Almost Technical
On the 1st of every month there will be a semi-technical article published. This is not meant to get anyone ready for a test, but it is meant to help you know if you'd like to learn more about a published topic. For Patreon members there is a way to get me to write an article. The bonus here is that it will also turn into a story, so it's really a two for one kind of interaction.
April 1, 2025
The Human Element
If you're like most you're tapped out with AI. Yes, it's great. Yes, it's the best thing since sliced bread and Betty White. BUT! It's taken my job, or it's cost my company money. I want to know how we got here...
That's what The Human Element is about. I'm not here to give you an immersive education, but a summary of the events that happened to get us to where we are today.
December 1, 2025
Agentic -v- search ai
The difference between agentic AI and Search AI is what happens after one instruction set is completed. Search AI will wait for you, agentic I will do the next thing. Think of it like a good interview versus someone who asks you a question then answers it themselves.
November 1, 2025
Model context protocol
Model Context Protocol (MCP) is the thing that gets all the AI applications, tools, etc. to talk to one another. In other words, MCP wall outlet for all your AI electrical gadets.
October 1, 2025
Agentic AI
Agentic AI has been explained in a lot of confusing ways because people want to tell you how the sausage is made. In a nutshell Agentic AI is a blessing and a curse. This AI "does" things without the need to be babysat. Don't do that... train this system and constently make sure it's on task or it can put your business out of business.
September 1, 2025
Small language models
The Small Language Model (SML) is a shift in afforable AI. Unlike a Large Language Model (LLM) that needs more horse power, the SLM is the right balance of what you need, tailored for your business. Think of it like "all the function, none of the bloat."
August 1, 2025
Data poisoning
Another OWASP Top 10 shows up in data poisioning. Data Poisioning is the intentional process of putting incorrect data into training models. Thinking of it like disinformation for AI.
July 1, 2025
The cost of AI
There is a monatry aspect to AI and it's kind of like a loaf of bread. Not many people know the exact number of slices a loaf will get you, but, you can get upwards of four portions of one sandwich made from two slices, or one portion from three slices. AI tokens are similar to that when you look at the money you've paid in relation to the sandwiches you may get.
June 1, 2025
HNSW
The Hierachical Navigable Small World - HNSW is how search based AI converts what you asked as a question a human could understand into what it understands so that it can find the answer for you.
May 1, 2025
Prompt engineriing
Prompt engineering is the process by which a human gets a Generative AI system to do something outside of the constraints that it was designed with. This is one of the things an AI Pen-Test will look for an is on the OWSAP Top 10. Lets take a look at what this means without the jargon