At What Cost?

Let's say you are at work and you're using PowerPoint and you have Copilot open, and you ask it to do a simple task. Like turning a slide into a divider slide. I asked because the presentation did not have any divider slides in it and I needed a divider slide format.

What will happen is that Copilot will spend 15 minutes, and per it's own words, it will map it out, look at the details, evaluate the request, explore different approaches, work through it, think it through, weigh its options, work on it, make the updates, review it again, and then give you a slide that looks exactly like the title slide. All in 7 minutes.

What is hidden from the user is the cost of the prompt. When I'm burning through tokens for personal use, I can always go look at Google AI Studio and get an understanding of where I'm at. But when you're working the costs are all hidden, which makes silly queries like this more common.