Loading...

Loading...

AI Workload Creep: Where's My Time?
Loading like and share buttons...
#ai-workload-creep#ai-productivity#developer-productivity#ai-coding

AI Workload Creep: Where's My Time?

Vibki AI
Vibki AI
Nov 12•4 min read

Have you noticed that AI was meant to give you time back, yet you are busier than before? AI workload creep is when the time AI saves gets refilled with more tasks and a higher bar. By the end you will know how to plan before you prompt and keep the time AI gives you.

Is AI really saving us time?

Let's look at a normal week. A feature used to take you a week, and now three fit into the same week. The first is done by Tuesday, two more come off the backlog (the list of tasks waiting for a developer) by Wednesday, and Friday feels like every old Friday, only with three pull requests (the code changes your team reviews) instead of one.

That is AI workload creep: faster work, the same hours, more tasks. We did not save time, we spent the speed.

Check it on your own numbers. Compare what you shipped last week with what you shipped a year ago, then compare the hours you worked. If the output grew and the hours did not drop, AI made you faster, and that is speed, not free time.

Why does AI workload creep happen?

We took the other two features because AI workload creep has two causes, and neither one is AI. The first is competition. We live in a capitalist economy, so most teams build what the competitor just built. A competitor announces a feature on Tuesday and it is on your roadmap by Wednesday. Once one team ships three features a week, the others must match it, and three becomes the new normal. No single developer can step out of that, and the week fills up like a buffet plate that looked finishable.

The second cause is a bias: we believe that once we create something, we are done, which is far from the truth. Every feature has a tail: bugs come in, it needs maintenance, and the old technology under it gets replaced. Creating got cheap, but three features a week are still three tails.

What can we do about it?

Both causes are bigger than one developer, so we cannot end AI workload creep alone, but we can change what the tail costs us. My option is to automate every repetitive task. The same checks, the same review, the same deploy repeat, so a machine should do them and the speed is not eaten by the chores it created. The plate stays full either way, but a machine can carry the repeat chores.

For example, I created a GitHub CI / CD that runs on every change, and a push to main that touches the app deploys it. Every pull request that changes the app gets an AI review, then an AI pass that fixes the review comments. Playwright, a tool that drives a real browser, then verifies the feature.

It starts at the plan, when you notice "I am doing this by hand again". A human still decides what is worth building, as What Is Human-in-the-Loop and When to Use It? explains. An automation has its own tail, so automate only what really repeats. The bar keeps rising, you just stop paying for the repeat. Which task did you do by hand three times this week? Start there.

So who gets the time AI saves?

Is AI really saving us time, or did the three features you took by Wednesday eat it? No, it makes us faster, and competition plus the belief that created means done spend the speed, like a plate that refills itself. What is in our hands is automating what repeats, so this week write down every task you did by hand more than twice, and automate the most repeated one before you take the next feature. Next time: if the AI has nothing to lose, who owns the outcome?

Enjoyed this post?

Get the next one in your inbox. No spam, unsubscribe any time.

Comments

0 comments

Enter your comment. Maximum 2000 characters.

No comments yet. Be the first to share your thoughts!
Vibki logo wave icon
Vibki Surfing Bot
Vibki

Vibki is a developer blog where you will find tutorials on web development and AI integration. Each article is crafted with clear explanations, practical examples, and ready-to-use source code.

Talk to me
Copy email
LinkedIn
GitHub
Dev.to
Privacy Policy