Reactive time vs proactive time

At work you can be reactive or proactive.

Before LLM-powered workflows, our reactive time was mostly spent reading inputs coming from other people. And arguably a lot of time was wasted because it takes time to switch context, and our colleagues might not ping us effectively.

You could have one colleague asking you to review a pull request and another asking you to review a design document for a future service you would implement. And both of these tasks are very different. And don't get me started on giving status updates.

Now that workflows can suggest things, we're in a place where we can react to the work flowing. And all of a sudden, this reactive time becomes extremely valuable.

Another benefit of LLM-powered workflows is that they can generate content faster than humans do. So you could be spending 30 minutes only approving one single step of the workflow. It's a bit like someone working in a factory. The assembly line is the workflow, and you sitting at the same place is the step you're approving or rejecting. If you stand up and go sit somewhere else, this can disturb the assembly line as a whole.

There is a LOT of thinking that has to be made about when to suggest WHAT to somewhat. Because the best "tinder experience" is one where the type of things you're being suggested are changing continuously.

If we keep the assembly line analogy: imagine standing up and walking 500 meters to do a completely different task than what you were doing.

For example if you approve an email about to be sent in the context of a given campaigns, it's probably smart to continue the campaign as much as possible, before switching to another type of work.

Don't be too scared about this distopian view of the world where we just sit and approve emails being drafted all day. Because if the work becomes repetitive, it's easier to automate than it is to automate manual jobs.