If you have integrated infrastructure, building, or rail transport needs, you've come to the wrong place. Here's what you're looking for: johnholland.com.au
If not, there's either a 1 in 1.3 trillion chance you've randomly stumbled here or you meant to visit. Either way, welcome.
I hire engineers for early-stage startups. Mostly software and AI people, mostly pre-seed to Series A, mostly Sydney and Melbourne. Retained, which means I work with a few founders at a time rather than firing CVs at everyone and hoping. The longer I do it, the more I notice that the easier something becomes, the more of the real thing it tends to take with it...
It's the same in every aspect of life really.
I can get a playlist assembled in seconds, full of songs manufactured to go viral on TikTok, but I still love stumbling across something I wasn't looking for in the bargain bin at the record store. If I wanted to, I could match with a hundred people without leaving the sofa, and be written off on a first date because I had spinach in my teeth. My wife has had to put up with far worse, but through that pain, we've gotten to really know each other. I can read a CV that says all the right things and was written, end to end, by a machine but the conversation with the person behind the machine reveals they can't actually do any of it. It's all rather convenient but it all slowly removes the bit that actually matters.
I don't think the answer is less technology. It's how I make my living — that would be ridiculous. But there's a difference between technology that opens you to the world and technology that keeps you from it.
I started playing around building some small things on the side, not really with any of this in mind. It was only as I got further in that I realised the same thread ran through all of them. Some are half-built, some will never be finished. But the point was never really the thing. It was learning how to build, so that when the right thing lands, I'm ready.
I've started writing some of it down too. Sharing thoughts tends to lead somewhere interesting, and it gives someone the chance to stumble across it. Treat this page a bit like the bargain bin in the record store.
This is where it all started. DOTS connects you with other people going to the same gig. You chat, connect over the music and meet up with a group beforehand so you're not standing there alone.
I built it because when I first moved to Australia I had no one to go to gigs with. So I started working on a solution. I pitched the idea to 800 people, interviewed ten. Everyone said they'd use it. I met with founders in adjacent spaces, spoke with Sydney Music. I still think someone should build it — it's probably my favourite project. To do it properly I'd have had to quit my job and the timing wasn't right. Maybe one day.
Like the idea and want to run with it? Happy to share how I think it could make money without losing its grassroots feel. I ran out of road, but would like, in some way, to help make it happen.
I started using ChatGPT to generate my workout plans and started seeing results. The problem was copying and pasting huge chunks of text on my phone and losing context every session. So I built something to fix that.
That got out of hand. I went down a rabbit hole of scientific studies into periodisation and training theory and that's what the logic is built on. I built the UI for me, but as I was doing it I thought I'd treat it as a real project and started thinking about UX that strangers might actually use.
Forge manages proper periodised programming. Year goals, six-week blocks, weekly structure that remembers what you lifted last time. The output is your actual progress. Something tangible in the real world. I'm rebuilding it again, properly this time, and might keep going if anyone other than my mates starts using it.
I still use it every day.
I spend most days having the same conversations with different people. Candidates: figuring out their USP, showing measurable impact, thinking in first principles for interviews. Companies: who do you actually need, what's your story, why should someone join you, and what does a process look like that actually secures that person.
I figured I could make something that gets the point across and have some fun with it. The tech can handle that part. The part where you get your foot in the door at companies that aren't advertising yet or find the people where the real signal is hardest to read — that's still where I come in. A fun project for me, a useful game for everyone else. If it leads to more human conversations, it's done its job.
Still only Chapter One but part of a bigger project. Built by someone who's seen the game from the inside, with the bits worth knowing pulled out.
Where I do my best work is having conversations with early-stage startups long before they need to hire and mapping out what's actually needed based on where they're going. Once a role is advertised, it's often too late or a lot of time has been wasted.
I built PROSPECT to surface startups showing forward hiring signals before any roles are posted. It scores them, keeps a live pipeline. The idea is to get me having real conversations with the right people before they know they need to hire. Completely personal use, built in evenings, runs for about a tenth of what the off-the-shelf equivalents cost. There's a worldview point in there about what software actually costs to make, but that's a separate conversation.
This is the long-term one.
My wife and I have been talking about owning a piece of land and putting people on it for as long as I can remember. I tend to get excited by things and lose interest but this one has stayed. Maybe semi-retirement. Maybe sooner if the right opportunity comes along.
I'll probably use tech to build and market it. But this is the 100% get-away-from-tech idea.
I'm not a software engineer.
The things on this page were built in evenings and weekends over the past year with curiosity and tools that weren't available a couple of years ago. None of them are products in any real sense. They're experiments and some of them work. All of them have led to genuine insight about the industries they touch, which was mostly the point.
The reason I keep building isn't about the tools. It's that having to make a thing actually work forces you to be specific about what you think. You can't write a half-formed idea into code and have it run. It's partly why I'm writing this now — the discipline reveals where the thinking is missing.
Maybe in my lifetime, we won't be recruiting humans at all. Maybe next year we won't. Who knows?
But right now, humans are still the heart and soul of a business and while that's the case it feels quite important to still be having human conversations.
For the record, since this is technically the part that pays the bills: twelve years doing this, a thousand or so CTOs and engineering leaders in the network built up over that time, and a habit of getting to the right person before the role officially exists. That's the part that's worked.
iiMost companies would like to automate conversations away and just have the perfect human turn up on day one. The idea is to install tools that pattern-match CVs against job descriptions, both of which are now written by language models, and then have an AI interview a candidate using AI to answer their questions.
The problem is, the easier signals to collect are the weakest, and the strongest signals are the hardest to get at — things like whether someone can think under pressure, what they chose to protect when resources were tight, what their peers say when nobody's listening. The people who do well in these areas often don't shout about it online, there is no data to collect other than speaking to other people with these traits and asking who the people are that they rate. It's a spider's web built on trust with little to no digital footprint.
It's a spider's web built on trust with little to no digital footprint.
It's also why the recruiting work I find genuinely interesting is the part the tooling can't touch (yet).
iiiI will literally automate everything that can and should be automated. Some things can't and some things shouldn't — and those are the reason companies still pay me to find their humans.
So every day I lay another piece of my own job on the altar, an offering to the gods of automation, in the devout hope that if I sacrifice enough of myself they'll one day grant me passage to utopia — universal high income, endless weekends, never a CV again. Until the rapture comes, I wake up grateful there's still some small thing only a human can do.
He spends a whole page moaning about tech hollowing things out, then uses a machine to help write it.
Guilty. Sort of. I didn't ask it to write the thing, I asked it to help me sieve through a year of thoughts scattered across models, notes and 1141 screenshots. Then I read what came back, felt a small flicker of despair at how confident and lifeless it was, deleted the lot, and wrote it again myself. If nothing else it showed me what I didn't want to say. A service for which I remain, grudgingly, grateful.
Same goes for the CV at the bottom of this page. Use AI to write yours — I have. Just remember there's a moment, about forty minutes into an interview, when a real human leans forward and asks you to talk them through the five years of experience you claimed in turnip.js, and no machine on earth can help you then. You'll be entirely alone.
Anyway. Here's my CV. I've spent this whole page explaining that the thing is dying. Make of that what you will.