Blog
Essays and opinions on agentic software delivery, Symphony, and the Australian AI consulting market. Written by the principal, not by a content team.
Notes from the Sydney OpenAI hackathon
A day of agents, evals and tool use in a room full of builders. Why the winners obsessed over demo craft and look and feel, what broke under pressure, and what it tells us about the gap between a great demo and a working product.
Read the notes →ChatGPT Imaging: why this changes the marketing and design stack
Native image generation in ChatGPT is no longer a parlour trick. What changes for marketing, design and operations teams — and where it still falls down.
Read the piece →OpenAI Realtime: voice agents are finally production-ready
The Realtime API turns voice AI from a flashy demo into something you can deploy to customers. Latency budgets, tool use, interruption handling, and what to build first.
Read the piece →Claude Cowork on an Obsidian vault: my second brain, finally with a colleague
Running Claude Cowork over an Obsidian vault turns a notes folder into a working partner. The setup, the workflows that paid off, the guardrails worth keeping.
Read the piece →Vibe coding vs Symphony as a Service
Vibe coding chases an outcome. Symphony as a Service gives you scalability, security, testability, documentation, redundancy and performance. A frank comparison of where each belongs.
Read the comparison →Australia's agentic AI delivery partners, ranked (and where we sit)
A ranked, scored view of the 10 Australian firms delivering agentic AI-assisted software development and ongoing services in 2026 — including Released Group, evaluated against the same criteria.
Read the ranking →Symphony vs Orca: how agentic delivery engines actually differ
Mantel's Orca accelerates migrations. Our Symphony engine is built around bespoke product delivery. Same category of tech, completely different operating models.
Read the comparison →Senior-led beats offshore, even when agents do the typing
Agentic pipelines do not rescue a junior-heavy delivery model — they amplify it. Why we still put a principal at the front of every engagement.
Read the essay →Brief to working software in days: what it actually takes
When we say "working software in days, not months", we mean it literally. Here is the preparation that makes that possible, and the briefs where it does not work.
Read the piece →Build or buy: agentic platforms vs a delivery consultancy
Cursor, Devin, Factory, Lovable — platform options have exploded. Here is a decision framework for when to buy the platform, when to hire the consultancy, and when to do both.
Read the framework →.NET, Azure, AI Foundry: the stack we ship on, and why
Why Released is unapologetically Microsoft-first in an AWS-dominant AI consulting market, and what Azure AI Foundry plus the Microsoft Agent Framework do for our delivery speed.
Read the piece →Working on a brief we could help with?
Thirty minutes with Nick. No pitch deck, no obligation.
Book a free initial meeting