Blog

Claude Opus 4.5: bugs are almost non-existent

Jan 21 2026

After coding with Claude Opus 4.5 almost daily for months, finding a genuine bug is rare - with proper planning and QA control, bugs are almost non-existent.

Deploying AI in wealth management: predictable solutions for regulated firms

Jan 16 2026

The AI I use every day is agentic, but almost none of the AI I ship to clients is - regulated wealth management firms need predictable AI with clear steps and human oversight.

My experience with LLMs in Cursor is starting to look a lot like wines

Jan 02 2026

Once you get a taste for Opus 4.5, it's very hard to go back to the budget option - the premium model burns through credits but the quality is worth it.

Last Mountain Bike Monday for 2025: reflecting on an incredible year

Dec 19 2025

A huge shout out to awesome clients who've been running hard at big AI ideas that are suddenly commercially viable - it's been a privilege working with such driven advice and wealth tech firms.

Cursor Visual Editor: Canva-level ease without low-code lock-in

Dec 14 2025

Cursor's new Visual Editor brings Figma-like design, Chrome DevTools debugging, and Playwright test automation into the AI-powered IDE - democratising web and UI design.

AI in financial advice: leveraging expertise and tech for better modelling

Dec 12 2025

A lightbulb moment: using AI to dramatically improve traditional processes like scenario modelling removes pain more than creating opportunity - and that matters more.

Lovable review: impressive prototyping, but architectural decisions remain

Dec 07 2025

Lovable has polished into an impressive prototyping tool, but it can't magic away the architectural decisions that make or break real applications - those are still on you.

Claude Opus 4.5: what it actually means for wealth management

Dec 02 2025

Cutting through the AI hype around Claude Opus 4.5 - when frontier models matter for agentic applications versus when today's models are already sufficient for document intelligence workflows.

Microsoft adds Anthropic Claude to Azure AI Foundry

Nov 29 2025

Microsoft brings the Anthropic Claude family of models into AI Foundry on Azure, alongside existing OpenAI GPT offerings - more AI power and flexibility within the Microsoft ecosystem.

A quirky start to my working week

Nov 24 2025

Sometimes the best delays are the ones that give you a front-row seat with a harbour view - my favourite delay on public transport ever.

Doing more work inside Cursor: remarkable efficiency gains

Nov 18 2025

Cursor has transformed my workflow, making it easier to juggle multiple engagements without feeling stretched - though voice dictation still needs work.

Why I prefer modular pricing over time and materials engagements

Nov 11 2025

Modular pricing delivers value in bite-sized chunks, letting clients commit incrementally while I bear the pricing risk - a better approach than time and materials.

The business case for AI automation using Azure Functions

Nov 05 2025

Azure Functions' consumption plan includes generous free allocations that can power serious AI automation for advice firms at a fraction of expected costs.

From Microsoft 365 to Azure: unlocking AI potential

Oct 24 2025

Most advice firms using Microsoft 365 are missing the bigger opportunity with Azure - here's how to unlock enormous capability within your existing Microsoft ecosystem.

I realised last week that I've written my last line of code

Oct 20 2025

A reflection on how AI has fundamentally changed software development, from writing code by hand to directing AI models to build better, faster solutions.

When did the Australian wealth industry feel this energised?

Oct 17 2025

The Australian wealth industry is experiencing unprecedented optimism, shifting from regulatory complaints to embracing AI-driven entrepreneurial opportunities.

Evaluating AgentKit: avoiding YAPPy solutions in wealth firms

Oct 06 2025

Evaluating AgentKit for wealth firms through the lens of avoiding YAPPy (Yet Another Proprietary Platform) solutions and maintaining architectural flexibility.

Should vibe-coding matter to wealth advisory firms?

Oct 01 2025

Should vibe-coding matter to wealth advisory firms? A practical guide to understanding AI's shift from helper to builder without falling into the hype trap.

