LLM outputs, indexed
A vehicle for logic
What a model’s output can become, from a read-once answer to a tool others build with.
Others use the underlying system to do an operative task, or a number of tasks, or to come up with new tools.
- Best used for
- Operative, repeatable tasks
- Running a number of tasks
- Making new tools
- Who interacts
- Operator, builderUses it to get work done, or builds on it
- Made of
- Code
- Data
- Any as input and output
Others can interface with it and explore different scenarios on a specific idea or problem. It allows for a wider spectrum of messages.
- Best used for
- Exploring scenarios
- One idea or problem, seen from many sides
- Messages a page can't hold
- Who interacts
- ExplorerChanges inputs and sees what follows
- Made of
- Text
- Image
- Sound
- Code
A fixed record that holds up on its own, outside the conversation that made it.
- Best used for
- Strategic decisions
- Logs of events
- Rules, or a set of rules
- Who interacts
- ReaderReads it, cites it, follows it
- Made of
- Text
- Image
- Tables
The plain reply in a chat. Read once, then it stays in the conversation.
- Best used for
- A quick question
- A one-off explanation
- Who interacts
- YouOnly whoever asked
- Made of
- Text
- Image
Select a level to open it. Scores run from 1 to 4.