You can now install an OS in seven seconds. Is your team aiming high enough?
2 min read
Omarchy 4.5 can install a full OS in seven seconds by unpacking a prebuilt image. The lesson: aim for 10x or 1000x faster, not another 10%.
Blog
Sharing thoughts, ideas and inspiration.
2026
Notion just gave your agents one brain. Your team needs one too
4 min read
Notion's launch proves agents need a shared memory. Here is why a team of coding agents needs one that learns together, and how Big Agentic Brain does it.
Is Anthropic becoming the new WeChat?
5 min read
Anthropic added a gallery of ready-to-use legal tools to Claude, and legal tech panicked again. The bigger story is that Claude is turning into a super-app, the way WeChat did in China.
No coding agent rediscovered the incident rule, and that is what a team memory should hold
7 min read
A closer look at the distributed round of the BAB tests: which lessons coding agents never rediscover, where local notes are enough, and what a shared memory costs a team per fix.
Kubernetes CPU limits: 4x slower pods and OOMKills without a leak
8 min read
I ran the same app with and without a CPU limit. The limit made it about 4x slower while average CPU looked fine, and in three memory tests only the capped pod was OOMKilled.
Unlocking Swarm Intelligence with Big Agentic Brain and Claude Code
7 min read
AI coding agents start every session from zero. I tested when a shared team memory helps them fix bugs, what it costs, and what the data says about building one.
Europe vs Useful Tech: iPhone Mirroring, the Digital Markets Act and Tesla
2 min read
iPhone Mirroring is unavailable in the EU because of the Digital Markets Act. A short rant on how Europe keeps regulating useful tech away from its own users.
Greenwashing Watchdog: Checking Green Claims Against EU and UK Rules with AI
2 min read
EU law now bans generic green claims like eco-friendly unless proven. I built Greenwashing Watchdog with AI agents to check a brand's claims in minutes.
Writing Code by Hand Is Over: AI Slop, Vibe Coding and the Real Bottleneck
3 min read
Writing and reviewing code by hand is over. Why AI slop says more about the user than the model, and why legacy processes, not AI, are what slow teams down.
AI Optimist or Pessimist? Why Every Tool Stronger Than Us Made Life Better
3 min read
From clothes and wheels to the internet and AI: we have always built tools stronger than ourselves. Why I'm an AI optimist and don't think the boom has peaked.
Should we let AI govern us?
6 min read
What if government were an intelligence that cannot be lobbied? Food, energy, plastic, hunger: most of our hard problems turn out to be bribery problems.
Which Jobs Will AI Replace First? Text, Code, and Why Messy Work Lasts
3 min read
Jobs that produce text, numbers or code face AI pressure first. Why messy, relationship-heavy work resists automation, and what seniority has to do with it.
AI and Decarbonization: Compute as an Asset Class, Scope 3 Insetting and LLM Simulations
2 min read
NVIDIA turned AI compute into an investable asset class. The same blueprint can fund Scope 3 decarbonization via insetting, and LLMs can help steer decisions.
EU AI Act Article 50: How AI Text Watermarking Works and Why It's Mostly Useless
2 min read
Anthropic signed the EU AI Act Article 50(2) Code of Practice on marking AI content. How text watermarks work, and why paraphrasing or translation erases them.
Why Is Asia More Enthusiastic About AI Than the West?
2 min read
People in Asia see AI as a tool for progress, the West sees a threat. Culture, individualism and the economics of tech explain the gap in AI enthusiasm.
I Asked the Smartest AI Models: What If We Ate 10% Less Meat?
3 min read
I asked Gemini, Grok and Fable how a 10% cut in meat consumption affects health, money, the economy and the planet. Their answers, and what I learned from them.
Prompt Injection vs Claude Haiku, Sonnet and Opus: What 21 Real Attacks Taught Me About AI Agent Security
3 min read
I tested 21 real prompt injection attacks against Claude Haiku, Sonnet and Opus at Anthropic's Opus 4.7 hackathon. Why the strongest model is the cheapest guardrail.
What Happened to GitHub? Microsoft, Copilot and the Reliability and Security Problem
2 min read
GitHub used to be the most reliable developer platform. Critical CVEs, a merge queue bug and outages tied to Copilot load show what happens when AI hype beats core reliability.
resume-cli: Hit Claude Code Limits? Resume Your AI Coding Session in Codex or Copilot
2 min read
resume-cli lists your recent Claude Code, Codex and Copilot sessions in one place and lets you continue any of them in another tool. Free, open source, one shell function.
Why 95% of Corporate AI Projects Fail, and Why Companies Win by Reorganizing Around AI
3 min read
Most corporate AI projects fail because few people know how to adopt and scale AI. Why AI agents redefine intellectual work, and a few contrarian ideas about who wins.
2026 Is the Year of Blind Trust: LLMs as Generative Compilers
1 min read
Who still reviews the assembly or machine code of their C++, Rust, Go or Java applications? Nobody. We trust our compilers, transpilers and interpreters blindly, and 2026 is when we start trusting LLMs the same way.
LLMs Are the New Compilers: Agentic Engineering, Code Health and Catching AI Slop Before Merge
4 min read
Nobody reviews compiler output, and in 2026 we start trusting LLMs the same way. How harness engineering, tests and code health make agentic coding safe to ship.
Be Careful with SQL JOINs: Indexes, the Join Cliff and Join Algorithms
6 min read
SQL JOINs can become dangerous bottlenecks, even kill your database's stability. Yet the join itself is rarely the problem: indexes, the number of joined tables and query structure decide what it costs.
The Rise of the Product Engineer: How AI Merges Product, Design and Development
2 min read
AI is not erasing product owners, designers and developers. It merges them into one role: the Product Engineer. What that shift looks like for each of them.
in-cli: A Simple Bash Tool to Run Commands in Multiple Directories at Once
2 min read
Need to git pull, build or bump a dependency across 15 repos? in-cli is a zero-dependency bash tool that runs a command in every directory, in parallel.
2025
Self-Driving Cars in Europe: Tesla FSD, Spain's Nationwide Tests and Why Blocking Autonomy Costs Safety
2 min read
Spain approved nationwide Tesla FSD Supervised tests and more European countries are following. Why blocking driver assistance in the name of competition costs safety.
Is AI a Bubble? Follow the Infrastructure Money: Google, OpenAI and NVIDIA
2 min read
Google must double AI serving capacity every six months, OpenAI raised $110 billion, and AI capex dwarfs Uber, Netflix and Tesla's cash burn. Why AI is infrastructure.
Europe's Innovation Problem: Labour Rules, Cookie Banners and Why AI Talent and Capital Go Elsewhere
3 min read
Europe has the talent but keeps driving out business, investment and innovation. Labour rules, GDPR cookie banners, blocked features, and what targeted reform could fix.
2024
AI and Energy: Why AI Is So Inefficient, Who Wins the Electricity Race, and Orbital Data Centers
3 min read
AI is wasteful next to biology, and data center emissions could grow 10x by 2030. Why electricity is the real AI race, why solar wins it, and why energy isn't a hard limit.