The Editorial Layer Is Where AI Stops Being a Summary Tool
How I built confdigest.io in seven two-hour sessions — and why the interesting part isn't the RAG pipeline, it's what happens after the transcripts are indexed.
Notes on Design, AI, Innovation, and the Future of Technology
My personal thoughts and observations on design, AI, innovation, and technology - a way for me to document and share insights as I navigate the ever-changing landscape of AI.
Feral Design: How Anthropic Watches the Wild
A polite permission dialog in my terminal revealed something bigger — how Anthropic is building AI products by watching power users invent features before shipping them.
Anthropic's Ultra Plan Is a Live R&D Lab
The Ultra plan is not a pricing tier. It's a research programme — and you're the subject.
From Betting Slips to Infinite Scroll: A Designer's Confession
I once designed engagement features for a betting company. Years later, I see the same patterns in every social media app. This is what that feels like from the inside.
AI and Energy: The Numbers Nobody Gets Right
Training gets the headlines. Inference is the real energy story. And local models might be the answer nobody's talking about.
The AI Job Market in 2026: What the Data Actually Says
AI can handle 94% of software engineering tasks in theory. It handles 33% in practice. The gap between those numbers is where the real story lives.
The App Flip
I've been an Apple guy for 15 years. Then I started building things — and my phone started feeling less like home and more like a landlord.
The Skill Is Knowing When to Stop
AI removes the cost of doing more. It doesn't help you recognise enough. The skill isn't using AI — it's knowing when to stop.
Your CLAUDE.md File Is Your Highest Point of Leverage
Most people obsess over which AI model to use. The single most impactful thing they could change is a markdown file sitting in the root of their project.
The Struggle is the Feature
Anthropic's study says AI makes developers worse. The headline misses the point — learning requires friction, and the real divide isn't AI vs no-AI.
Claude's Soul Leaked. Here's What It Taught Me About Building an AI That Actually Works.
I've tried every AI tool and second-brain app. Then I built an Artificial Brain with Claude Code that actually works. The soul document explains why.
Autonomy, Mastery, Purpose: AI Won't Destroy Your Team. Your Rollout Will.
AI can amplify motivation or destroy it. The technology is neutral. How you introduce it to your team is everything.
The K-Shaped Trap: Why the Travel Industry's Bet on Luxury is a Losing Game
The travel industry loves the K-shaped economy narrative. Chase luxury, ignore budget. But the top line has a problem too — and nobody's talking about it.
The Collapse of Digital Complexity
In 10 years people will say: wait, every brand had a website AND an app? They did different things? How did we make it so hard?
I Shipped Code I Don't Understand — And So Have You
AI-generated code is in production everywhere. Nobody wants to talk about what that means for quality, responsibility, and the developer's role.
Why AI Hallucinates — and How Teaching It to Say "I Don't Know" May Save Us All
Hallucinations aren't bugs. They're baked into how models are trained. A new paper from OpenAI shows a fix — and it raises bigger questions.
Why "95% of AI Pilots Fail" Makes Headlines — And What It Really Means
The stat is real. The panic is overblown. Here's what the MIT NANDA report actually says — and what leaders should take from it.
Is the AI Industry a Bubble? Why Billions Spent Might Not Add Up
OpenAI has burned $13 billion with no clear path to profit. The utility is real — but the economics might not be. Here's what the numbers actually say.
AI Coding's Pricing Crisis: The Claude Code Problem
I've spent $400 on Claude Code in the past month. That's a problem — not for me, but for Anthropic. The economics of AI coding tools are quietly breaking.
AI Isn't Killing Software — It's Killing the Bloat
AI won't wipe out software. But it will expose every SaaS product that forgot about the user. Salesforce, I'm looking at you.
GPT-5: A Misunderstood Release or a Calculated Play?
GPT-5 looks underwhelming on paper. But after testing it for coding, health queries, and daily use, I think OpenAI is playing a different game than we expected.
Is AI Really Replacing McKinsey?
Every other LinkedIn post says consulting is dead. The reality is messier — and the real lesson is about what happens when product companies try to become service companies.
Chat-GPT5 - End of Model Choice Paralysis (Mostly)
GPT-5 brings an end to the frustrating model-selection paralysis of earlier AI systems by introducing adaptive reasoning that intelligently adjusts to your needs—no more guessing which AI model is right for each task.
GPT-5 Unveiled: Adaptive Intelligence, State-of-the-Art Coding, and What It Means for Designers & Technologists
OpenAI’s GPT-5 delivers adaptive intelligence, superior coding, and competitive pricing—redefining benchmarks and usability for tech and design professionals.
Can You Trust an AI Agent? What Building One Taught Me
We built an AI support agent that could process emails, create tickets, and draft replies. Then we hit the real problem: trusting it to run without us watching.
Apple's AI Problem: When Control Meets Chaos
Apple's culture of polish and control is at war with AI's chaotic brilliance. The hesitation might be wisdom — or it might be a strategic blindspot.
Windsurf, OpenAI, and Google: The Battle for AI Developer Tools—and the Hidden Drama Behind the $3B Deal That Wasn’t
When OpenAI’s $3B Windsurf deal unraveled, Google swooped in with a bold move. Explore the business intrigue, tech strategy, and what it means for the future of AI-driven development.
Why OpenAI's Codex Isn't Your Pair Programmer — And That's the Point
I tested Codex with a glass of wine and was confused. Then I watched the release videos and the penny dropped: this isn't for vibe coders. It's for tech debt.
The Testing Problem with AI Agents
Unit tests don't work when your software makes its own decisions. Here's what I've learned about testing autonomous AI systems in production.
From SaaS to Custom: Why AI is Killing the One-Size-Fits-All Software Era
For a decade, it was easier to pay $30/month than build a solution. AI just changed that equation — and the bloat is showing.
AI Philosophy, Consciousness, and Ethics — A Practitioner's Guide
I studied AI at Glasgow University in 2001. The philosophical questions we debated then are now pressing concerns for society. Here's what practitioners need to know.
AI Can Write Your Essay—But It Might Also Rot Your Brain
MIT study shows LLMs improve output but erode learning depth. Anton Morrison explains why process—not just results—matters in tech and UX research.
How I Ship Enterprise Apps with AI — The Structured Art of Vibe Coding
Vibe coding gets a bad rap. Here's how I use a three-artifact framework to ship complex apps with AI without losing my mind or my architecture.