Machine in the middle
Large Language Models can’t start or finish anything. That requires a human. We’ve started calling it the 20:60:20 where the 60% is the middle part the machine can take on.
Within conversation theory there is the concept of intent1, as humans when we enter into conversation we have some sort of expected outcome, even if it’s only on an intuitive level. We reach the conclusion of a conversational unit once we’ve fulfilled that intent.
With human conversations the most successful interactions are those that have a clear expected outcome to both initiate and conclude a conversation. The same is true for Large Language Models you need to come with a clear intent that can be expressed clearly - the first 20% of the work - and you need to know when you’ve reached the limit of what the machine can help you with. At that point you’ll need to take it off, for the last 20%, to make it relevant for your needs.
I have a next js app that’s using tailwind css. I want to create a component called FeaturedCard. It needs to take a string (‘title’), number (‘price’) and string (‘category’) properties. The props need explicit types. The FeaturedCard has an image that is a snakecased version of the title. The card needs a 1px border and 4px padding, the image is above the title, the price and category need to be on a row at the bottom of the card. Please suggest some initial code that could allow me to achieve that with classes from Tailwind.
Can you tell me how I can get a website up-and-running.
Paul Grice, Erving Goffman, Noam Chomsky, Harvey Sacks and Erving Goffman all talk about this to one degree or another ↩