Contact Information
Australia
- October 27, 2025
The AI in Automation Puzzle: Why “Learning” is a Myth and How to Get Real Results
AI and automation are two distinct concepts that can work powerfully together, but they are not interchangeable. The excitement around AI has led many to believe it can replace the entire automation process, but this often leads to complex, expensive, and unreliable solutions.
Real LLM Streaming with n8n – Here’s How (with a Little Help from Supabase)
n8n, for all its power in workflow automation, is NOT natively built for streaming HTTP responses. Its sequential, node-by-node execution model is fantastic for many tasks, but it’s a fundamental blocker for true LLM streaming.
The good news? I’ve been wrestling with this and have landed on a robust architectural pattern that brings that smooth streaming experience to n8n-powered UIs, primarily by leveraging the power of Supabase Edge Functions and Realtime subscriptions.
The Shifting Sands of the AI Landscape
In navigating the current AI landscape, it’s become clear that relying on off-the-shelf platforms, even those offering some level of customisation like OpenWebUI, presents significant limitations. These platforms, while useful starting points, often bake in assumptions and rigidities, particularly around critical features like RAG, that don’t cater to the diverse and dynamic needs of real-world business applications. Similarly, the promise of quick-fix AI solutions often seen in online tutorials, while great for proof-of-concept demos, glosses over the true complexity and customisation required for robust, production-ready systems.
Understanding the Brains (or lack thereof) Behind Your Chat App: Why LLMs Aren’t What You Might Think
Large Language Models (LLMs) are incredible pieces of technology, capable of generating remarkably human-like text, answering complex questions, and even assisting with creative tasks. It’s easy to interact with a chat application powered by an LLM and feel like you’re talking to a truly intelligent, aware entity with memory and understanding. However, this is where a common misconception arises, and understanding the reality is key to having a good experience.
Case Study – provider agnostic AI chat interface
A client wanted a “chatbot” to interface with all of the providers (Google, OpenAI, Perplexity, Anthropic, and image generation models) so they could easily compare
Basic comparison between LLM providers
I’ve created this table of information to assist me in remembering the features and restrictions of each provider.
My issue with all AI providers / models
After testing multiple AI platforms, I realized they all had deal-breaking limitations—short conversation memory, prompt manipulation, censorship, rate limits, and unreliable file processing. Many chunk documents, making full-document interactions impossible, and frequently “hallucinate” answers instead of admitting uncertainty. Worse, AI providers prioritize storage efficiency over retaining context, making deep discussions frustrating. I needed an AI that could analyze entire documents, retain long chat histories, and offer real control—so I built my own system. Now, I have an AI that actually works the way I need it to.
From “AI in 7 Minutes” to Reality: The Hidden Costs and Complexities of Production-Ready AI
The marketing hype around AI often oversells simplicity, leaving many to grapple with the hidden complexities and costs of effective implementation. As you dive into the world of AI, you may find yourself entangled in a web of interconnected services, each with its own price tag. From API costs to infrastructure overhead and the significant human time investment, the reality can be sobering. Discover strategies to minimize expenses while maintaining functionality, and learn the harsh truths behind the enticing promises of “AI in 26 minutes.” Prepare for a journey that requires solid planning and patience.
The Illusion of AI Knowledge: Making LLMs Work for You Without Overreliance
Unlock the potential of AI without starting from scratch! Discover how large language models (LLMs) like GPT-4 operate on statistical patterns rather than true knowledge. Learn why simply uploading documents isn’t enough and how Retrieval-Augmented Generation (RAG) can enhance your brainstorming sessions. Explore practical strategies like fine-tuning, embedding systems, and specialized prompt engineering to tailor AI responses to your unique business needs. Whether you’re looking to streamline workflows or enhance relevance, our insights will guide you in leveraging existing models effectively. Dive in to transform your approach to AI and maximize its utility for your organization!