If have ever tried to write anything, you have probably run into the Writer’s block. Or suffered the blank page syndrome: you want to write something, but you keep staring at an empty page, unable to decide how to start. You may even have a clear ovearll idea of what you’d like to convey but can’t get it rolling. Those first few sentences are too hard.
I’m experiencing the same thing with agentic development. I have all the (agentic) power at my fingertips. And many things I’d like to build. And yet, I can’t find the time? energy? willpower? to get started.
Let’s call it blank prompt syndrome.
It’s indeed very similar to the blank pagee one. Once I start the conversation with the agent, explain the goal of my idea, define some basic architectural constraints,… all the subsequent iterations become really easy. Friction disappears.
Especially when you reach the point of being able to try a first prototype. It helps a lot to visualize that you have something. And place yourself in an iterative improvement loop. You have something concrete to react to.
And, most often, you can do it without being fully concentrated on the agent. Any small chunk of free time is good enough to ping the agent and put it to work. You could even do it during a boring meeting. Finally, a way to make some meetings productive 😉.
There is another parallel with writing that becomes useful. A common piece of writing advice is to separate writing mode from editing mode. First, get your ideas onto the page. Allow yourself to write something imperfect, explore a direction, and change your mind. Later, come back to improve the structure, remove what does not work, and polish the details. If you try to perfect every sentence before writing the next one, you may never finish the first paragraph.
I think we should give ourselves the same freedom when working with agents. Start in exploration mode. Describe a small problem and ask for a first attempt. Try it. Ask for a different version. Discover that half of what you thought you needed is unnecessary. Or that the useful part is something you had not considered at all. The first version can help you discover what you want. You should not have to know every requirement before you allow yourself to build anything. And even less look at the quality of the code being generated.
So my advice is to reserve a little focused time for the initial push. And then iterate to maximize the productivity when you are tired of other things or, as I said before, in meetings that could have been an email. Otherwise you’ll feel stupid twice: once for being in the meeting and twice for not doing anything useful during it.
Blank prompt syndrome is the perfect name for what happens when the cursor outwaits the brain. Your opening line about writer’s block hit home — every draft I’ve ever abandoned began with a perfectly blank page, not a shortage of ideas. What I appreciate here is that you treat the empty state as a syndrome rather than a personal failing; that reframing alone lowers the pressure. Lately the same freeze visits me at the chat window: the blankness simply moved from the page to the prompt box. Does the cure transfer too, in your experience — freewriting into the input until something coherent surfaces — or does the machine’s instant eagerness make us self-censor even earlier? Either way, naming the thing feels like half the battle. Saving this for the next stare-at-the-screen afternoon.