-
Human Slop vs AI Slop
The argument that AI code is slop and human code is clean is backwards. Humans have been shipping bad code for decades. Backdoor textareas, 1500-line methods, zero test coverage. AI just produces the same ratio faster. Your pipeline was already broken. Fix that.
-
Getting OpenClaw up and running
Discover how to get OpenClaw up and running. How easy it is as I walk you through a setup that lets you have your own virtual AI assistant ready, and willing to do your virtual tasks in a secure and controlled environment.
-
We Are a Campfire, Not a Family
The “we are a family” pitch in hiring is a trap. It demands loyalty and sacrifice without asking, and creates silence when things go wrong. The alternative is the Campfire model. People gather, contribute, and leave. The fire endures. Stop pretending otherwise.
-
AI Imposter Syndrome: The Wave You Can’t Skip
Every previous tech wave had an escape hatch you could skip and your career was fine. AI does not. 72% of mid-career engineers report AI fatigue. This piece breaks down the five-step spiral from loss of authorship to burnout and offers interventions that work: retrieval practice, AI free days, and deliberate skill use over passive rest.
-
PE has an inventory problem; AI is not the solution
Private equity faces a massive inventory crisis, with thousands of companies sitting unsold as easy multiples vanish. Instead of solving this pricing problem through true innovation, firms are falling into an “Efficiency Trap” using AI to cut costs while simultaneously destroying the premium value that makes assets attractive to buyers.
-
Beyond the AWS Default: The Cloudflare Alternative
The standard move is to default to AWS. It is the safe choice, but it often comes with an unnecessary tax of complexity. For many projects, a faster path exists. By choosing Cloudflare, you can trade deep infrastructure layers for edge velocity, cutting through the noise to deliver a simpler, more scalable solution.
-
Roadmap to LLM Filenames: From Quantization to Architecture
Navigating the world of local LLMs can feel like decoding a secret language. This guide breaks down common model filenames—explaining terms like Q4/Q6, MoE, and MTP to help you understand how resolution, architecture, and prediction logic impact performance, so you can choose the perfect balance of intelligence and speed for your hardware.
-
The Risk Premium: Why Your Next AI Integration Should Be Local
Stop worrying about API costs and data leaks. Discover why moving AI to the edge, using local, open-source models like Gemma 4, offers the ultimate solution for data sovereignty. Learn how to build secure, scalable products by embedding intelligence directly into your infrastructure, ensuring your proprietary data never leaves your network.
-
How to run AI locally
Let me walk you through how to run AI models locally on your own machine, removing the logistics and figuring out which type of models you can actually run. You will be surprised at how easy and quick it is.
-
Enterprise Stability: Who’s Cutting Your Grass?
Sometimes we forgot who keeps the lights on and the level of effort required. Maintaining enterprise stability is not effortless and we even in this AI age, it is not a release and forget exercise.
-
Jack Clark talks AI sense
Jack Clark was featured on Rest is Politics podcast, which turned out to be an non-hyped sensible conversation with an honesty seldom seem from the AI founders.
-
Podcast Episode: Prompting, AI, And Business Chaos
An experiment of an AI generated podcast from a series of previously published articles. You decide.
the legend that is, Alan Williamson
For over 24 years, I’ve shared my thoughts here on this blog whenever the muse strikes.
It’s unpredictable, honest, unedited, and occasionally witty. I’m living the best years of my life, working harder than ever, learning each day and enjoying every second.
