A few months ago, I overhauled our contracting infrastructure to support AI generated contracts, DOCX templates that fill in the things we’d need, and correct-but-unsupported DOCX detection1
Two years ago, taking on a big daunting project would’ve been a measured decision. AI has made building things faster. That changes the calculus of what projects you can take, and how ambitious you can be with your bets. A big overhaul doesn’t take weeks. It can be done overnight, which means you can experiment much faster.
This is all true, and lots of people have talked about it. I don’t want to talk about it here.
What I do want to talk about is the unexpected surprise of how AI has made me braver.
The fear hasn’t gone anywhere #
Every time I start work on something ambitious, there’s a knot somewhere in my tummy. I call it the fear of the unknown. Not knowing exactly how I’ll do something leads to uncertainties. And uncertainty is … uncomfortable.
That hasn’t changed. I don’t think it ever will.
It’s something I’ve grown to recognise and take in stride. Working with AI over the past year though has made this feeling 10x better. The discomfort is minimal because I’m extremely confident in getting out a shitty working version thanks to AI.
And once I have a working shitty version, the uncertainties are gone, and I’m just a few iterations away from a working wonderful version.
AI compressed the distance between a scary idea and a working prototype.
Repeated exposure #
In an earlier post talking about this feeling, I mentioned how repeated exposure to new and uncertain ambitious projects makes me more comfortable dealing with them.
AI speeds up that exposure loop, because huge projects take 10x less time, so I’m doing far more of them.
I have more evidence that I can handle whatever comes up. I’ve watched myself build things I thought were too hard, over and over, and each time the gap between “this feels impossible” and “okay, it’s working” got shorter.
Bravery #
The contracting overhaul was just one example. There have been several more over the past few months, from cost attribution & tracking across all organisations we work with, to a new payment processing infrastructure, to cheap content monitoring 2, to another top secret big project I’ll write about in a future post.
I say yes to a lot more projects. I feel good knowing I’ll get them done this week 3. I can take on more challenging projects, and not break a sweat.
I love working with AI.
Calm under fire #
A second order effect I didn’t expect is staying calm when things go wrong.
I think it comes back to uncertainty + time pressure. Things are broken, so you want to fix them ASAP, but also you don’t always know what’s broken. This is stressful.
Now when something breaks, I still feel the urgency. But the fix cycle is so much faster that the urgency doesn’t compound into panic. I can have an agent look at logs, traces, and propose what’s wrong, and I can easily verify it. In most cases, verifying the issue is far easier than figuring it out.4
Looking back, being productive is nice. But imo feeling braver matters more.
Did you know there are legit DOCX to DOCX converters that convert google docs DOCX into microsoft word DOCX? And some services refuse to operate on one or the other. Insanity.
↩ - This will also get its own post soon. But the TL;DR is how do you cheaply analyse content created by 10,000+ creators daily to find your brand mentions.
↩ - Not all of them of course, hyperbole.
↩ - Classic P vs NP problems.
↩
You might also like
How I'm Productive with Claude CodeAgentic DebtWhat I learned about burnout and anxiety at 30How to setup duration based profiling in Sentry