Beyond the Sci-Fi Tropes

It is impossible to talk about the future of technology without talking about Artificial Intelligence (AI). But before we dive into the philosophical debates or worry about robots taking over the world, we need to understand what AI actually is right now.

Much of the media coverage around AI makes it sound like magic, or like there is a tiny human brain trapped inside a computer. In reality, modern AI is mostly advanced mathematics and pattern recognition.

Machine Learning and Training Data

Most of the AI tools making headlines today (like ChatGPT, Midjourney, or recommendation algorithms) are based on Machine Learning (ML).

Instead of programming a computer with strict, step-by-step rules (e.g., “If X happens, do Y”), developers feed a massive amount of data into an algorithm. The algorithm analyzes this data, finds patterns, and learns to make predictions based on those patterns.

For example, if you want an AI to recognize pictures of cats, you don’t write code explaining what a pointy ear or a whisker looks like. Instead, you feed the system millions of images labeled “cat” and millions labeled “not cat.” Over time, the math adjusts itself until the system can accurately guess whether a new, unseen picture contains a cat.

Large Language Models (LLMs)

Tools like ChatGPT are powered by Large Language Models (LLMs). These are machine learning systems trained on an unfathomably large amount of text—essentially, the entire readable internet.

When you ask an LLM a question, it isn’t “thinking” in the human sense. It is using complex probabilities to predict what the next most logical word in the sentence should be, based on all the text it has ever “read.” It’s basically an incredibly advanced, highly sophisticated version of the autocomplete feature on your phone keyboard.

The “Black Box” Problem

One of the biggest challenges with modern AI is that the systems are so complex that even the engineers who build them don’t always know exactly why the AI made a specific decision. This is known as the “black box” problem. The math works, the predictions are accurate, but the internal logic is opaque.

Understanding that AI is currently just highly sophisticated pattern recognition helps strip away the hype. It is a powerful tool, yes, but it is a tool built on data—and the quality of that data is everything.

Agentic AI

While traditional AI like ChatGPT waits for your prompts to generate text, Agentic AI represents the next leap forward. Instead of simply answering questions, an AI “agent” is given a goal and allowed to act autonomously to achieve it. It can break down tasks, search the internet, use software tools, and make decisions on its own. While this opens up incredible possibilities for automation, it also introduces significant new risks regarding security, accountability, and the loss of human oversight.