RSSAmplifier

Blog

- Technology Blog Posts

Technology, coding, public cloud, cyber, banking, AI, crypto and culture

andrewbaker.ninjaRSS feed ↗11 posts

Latest posts

AI Context Is Not Memory: Why Agents Need to Forget

1. The follow up question In my previous post I explained why context is not free, and why a smaller AI model carrying a huge context can sometimes cost more to serve than a larger model with a short one. The response I heard most often, in different words each time, was the same question. […]

Context Windows and KV Caching Explained: Why They Drive AI Inference Costs

1. What a token actually is Before any of this makes sense, it helps to be precise about what a token is, since the whole article rests on it. A model does not read text the way a person does, character by character or word by word. Text is first broken up into tokens, which [ ]

Security Vibing: Why Teams Keep Sacrificing to the Risk Gods Instead of Managing Real Risk

Something interesting happens inside large organisations when a new technology appears and nobody is quite sure how dangerous it is: people start sacrificing things. A feature gets disabled, a setting gets locked down, and a capability gets removed. Another approval step gets inserted into the process, or someone quietly updates a policy. Everyone feels slightly [ ]

Open Weights: Why Would Anyone Give Away the Most Expensive Part of an AI Model?

1. A strange thing is happening in AI There is something slightly strange happening in artificial intelligence. Companies are spending enormous amounts of money buying GPUs, building data centres, collecting data, employing researchers and training increasingly capable models. Training a serious frontier model is not like writing an ordinary piece of software. Much of the [ ]

The AI Big Five: How Your Business Survives What Is Coming

I was in the Kruger Park recently, where everyone talks about the Big Five: lion, leopard, elephant, rhino and buffalo. They are the five animals everyone wants to see, and they are big, powerful and dangerous, built by evolution to survive in an environment that punishes weakness immediately and without appeal. It got me thinking [ ]

Leadership Success: Are You Earning It, or Borrowing It?

1. Two million users A product review lands on your desk. It shows 2 million users, and the team presenting is proud. Rightly so. Two million is a large number by any measure. Then you look at the parent organisation. It has 8 million active customers, and the feature has been sitting prominently inside the [ ]

AI and Organisational Design: When Capability Belongs to Opportunity and Purpose, Not the Org Chart

1. What I saw in Kruger I was walking through Kruger National Park recently when I noticed small pieces of metal sticking out of the ground, the remains of old fence posts. I asked the ranger what they were, and he explained that Kruger had not always been the enormous open ecosystem we see today. [ ]

How to run Claude Code on OpenRouter with alternative models like DeepSeek: the ANTHROPIC_BASE_URL guide

1. Why this is worth being precise about Claude Code s model can be changed through an environment variable, and that variable can point at a gateway such as OpenRouter rather than at Anthropic s own API. This is how people run Claude Code against DeepSeek, GLM, or any other backend a gateway serves. The question comes [ ]

Centralization vs Decentralization: Why Recentralizing Teams Fixes the Wrong Problem

1. Two arguments that sound contradictory and are not Most organisations end up holding two opposite opinions about centralisation at the same time, often within a few weeks of each other. 1.1 The case for pulling a function back together Somewhere, a function that used to sit centrally has been pushed out into the business [ ]

Claude Code Without Model Lock In: Ghostty, OpenCode and OpenRouter

CloudScale AI SEO summary What this is: a practical build guide for running an OpenCode agent behind Ghostty, routed through OpenRouter, as an alternative to a single model coding agent such as Claude Code. Why it matters: token pricing across OpenRouter s catalogue varies by a factor of four to over a hundred for the same [ ]

From Zero to Exploit Detection: A Practical Guide to Nuclei Templates for Pen Testing

Vulnerability scanning used to mean choosing between a handful of expensive commercial scanners or writing your own scripts for every check you cared about. Nuclei changed that calculation. It is an open source scanning engine built by ProjectDiscovery that runs YAML defined templates against targets, and the template library has grown into one of the [ ]