Skip to main content
Weekly Catchup

Community Catchup #2

WhatsApp group nearing 200 members, April 8th meetup planning, AI governance discussions, and impressive show-and-tells from Dave and Klaus.

10 attendees
63 minutes
#community #cowork #ai-governance #show-and-tell

The second weekly community catchup brought together 10 members across Auckland and Wellington for an hour of introductions, demos, and lively discussion about AI adoption in New Zealand.

Highlights

The WhatsApp community is approaching 200 members, and the next in-person meetup is locked in for April 8th at 2 Degrees in Auckland, this time focusing on Cowork — Anthropic’s sandboxed coding environment that bridges the gap for non-developers.

The Power of 10 Minutes

Dave stole the show with an interactive machine guarding tool he built for WorkSafe New Zealand — all from publicly available legislation, created in just 10 minutes using Claude. His technical expert colleagues were stunned: “We’ve needed this for a long time.” Meanwhile, their traditional development team had spent a year building similar internal tools.

Show Don’t Tell

Klaus brought multiple demos: an article-to-video pipeline using ElevenLabs and AI-generated diagrams, an interactive governance decision canvas for NEAR Protocol, and a weekend hack to bulk-manage Twitter follows. His message was clear — “I feel unleashed these days. Whenever I have an idea, I just go and build it myself.”

The Adoption Question

Hamish shared his “bottom-up” strategy for enterprise AI adoption: find the curious 8% within an organisation and empower them, rather than trying to move everyone at once. This sparked a broader discussion about governance, data privacy, and the reality that enterprises need proper contracts — not just credit card subscriptions — to feel safe with AI.

Looking Ahead

The group discussed hackathon opportunities through Adam’s Anthropic ambassador role, the need for more diversity in the community, and plans to share meeting summaries more broadly via WhatsApp and LinkedIn.

Topics Discussed

April 8th Meetup Planning

Next Auckland meetup locked in for April 8th at 5pm, hosted at 2 Degrees again. Focus will be on Cowork — bridging the gap between developers and non-technical users.

AI Governance & Data Privacy

Discussion on enterprise AI adoption challenges — GitHub's default training-on-code toggle, terms of service tracking, shadow IT risks, and the gap between enterprise contracts vs credit card subscriptions.

Bottom-Up AI Adoption

Hamish shared his approach of finding the curious 8% within organisations rather than trying to move everyone. Focus on motivated teams who experiment, rather than top-down rollouts.

Fast Fashion Software

Jason raised the concept of disposable software — individuals creating mini apps for their specific needs. Adam shared a real example of rebuilding his solar inverter dashboard via WhatsApp in minutes.

AI Slop & Technical Debt

Klaus raised concerns about AI-generated code debt accruing faster than human-written code. Story of someone inheriting a 2-month AI-built project that was a mess.

Reinforcement Learning & Self-Improving AI

Jake discussed Karpathy's auto research project, potential for auto reinforcement learning, and the importance of evals/metrics for self-improving systems.

Show & Tell

Dave

WorkSafe Interactive Machine Guarding Tool

Fed NZ machine guarding legislation into Claude and built an interactive decision-support widget in 10 minutes. WorkSafe branded, hosted on GitHub Pages. Subject matter experts said 'we've needed this for a long time.'

Klaus

Article-to-Video Pipeline

Built a pipeline that ingests long-form articles, partitions them into 5-minute presentation scripts, generates ElevenLabs narration, creates explainer diagrams via Nano Banana, and stitches together MP4 videos.

Klaus

Excaliburate — Twitter Unfollow Manager

Weekend project to bulk manage 8,000 Twitter follows — rank by follow-back status, last active date, and bulk unfollow.

Klaus

NEAR Protocol Decision Canvas

Interactive decision support tool for on-chain governance votes — live sliders to model inflation rate impacts on fee burning and treasury over time.