Have you ever found a tool that looks so cool and promising that you wish you had a problem, or excuse, to play with it?
This happens a lot with AI nowadays. How do we find balance?
You have decided what you want to do or maybe who you want to be. Yet, you are stuck researching the perfect tool because you instinctively believe doing so will prevent you from failing. Today I’ll share some ideas on how to invite AI to your workbench without getting lost in the sea of possibilities.
The key is to approach your problems as a scientist. So you rely more on iteration and experimentation rather than dogma.
Back in 2025, I was planning the development of my first iPhone app. I completely fell into the shiny object syndrome. I already had everything I needed to get started. However, I started checking tools for customer support and community building. I spent time and energy comparing the best tools.
It was exciting to think about all the possibilities. However, to this day I haven’t used those tools. At this point the good old email inbox works just fine. I was solving a problem that didn’t exist.
One year later, thanks to AI there are even more possibilities. From well-established tools, multiple vibe-coded tools, and DIY tutorials.
The shiny object syndrome
The shiny new object syndrome is procrastination disguised as preparation. It feels more satisfying to research and learn about the greatest and latest tools than doing the actual work.
I believe we fall into this trap because dreaming is fun and the possibilities are endless. On the other hand, doing the work is about facing your limitations, recovering from failures, and iterating.
To move away from the shiny object syndrome:
- Notice when you are falling into the loop
- Keep in mind the bigger picture, what is the real problem you want to solve and your intention? What needs to be done right now, instead of at a future phase?
Pick the best tool for the current problem
Once you have your intention clear, you can focus on one problem and pick the best tool you can find at the moment. Keep in mind AI is not always the answer, sometimes you need a tool that’s more deterministic, other times you need a mix.
Here are some areas where you can use AI for experimentation. Remember, the goal is to become the best expert in the world in using AI for a task you know well:
- Decisions
- Learning something new
- Boring, Repetitive Work
When to switch or upgrade tools?
Once you have picked the tool and established the workflow, commit. And for that purpose ask yourself “When is it worth changing to a new tool?” Depending on the use case:
- When hitting a wall: Ideally you’ll find out if the tool matches your needs while testing it. Experiment for at least 15 minutes, and avoid tools that lack a track record.
- For major benefits: AI is a competitive field, maybe a new tool comes out and it’s cheaper or more efficient than the one you are using. Experiment for 1-2 weeks to confirm if making the permanent switch is worth it. Remember that if the feature is a game changer, your current tool might catch up in a couple of months or less.
- After a periodic check: Decide how often you want to evaluate other tools. For instance, David Sparks(productivity expert and podcaster), commits to his task management app for one year. This way he doesn’t fall into dogma but is not constantly making time-consuming changes either.
How to get started?
If you are unsure on where to start, I created a cheatsheet organized by problems you’ll probably recognize, a prompt you can copy for each, and the stack I’m actually using.
Remember, there’s a clear difference between looking for an AI tool to solve a specific problem, versus trying all the new AI tools with the hope a problem/solution will materialize along the way.
Use AI intentionally. You don’t need to try every new tool or model. You’ll gain relevance and credibility by doing the work that matters empowered by AI.
This is part of my “Principles for Working with AI” series. See the first part here: AI FOMO: You’re Not Falling Behind.
Leave a Reply