Week 32 of 2026 - Ambition with AI

It’s insane to think that it’s now August already. Ireland has been through a July drought, all but confirming to the last holdouts that indeed, climate change might actually be real. And importantly, we’ve well-and-truly crossed the threshold into the second half of the year. Q3 is in full swing for companies; kids are winding down the weeks to school’s inevitable return; and adults are getting the best out of their newly acquired holiday attire before it’s retired for twelve months, upon which time you ask how you fit into it just one year prior!

But this post has nothing to do with that. It actually stems from a text chain I had with someone also in the tech world. We had a good aul natter about the mysterious thought processes of our billionaire overlords, as well as the general sense of direction tech has taken. It’s no longer fun, quirky, friendly or solving real issues. It’s just big, often antiquated, and more often than not, a bit shit.

And while AI firmly earns it’s place in the midst of this mire, it also unlocks a lot of interesting use-cases. Speaking purely from a tech perspective, I’ve vibe-coded an entire platform quietly for a few months that I’m quite proud of. It’s not prime-time ready yet and I’ve not done much to get new users, but it is excellent given the relatively short period of time I’ve spent on it. And other things I’ve done with AI have also been either really quick techie things for myself, or wildly ambitious like the aforementioned platform.

AI, in the hands of weirdo nerds, is excellent.

I remember being at a meetup for tech weirdos about 20 years ago. It was a regular event, every month, where tech weirdos met up and just had a few pints. Ostensibly paid for by a stalwart of the Irish tech scene as a pseudo philanthropic endeavour to build a community in Dublin. Before everyone in that audience wound up in huge jobs paying obscene salaries. We needed that pint, because we were loners otherwise. And because we couldn’t really afford pints because our salaries were pathetic.

On one such evening, I was talking to 2 Microsoft engineers. If I recall correctly they actually worked on the Office 365 suite, before it was called anything remotely like that. And as a result of my getting a job in a high-growth startup that explored hiring engineers in Ireland, they wound up coming to be my colleagues a few years hence. Which is why the power of community is so important, but we’ll come back to that later. That conversation stuck in my mind because we were talking about AI, and automation. In very loose terms. It was 20 years ago and we were probably still mourning the loss of Clippy.

But I noted that engineering is such a powerful job at the time. Engineers held the keys to so many doors. But that wasn’t right. Engineers are rarely the ones who should be judge, jury and executioner when it comes to building experiences for the average human. And at some point, computers themselves will get pretty good at building software for themselves. Ideally with a level of sophistication and understanding of the bare metal of what’s happening that the sheer amount of efficiency will appear as magic to us mere mortals. Computers, one day, will have the same difficulty explaining to use idiot humans how it’s achieving things as we would trying to explain electricity to a chimpanzee.

I was wrong on a lot of this. We now have “AI,” but we’re a long way from having grains of sand think. It’s not particularly efficient of sophisticated. It’s just a conglomeration of all human existence (print media, books, white papers, the internet) shoved onto RAM and given fancy bubble sort. It’s just T9 predictive text with more bells and whistles, really.

But the AI we have, when applied correctly, does unlock a lot of ambition. Automating little tasks isn’t the juice worth the squeeze. We could already automate a lot of tasks on phones and desktop OS’ before; most people just didn’t pay attention to those features. But now when you add “AI” to the name, a level of pizzazz or showmanship enters the room that would make Monaco blush.

If you leverage Claude to code something big, bold and ambitious, you have my attention. Go build cool shit. Make it work properly and deploy it with real security checks. But let this weird iteration of AI unlock ambitious thinking. So many programming frameworks over the last 20 years, since my fateful conversation with soon-to-be former Microsoft engineers, have made engineers and product teams so much more efficient that they could think bigger. AI should have the same impact on folks.

And if that’s what we use AI for, rather than making fake comedy skit videos with hyperrealistic graphics on our social media doomfeeds, then these morally corrupt leaders won’t need to consume so much RAM that a Steam Machine games PC doesn’t cost the same as a car deposit or drain so much resource that it advances climate change aggressively.