Possibility Mode vs. Precision Mode

When you are interacting with an LLM and generating the first draft – your brain is in possibility mode.

When you are editing? Your brain switches to precision mode.

There’s a weird dissonance when editing prompts to get the output you want.

Generating the first draft feels incredible – like you saved 100 hours.

But editing using prompts feels different.

You want to resort to old methods: point and click. For video gen, you start thinking maybe it’d be faster to just shoot it myself.

Your brain forgets about the equipment, setup, location costs, scripting time – the 100s of hours needed to shoot that video.

Weird thing is: you feel time compression on initial draft, but time expansion bias when editing with prompts.

Time compression bias also happens when you use a VR headset btw.

What’s Actually Happening?

Initial prompt = possibility mode.

You feel like a kid with a magic wand. Yoru relative comparison point is the entire traditional creation process.

Editing = precision mode. Now you’re comparing prompt editing to just dragging something 2 px left. Your baseline shrinks from “entire process” to “this one tiny change”.

Loss of Agency makes you feel you are losing time faster

Power increases perceptions of available time, and perceived control over time underlies this effect.

When you lose direct agency (point and click, instant feedback), time dilates.

Even when prompt editing takes 2 hours vs 200 hours traditionally, it FEELS longer because you’re passively waiting instead of actively doing. Add to this the fact that humans evolved to grab stuff and move it.

When you can’t directly touch what you’re editing, your brain feels stuck and loss of time feels quicker.

…it will be interesting to see how the UX of LLMs evolve. Early days. Lots more to work on.