. They still get tons of value, and the answer is an extremely simple, streamlined experience for the many, with the cutting-edge features easy to reach for the people who care. New things are tried on that group first, through the desktop app and Codex, and ChatGPT Work is now stepping out to a much broader audience.
An interface that adapts
For billions of users, the product only truly works when nobody has to think about a switch or a mode, so whatever proves out has to be distilled into the main experience. The interface also has to shape-shift. A first-timer needs to see how it adds value to their life or work, and a daily workhorse user wants it out of the way.
Why do most users see only a sliver?
A general-purpose tool offers no menu of what it can do, so people use it for the tasks they already imagine. New capabilities arrive faster than habits form, which leaves a gap between what a product can do and what most people try. Early adopters discover new uses and share them, and those uses spread outward slowly. Example prompts, templates, and suggestions are common ways products try to narrow that gap.
How do apps serve beginners and experts together?
Designers call the approach progressive disclosure: basic options appear first, and advanced ones appear only on request. Word processors are a classic case, where beginners use menus while daily users rely on keyboard shortcuts. The same program serves both groups because the advanced layer stays out of the beginner's way. The term was in use in human-computer interaction research by the 1980s.
Why do modes confuse people?
Interface researcher Jef Raskin argued in The Humane Interface, published in 2000, that modes cause errors because the same action does different things depending on a hidden state. Researchers call these mode errors. A well-known example is typing in a text editor that is in command mode instead of insert mode. Raskin held that an action should do the same thing every time it is performed.
More from Lenny's Podcast: Product | Growth | Career
