Deep dives into AI engineering, LLMs, agentic systems, and building production AI applications
An analysis of how AI induces cognitive offloading in education, conflating 'doing' with 'understanding', and creating a new class divide between those who wield AI and those replaced by it.
Explore why the AI industry's pivot to 'fundamental research' is a strategic retreat from the hard realities of productization. This post argues that the future of AI depends on disciplined engineering, not just theoretical breakthroughs.
Current AI is a sophisticated parrot, not a true learner. Discover why neuro-symbiotic frameworks and growth loops are the real path to adaptive cognition, moving beyond the illusion of deep learning.
A deep-dive into why the AI industry is shifting from a focus on scaling foundational models to building value at the application layer, driven by the new realities of cost, speed, and model commoditization.
Discover Chrome's Built-in AI. This deep dive explores speed, cost, and usability advantages, testing the limits of embedded AI in browsers. Leverage in-browser LLMs for interesting AI solutions.
Next generation AI powered privacy automation tool. Analyse any web surface for privacy leaks. Why we are building it.
Human nature, large corporations, politics all mixed into a juicy cocktail about the dynamics of power in big tech. With swords.
Most MCP servers are prototypes. Here's the engineering checklist for building production-grade infrastructure.
I built a simple version of OpenAIs voice functionality, using Groq, Llama3, NextJS, DeepGram, SpeechRecognitionAPI. Talk, listen and converse with LLMs.
A critical analysis of Agentic AI frameworks.
Generative AI is misunderstood as a sentient tool that is capable of expert, highly qualified output. The truth is, Generative AI is capable of performing enormous amounts of work, to an intermediate (at best) level.
Claude 3.5 Sonnet artifacts are to structured output such as code generation, what vector retrieval is to rag. It is the search and retrieval system for structured output.
Edward is a full-stack developer and technology leader with 15+ years of experience
Anthropic's code execution with MCP isn't innovation. It's an admission the original paradigm was broken. Here's what comes next.
Discover the journey of building a cutting-edge AI data privacy tool with five integrated AI APIs. Explore the challenges, costs, and lessons learned in using AI for advancing data privacy.
We now have a cheap, fast and highly capable alternative from Google to build agentic flows.
Chapter one. How did we get here. The beginning of a journey through tech.