Part two: as AI inference margins collapse, who actually captures the value? The hardware supply chain and consumers win, model inference commoditises, and the frontier labs' escape routes are managed agents and staying ahead.

GLM 5.2 is the first open weights model I'd call a genuine competitor to Opus and GPT for agentic work - at ~15-20% of the price. Part one of why AI inference margins are about to collapse.

Profiling a MoE model to find which experts matter for a specific task, then quantising the cold ones hard. The result: near-Q4 quality at near-Q2 size for local models.

How KV cache compression - from MQA and GQA to MLA and linear-attention hybrids - quietly unlocked the long context windows that make modern agentic LLMs possible.

xAI is renting huge amounts of GPU capacity to Anthropic and Google. Financial engineering ahead of the SpaceX IPO, a real compute shortage, or a genuine datacentre advantage? Probably all three.

The UK is obsessed with building AI datacentres at home. But the arguments for sovereignty - latency, tax, control - mostly don't hold up.

I joined Radical's Built for Turbulence podcast to talk about what AI agents are doing to the economics of software, the Figma Trap, and why running human-written code without AI audit is going to start looking reckless.

Google's Gemini 3.5 Flash was the headline model at I/O - fast, but expensive and middling at coding. Why it makes more sense as a model built for Google itself, the TPU advantage, and Google's real weakness in coding agents.

Managed agents (cloud-hosted agent harnesses) are powerful, but locking yourself into a frontier lab's platform now is risky - here's why and what to do instead.

Open weights models keep frontier labs honest on price. If they disappear, we end up with a handful of oligopolists extracting consumer surplus.

A fictional scenario about what AI changes for cloud security, written because the technical version of the argument doesn't land with anyone except engineers.

A back-of-envelope tokens/sec estimator for running LLMs locally. Pick your memory, model size and quantization to see how fast it'll actually run.

Figma's reliance on non-designer seats made it uniquely exposed to AI. Claude Design's launch deepens the problem.

I built a small tool to visualise how Mixture of Experts models route tokens through different experts. It's genuinely fascinating to watch.

What Anthropic's Mythos research preview tells us about the trajectory of frontier models, sandbox escapes, and the cybersecurity risk ahead.

AI compute demand is growing exponentially while supply constraints bite hard. The next 18-24 months are going to be defined by shortages, rationing and price discovery.

A cascading wave of supply chain attacks has hit npm and PyPI in under two weeks. LLMs are making it worse, and current mitigations aren't enough.

How I used Claude Code to fix Linux desktop issues, get 'garbage'-rated Windows apps working in Wine, and what it means for software ecosystems

Anthropic's 1M token context window on Opus 4.6 and Sonnet 4.6 is a genuine breakthrough - and they're not even charging more for it.

Using Qwen 3.5 open weights models to OCR scanned PDFs - locally on consumer hardware or via OpenRouter for pennies

The viral claim that Anthropic loses $5,000 per Claude Code subscriber doesn't survive basic scrutiny. Let's do the actual maths.

Claude Code has 2-3 million users. That's 1% of knowledge workers. The compute math gets scary from here.

On-device AI agents sound great in theory. The maths on KV cache scaling, RAM budgets, and inference speed says otherwise.

Why I replaced SaaS code review tools with OpenCode running in CI/CD pipelines - cheaper, more secure, and works with any Git provider

I benchmarked 19 web frameworks on how efficiently an AI coding agent can build and extend the same app. Minimal frameworks cost up to 2.9x fewer tokens than full-featured ones.

A step-by-step guide to converting complex Excel financial models into Python using Claude Code, enabling faster scenario testing and integration with other tools

Anthropic's red team found 500+ critical vulnerabilities with Claude. But they focused on maintained software. The scarier problem is the long tail that nobody will ever patch.

I cloned Linear's UI and core functionality using Claude Code in about 20 prompts. Here's what that means for SaaS companies.

A step-by-step guide to extracting your brand design system and generating on-brand PDF reports and slide decks using coding agents.

A simple trick to keep your CLAUDE.md and AGENTS.md files updated using the agent's own chat logs - turning a tedious chore into a 30 second job.

Anthropic's 'legal tool' that triggered a $285bn selloff is 156KB of markdown. The panic reveals a hard truth about the future of software.

A bifurcation is happening in AI adoption - power users shipping products in days versus everyone else generating meeting agendas. Enterprise tool choices are accelerating the divide.

I used to be a TDD sceptic - too much time writing tests for features that might get deleted. Then coding agents completely changed the economics of software testing.

Permission systems, Docker sandboxing, and log file secrets - why current approaches to securing coding agents fall short and what we might need instead.

Why DRAM shortages, not capital, will define AI infrastructure growth through 2027

Comparing token efficiency across 19 popular programming languages using RosettaCode data - from Clojure to C, there's a 2.6x difference.

Using Claude Code to port 120k lines of Pascal and 68k assembly to modern C# - and what this means for cross-platform development

MCP tools eat thousands of tokens. A simple CLI with instructions in your CLAUDE.md file uses 71 tokens and works brilliantly.

Travel agents are the classic example of an industry killed by the internet. Software engineering is facing the same disruption, but the timeline is compressed.

Critics are judging models trained on last-gen hardware. There's a 6x wave of compute already allocated - and it's just starting to produce results.

Using ASTs and AI agents to reverse engineer minified JavaScript in minutes instead of weeks

Software ate the world. Agents are going to eat SaaS.

Agentic coding tools are dramatically reducing software development costs. Here's why 2026 is going to catch a lot of people off guard.

Gemini 3 Pro's design capabilities and Opus 4.5's reduced babysitting needs represent a subtle but significant leap that traditional benchmarks completely miss.

Using IPv6 with Cloudflare to run multiple services on a single server without a reverse proxy

A practical approach to managing production infrastructure using git-tracked markdown files and Claude Code for small teams

Software engineers underestimate the scale of Excel usage. With agents now able to work directly in spreadsheets, we're looking at transforming how billions of dollars in business processes are managed.

Looking at actual token demand growth, infrastructure utilization, and capacity constraints - the economics don't match the 2000s playbook like people assume

A non-technical CFO built a production operations dashboard with Claude Code that had failed with low-code tools and agencies. This shift in who can build software is going to change everything.

I built a tracker to monitor the growth of MCP servers in the wild - turns out the ecosystem is growing faster than I expected

Insights from MCP Dev Summit Europe on agentic discovery, client compatibility challenges, and the emerging field of agentic experience design

Why GitHub Actions runners are slow and how bare metal servers can make your CI/CD 2-10x faster while costing 10x less

Google's Gemini AI Studio API has been suffering from severe reliability issues with little transparency about the problems on their status page.

From magical to frustrating in months. Why AI coding agents feel like dial-up internet and what ultra-fast inference could unlock for developer productivity.

How to give AI coding assistants complete visibility into APIs and third-party libraries using static analysis instead of basic text search.

Deconstructing the real costs of running AI inference at scale. My napkin math suggests the economics might be far more profitable than commonly claimed.

Testing Claude Code beyond software engineering - using it as a tax agent to analyze documents and navigate complex tax scenarios in real-time.

Testing Model Context Protocol support across OpenAI, Anthropic, and others. The reality of cross-platform MCP implementation in 2025.

Starting a blog about AI-assisted development, MCP integrations, and building production software with modern tooling like Claude Code and Cursor.