Talking to the World Around Us: My AICON Belfast 2026 Talk
At AICON Belfast 2026 I asked what happens when we can talk to the world around us. Five lessons from running an AI guide live in three national museums.
Artificial Intelligence
Deep-dives on building AI systems that survive contact with production — architecture, evaluation, guardrails, and the engineering that sits between a model and a product.
2026
Moats, Money and a US Presence: What SF Founders Taught Me
What SF Tech Week 2026 taught me about building a moat when frontier labs can copy you, raising pre-seed money, and why every founder said to go to the US.
Building a Production-Ready MCP Server: How I Gave an AI Agent the Run of Our Platform, and Why I Removed Every Delete Button First
A Model Context Protocol server that exposes an entire production SaaS platform to an AI agent — content, sessions, devices, analytics, billing, jobs. Seven layers, 289 tools, and no delete button anywhere on the surface.
Software Engineering's Great Re-bundling
AI made producing software artefacts cheap and verifying them expensive. That single split is why engineering roles are collapsing into fewer, wider ones.
From Model to Product: Foundations for Practical AI Engineering
Everyone is building with AI today. Very few are building systems that actually work in production. This article explores the real gap between models and products, and why thinking like an AI engineer matters more than ever.
2025
Building a Production-Ready Prompt Engine: How I Designed a System That Handles Everything from Dynamic Templates to Multi-Layer Guardrails
A behind-the-scenes look at how I designed and built a production-ready prompt engine - from dynamic templates and reusable components to multi-layer guardrails, structured output, and parallel execution.
Bringing Computer Vision to the Edge: How I Cut Real-Time Similarity Search Battery Usage by 50%
Building real-time computer vision on low-end mobile devices comes with unique challenges - especially when every second of inference drains battery. Here's how I optimised similarity search using HNSW and hybrid re-ranking to cut battery usage by over 50%.
Software is Changing (Again): Key Takeaways from Andrej Karpathy’s Talk at YC AI Startup School
Andrej Karpathy’s talk at Y Combinator’s AI Startup School explores Software 3.0, the rise of LLMs as programmable computers, vibe coding, and building partial autonomy apps. Learn how AI is transforming how we build - and who gets to build - the future of software.
Why SEO Matters More Than Ever in the Age of AI (And What You’re Missing Out On)
In a world where anyone can build a product fast, visibility is everything. Discover why SEO is more crucial than ever in the AI era - and how it can make your startup stand out.
OpenAI’s New AI Engineer Can Code and QA Without You - Should You Be Worried?
OpenAI’s CFO has revealed plans for an AI agent capable of building, testing, and documenting software all on its own. What does this mean for developers - especially students and juniors just starting out? Here’s what you need to know.
OpenAI Unveils GPT-4.1: A Quantum Leap in AI Performance and Efficiency
OpenAI’s latest release, GPT-4.1, redefines AI capabilities with a massive 1 million-token context window, enhanced coding and instruction-following skills, and cost-effective performance. Discover how GPT-4.1 and its Mini and Nano variants are set to transform AI applications across industries.
This Shopify Memo is Going Viral - 8 AI Truths That Will Reshape Every Company
Shopify’s leaked AI memo isn’t just a company directive - it’s a call to arms for every business leader. If you’re not thinking AI-first in how you hire, build, and scale, you’re already falling behind. Here’s a breakdown of the memo’s most urgent insights and how to act on them today.
Ghibli-Style: How One Magical Look Captured the Internet’s Imagination
With soft brushstrokes, dreamy colors, and heart-tugging nostalgia, Ghibli-style art has taken the internet by storm. But what exactly is “Ghibli-style,” and why did it melt OpenAI’s servers? Here’s a look into the origins, the viral trend, and the whimsical magic behind it all.
DeepSeek: The Quiet Math Tweak That Might Redefine the AI Future
It didn’t arrive with fireworks or fanfare. DeepSeek slipped into the scene with a simple mathematical change - and suddenly, it was the name on everyone’s lips. But what’s really behind the hype? Let’s unpack how a humble paper sparked global buzz, and what it means for the future of AI, geopolitics, and the race for smarter, smaller, and more efficient intelligence.
Attention is All You Need - The Paper That Changed AI Forever
Before 2017, AI models were powerful but painfully slow and complicated. Then came a research paper titled “Attention is All You Need” - a game-changer that flipped everything we knew about AI on its head. Here’s how it reshaped the future in the simplest way possible.
UI-R1: Teaching AI to Navigate Your Screen Like a Pro
What if your phone could predict your next tap or swipe? UI-R1, a new AI model trained with reinforcement learning, is bringing us closer to smarter, more intuitive virtual assistants. By learning from visuals and language - not just massive datasets - UI-R1 shows how AI can interact with real-world mobile apps more accurately and efficiently.
How Do Language Models Learn Facts? Inside the Mysterious Memory of AI
How do large language models actually learn facts? A new study by Google DeepMind and ETH Zürich uncovers a surprising three-phase process - from slow starts to sudden insights and unexpected memory loss. These findings reveal why your AI assistant sometimes nails the answer - and sometimes confidently makes things up.
Why Do Multi-Agent AI Systems Keep Failing? A Look Into the Latest Research
Multi-Agent AI Systems promise intelligent collaboration between agents, but why do they so often underperform? A new UC Berkeley study digs deep into the causes of these failures - highlighting critical design flaws, coordination issues, and verification challenges - and offers a roadmap for building better, smarter systems.
Generative AI: A New Revolution Reshaping the World
Generative AI is reshaping our world in ways reminiscent of the Industrial Revolution. From automating routine tasks and transforming job roles to revolutionizing the IT landscape and empowering solo entrepreneurs to launch unicorn startups, this new era is redefining creativity and innovation. Dive in to discover how embracing AI could be your ticket to a future full of endless possibilities.
Understanding Large Language Models (LLMs): The Basics of Their Math, Training, and Inference
Large Language Models (LLMs) have transformed the world of artificial intelligence, enabling machines to generate human-like text, answer questions, and even write code. But how do they actually work? This article breaks down the key concepts behind LLMs in a way that is easy to understand, with just enough math to show how things come together.
The Role of Generative AI in LegalTech
Generative AI is revolutionizing the legal industry by automating tasks such as legal research, contract drafting, and litigation support. Its applications lead to increased efficiency, accuracy, and access to legal services, while also presenting challenges like ethical concerns and regulatory issues. Despite these hurdles, AI’s integration is set to reshape legal practices significantly.
The Quest for AGI: Towards Artificial General Intelligence
Artificial General Intelligence (AGI) denotes a hypothetical AI with human-like cognitive abilities, capable of solving diverse problems autonomously. Its attainment poses significant technical and ethical challenges, reflecting debates on consciousness and alignment with human values. While progress in AI continues, true AGI remains uncertain, potentially decades away or evolving gradually.
The Future of Artificial Intelligence
Artificial Intelligence is rapidly evolving, particularly through advancements in generative AI, large language models (LLMs), and autonomous AI agents. These innovations are reshaping industries by automating tasks, enhancing productivity, and augmenting human capabilities. However, limitations exist, such as biases and inaccuracies, necessitating careful integration and oversight. A balanced approach can maximize benefits while mitigating risks.