It’s increasingly likely that when you ask for feedback or thoughts about something at work, you will get something that a robot wrote in return.

At a basic level, I get it. It feels helpful to talk your feedback into ChatGPT or Claude and have it turn your ramblings into structured thought. The problem is that it’s simply too much information. The models are there to produce accurate output, which means they won’t take the risk of being confident about something that may be wrong.

There’s a problem with this because it has a ballooning effect. Before LLM’s, feedback from a colleague would help refine work that was being done. The purpose of asking “how does this look?” or “is there anything you would change” was to get closer and closer to the bullseye of a good output for that project or task. Each iteration, meeting, or conversation about a project should ideally make it better,meaning that we are closer to our goal.

Despite the promises of increased levels of productivity, the practical reality on the ground is that there are too many tokens being created for us to turn into anything useful.

This gets at the heart of why producing good work is actually hard: getting toward the right outcome over time means saying no to all sorts of other outcomes, even though they may be good ones.

The experience of feeling like a project is “going around in circles” is common, and usually the reason this is happening is because we aren’t willing to say yes to one thing and no to all the others.

You can’t move forward without leaving something behind.

This is why humans are great: we take risks. You may not think you’re a huge risk taker, but in the smallest ways each day we risk our reputations & our futures when we make decisions that move work forward.

You could be wrong about the direction for that new campaign. The design concept you thought of may not be the best option for the client. Every thing we commit to is a potential failure, and that’s why it’s hard to commit.

The other side of it, however, is that the risk comes with possibility.

It also might work.

The dance of risk is the human work.