When you accidentally Logic your way to best practice App building

I’m in the middle of building my app and I have pretty much all of the rough pieces of the backend and front end in place.

Refinement is needed but it’s definitely starting to take shape.

I’m rate-limited by the coding agent I’m using to build the app. For anyone that doesn’t use AI to build stuff, rate-limited means I’m on a plan that has a certain amount of usage every few hours and I’ve hit my limit.

I’m (with my Mrs) running a busy household and I cannot justify the 10x price increase to upgrade my plan and get more access.

So I sit tight and wait until I can go again.

Tonight, (let’s call it a nondescript Friday night) my rate limit has another 45 mins to go until I can continue working - so naturally, I’m scrolling on X.

Nothing is catching my eye until I scroll past a tweet (are they tweets or x’s now?) that says:

“New @GoogleResearch paper builds a personal health assistant that reads a users data, answers health questions, and coaches daily habits…”

“Jeepers, that sounds a bit like what I’m building…”

Furiously begins to read in case the idea I’ve had is already done and done by a behemoth of a company….

Nope. My idea is safe.

Phew.

And this paper is actually geeky cool and super interesting.

Then an idea strikes like a bolt of lightning.

I download the 149 page paper and upload the file to the AI project I use to plan my app build.

So I do what every person in 2025 does.

I ask the LLM to compare what I’m building in my app (and how I’m building it) to what Google's Deepmind paper has suggested is the most effective way to approach building a personal health app.

I figure that I’m going to learn something about where my app build is weak and how I can improve it. That’s a win in my book.

I’m waiting for the AI system's response…

This could go either way 🤷🏻‍♂️

Holy handlebars. Let me read this again.

So according to this AI system, in some areas I am actually ahead of what Google's researchers have said is best-practice and it also called my chronic pain focus “top class”.

I don’t know about you, but after reading that I’m tempted to say that I’ll give any Cheshire Cat a run for their money in a “who’s got the biggest grin” competition? 😃

Let's climb down from the journey to the clouds here I tell myself.

These are just words on a page that an LLM has generated. I don’t even have this app built let alone have a user using it yet...

Can you hear that whooshing sound? That’s the sound of an anvil (me) that’s just been dropped out of a plane, plummeting back to reality and landing with a thud.

That feels better.

I often talk in clinic about finding the signal within the noise.

For a person with pain, that can be the hardest thing to do by themselves. That can be the hardest thing to do for me with someone - and I’ve been doing it for a living for 15 years!

This feedback seems like some signal within a lot of noise in my head.

Maybe I’ve somehow built something that is architecturally and technically good?!

The mad thing to me is that I’ve built this simply from 3 things:

1. a deep personal understanding of living with chronic pain,
2. experience drawn from thousands of hours of helping people who live with chronic pain
3. pure use of this Irishmans logic.

I allow myself to feel something approaching a sense of pride for a few minutes.

Make a cup of tea.

Rate limit is up in 3 mins.

My mind snaps back into work mode.

I put that little piece of signal in the small pile in my mind that represents “this thing might actually work out well David” and I get back to it.

The grind that is building an app continues….