I’ve been running Design Sprints for years since the method was first introduced as a book.

One practice that really stuck with me is “Always Be Capturing”.

The concept is simple: as the facilitator, you capture ideas on post-its or whiteboards the moment people share them. This small practice has some great side-effects. It saves the in-the-moment discussions and thoughts, shows the participants their ideas are important, it creates a visual summary that we can use as a “was this what you meant?” and probably the best part is that it creates a collection of ideas that we can organise later.

I use this during conversations, while listening to audiobooks, or podcasts. I’ve set up a quick shortcut on my phone that opens a text area and saves notes directly to my favorite note-taking app.

The beauty of this system is its simplicity. Capture first, then group, filter, and remix later.

No field has stayed unchanged in the face of new technology. It either transforms or eliminates jobs.

Don’t wait for perfect to start.

Creativity starts after your first idea ends.

The cost of entry for AI is lower now than it will ever be again. If we only use AI for quick answers we’ll struggle to justify higher costs in the future. Those who use AI as a thinking partner will see the value differently. They’ll happily pay increased prices because they’ve experienced AI as an affordable teammate rather than just a fancy search engine.

Every piece of text you write adds weight to the story you tell.

At some point that weight makes the story complicated and hard to understand.

Weight becomes the blocker for advancing. You have built debt.

This applies to generating code with AI too.

It’s powerful but can produce code in wrong places, fix unintended problems and make assumptions about what to write.

Just because AI can create a lot quickly doesn’t mean it’s creating what we need.

The real question isn’t “how many hours did you work?” but “did you deliver the value promised?”.

If the client is happy with the results then you’re delivering what they’re paying for.

Hourly billing punishes efficiency and expertise. Outcome matters, not the time spent.

People rarely enjoy all aspects of their job equally. This is particularly noticeable with fullstack developers who are expected to be proficient across the entire stack. From transforming Figma mockups into functional UIs to implementing complex architectural services. The range of skills required is… let’s call it extensive.

At some point the industry decided fullstack was the ideal. You should handle everything from pixel-perfect CSS to database optimization because… it’s efficient? More cost-effective?

Outside of solo entrepreneurs I’ve met remarkably few developers who genuinely enjoy both ends of the spectrum equally. Most have clear preferences. They either light up discussing UI interactions or get excited about backend architecture but rarely both.

This preference gap can explain why tools like vibe coding are gaining popularity. These tools reduce the parts of development work people enjoy least while still getting the job done.

You can build great things that feel effortless now. That simplicity exists only because you invested hard effort earlier.

Like grinding through side quests in a game before tackling the main storyline. All that seemingly extra work makes the final boss battle feel surprisingly manageable.

Tech influencers often preach ‘If it’s easy, it’s probably not special.’ But they miss the timeline. What feels easy today only seems that way because you already did the hard work yesterday.

I’ve found AI actually demands a different kind of thinking from me.

When I’m vague, it gives vague results. This forces me to clarify my own thinking and be more precise.

The cognitive work doesn’t disappear, it shifts to higher level of design and critical evaluation.

The challenge is using AI intentionally rather than letting it replace our thinking entirely.