The next chapter

Chris Barlow

07 Nov 2026 — 4 min read

AI promises intelligence on tap and a brighter tomorrow. So far it has mostly delivered intelligence on tap, fear and incertainty, and a lot of posts on LinkedIn and X.

The brighter tomorrow is the bit we're working on.

Marc Vincent and I started Agentics NZ in August 2025. Fifteen months later we did something radical for a tech community: we asked the members what they wanted. Thirteen of you filled in our anonymous survey (thank you, all thirteen of you). On 5 November we spent 90 minutes online working out what it all meant.

This is my quick take, with some history added. The proper version is Chris Gagne's Agentics NZ wayfinding: what members said, so please go read that.

How we got here

My first aha moment (like many of you) was ChatGPT in 2022. I've used large language models every day since mostly to work out how to apply these super powers in my work, to delegate things I don't like doing, and to find out where they break.

Along the way I picked up some excellent humans. I met Dr Simon McCallum in a lift on the way to a Victoria University AI in Society talk. Possibly the nerdiest lift ride in Wellington history. A lasting friendship later, his six animal model became Campaign Mode. Dr Paul Duignan, fellow meditator and creator of DoView, has been talking AI strategy with me for years now and has yet to tell me to go away.

The second aha came in late 2024, when I tried Lovable. I didn't code until coding agents arrived. Turns out I had loads of ideas lying around that I could suddenly make real. I was hooked.

In early 2025 there were no experts in this stuff, and I was hungry to learn. I went looking for my tribe and found it at Reuven Cohen's mind-blowing vibe-coding sessions, run by his Candian-led Agentics Foundation. Fellow-kiwi Marc Vincent discovered me in the discord chat. In August 2025 Marc and I started the New Zealand chapter.

When a field is very new, the best way to learn is to learn as a group, sharing what does and doesn't work. The pitch for Agentics NZ: AI needs fighter pilots, and formal training isn't keeping up. A hardcore group of enthusiasts turned up, which is either very encouraging or a sign we all need hobbies.

Later, triggered by the release of the latest/greatest model of the moment, Fable, came Unsorry. A swarm of agents run by members, turning unproven Lean statements into kernel-verified proofs. It went from idea to movement more or less overnight. The recipe was simple: a hard problem worth solving, a dead-simple way to join in, and a leaderboard (turns out everyone loves good-natured competition).

Who we are

We're builders, hackers, thinkers and leaders. Plenty of us are experts in something else entirely, like psychology, law, delivery or transformation; each of us seem to share a habit moonlighting as agent wranglers. We're earnest (sorry). We want to help our fellow Kiwis, we want New Zealand to reach its potential, and we like hanging out near the frontier.

Chris Gagne's write-up has the detail. Two themes stood out. We have energy and community in spades. What we're missing is focus: one clear purpose, and a win big enough to rally around.

So here's where we landed: we build in the open, solo and together, in service to Aotearoa New Zealand.

Our role

We have two jobs.

Education. Leadership, direction and support for people learning to work with agents. Think flight school, minus the crashes. Ideally.

Protection. Looking after the success and wellbeing of New Zealanders along the way. Trust, jobs and sovereignty matter as much as the shiny stuff. If you want to go deeper on sovereignty, Tom Barraclough's sovereign AI policy work is a good place to start.

We're better equipped than you'd think. There's a frankly unreasonable amount of talent in this group, a pile of open source tools already shipped (see the showcase), access to sponsorship, venues through our networks, and some serious hardware: a Strix Halo machine and a DGX Spark, both humming quietly and judging us. Between us, we also know a lot of people, some in high places.

My approach hasn't changed much. Listen. Lead by example. Doing beats talking.

What's next

One flagship project that solves a real New Zealand problem.

We're starting small. Two members are leading our first hackathon, online only, as a possible first step. No travel, no lanyards, just building. We're aiming for a hackathon every quarter.

If you've been waiting for a sign to get hands-on, this is it.