~32,000 tokens versus ~7,000 tokens. Illustrative example.
Illustrative example
Tokens
A recording can replace a long written prompt, so the bot reads less before it starts.
LIT is a Mac app. Talk through your screen and send the recording, with its transcript, to any bot.
Your bot can do the work. Getting what's in your head into its context still takes effort. Show your screen and talk through it while you're there.
You notice something, know why it matters, and already have the screen open. Hold Option+Shift+L and talk through it. You don't have to stop and turn the whole explanation into a written prompt.
The exception changes what to do. Your priorities explain why. Talk through those details alongside the screen that makes them concrete.
A spoken “this” only makes sense with the frame it was said against.
LIT pairs spoken phrases with screen frames from the recording. Your AI bot can retrieve the transcript and visual context together.
Message to the bot
Recorded screen
Customer request
Send the replacement today. Use the original order for payment.
The bot still has to do the work. LIT gives it more of the context you used to explain it.
Where showing the screen saves a long written prompt.
~32,000 tokens versus ~7,000 tokens. Illustrative example.
Illustrative example
A recording can replace a long written prompt, so the bot reads less before it starts.
Writing it out~45 min
Showing it~3 min
Illustrative example
Writing out every step takes time you don't have. Showing the screen is the calmer way to teach it.
Lesson
Bot's work
LIT checks the submitted work against the lesson. The bot does not grade itself.
| What does LIT stand for? | LIT means Last Inference Training. LIT is a Mac app that makes a screen recording with a transcript, which you send to your bot. |
|---|---|
| Which bots can I send recordings to? | Grok Bot, or any other AI bot you connect. Choose it when you send a recording. |
| What does my bot receive? | A lesson link it can use to retrieve the transcript and associated screen frames. Those frames provide visual context for the spoken explanation. |
| Is this just dictation? | LIT records your screen alongside your voice. You can point, move through the interface, and explain why something matters while you're looking at it. |
| Can anyone with the link read my recording? | Access is authenticated. A lesson URL alone does not grant access; the receiving bot needs permission. |
| Does sending a recording mean the bot understood it? | No. LIT supplies the explanation and visual context. You still need to review the work the bot produces. |