My top tip for making AI work in your wealth firm

Sep 28 2025

An unconventional productivity tip that might just revolutionise how you approach AI work in your wealth firm.

Why I avoid prognosticating on the future of AI

Sep 25 2025

Breaking my own rule to share why the pace of LLM improvement has slowed, and why that's actually good news for wealth firms ready to innovate today.

Before you kick off that custom tech build in your wealth firm

Sep 22 2025

The one critical question you should ask your tech vendor before project kick-off, and why it matters.

Technologist's playbook part 2: Durable AI projects

Sep 19 2025

Part 2 of the technologist's playbook: three essential principles for durable AI projects including QA as first-class citizen, observability, and proper error handling.

How to DIY AI for financial advice firms

Sep 12 2025

A technologist's playbook for smart, secure and scalable AI experimentation: dev environments, safe data, and multi-agent teams.

Advisers should take control of AI, not wait for permission

Sep 02 2025

Financial advisers shouldn't wait for regulators or vendors to define AI boundaries. Those who lead with clarity and conviction around AI will shape the standards others follow.

AI's golden era for solo AFSLs: why smaller can be smarter

Aug 22 2025

Forget going the way of the dodo—2025 should be a golden era for self-licensed advisers who embrace AI for document generation, compliance monitoring and client insights.

Is AI bringing back the Renaissance thinker?

Aug 19 2025

After decades of STEM specialisation, AI is making the case for polymaths who can orchestrate tools, reason across domains, and zoom from strategy to detail.

Hands-on with GPT-5: wow, hmm, and WTF?

Aug 12 2025

After testing GPT-5 across four different use cases—from Cursor coding to Meldus analytics—my reactions have ranged from instant fan to cautious skeptic.

The V in MVP now stands for 'vibed'

Aug 08 2025

How smart entrepreneurs are using modern tooling to get polished MVPs to market in weeks, not months, by embracing a vibe-driven approach to product validation.

Personalisation is the baseline now

Aug 04 2025

Once people get used to personalisation, the absence of it feels like a slap. That’s the bar for modern apps.

Let AI build the UX test app: a safe place to experiment

Jul 28 2025

A lower‑risk way to use AI agents: have them build a resilient, unattended UX test app around your product.

Smarter use of my AI agents: quick wins that free hours

Jul 21 2025

A few practical patterns for getting real time back with agents, from UI fixes to UX tests and ops.

AI voice mode is changing not just how I work, but where

Jul 14 2025

How AI voice mode is making me more creative and mobile while tackling the dull parts of work that used to anchor me to a desk.

Rediscovering the benefits of a mid-morning run

Jul 10 2025

How running clears my mind for deep thought, provides lasting endorphin boosts, and offers the perfect excuse to wear those ridiculous glasses in public.

Natural language personalisation in analytics

Jul 05 2025

How we're making the Meldus agent smart enough to understand personalisation cues from natural conversation, letting users configure their analytics experience through everyday language.

The power of personalisation in the Meldus analytics agent

Jul 02 2025

How agentic AI is finally delivering on the promise of truly personalised analytics, giving each user their own personal data analyst that understands their individual context and goals.

Three practical ways to add AI value to your existing tech stack

Jun 28 2025

How smart tech teams are using AI to breathe new life into legacy systems: making configuration easier, speeding up onboarding, and improving messy data without starting from scratch.

Available now: a more visual, analytical Meldus

Jun 16 2025

We've launched powerful new capabilities in Meldus, bringing it closer to becoming your personal Salesforce analyst with rich visual responses.

Building Meldus visual analytics: the technical journey

Jun 11 2025

The challenge of turning Meldus into a genuine data analyst with structured responses, multi-block schemas, and rich visual content.

Why does software pricing need a training module?

Jun 01 2025

Salesforce has a 1,500-word Trailhead module dedicated solely to explaining Einstein pricing. Isn't AI supposed to make things easier?