You may have noticed something remarkable happening in the technology world since last December, something so astonishing it leaves you agog. Well, let me let you in on a little secret. It's not AI, as you might have guessed. No, it's something far more nuanced: Machine Learning, and more specifically, its uncanny ability to predict outcomes.

There's an irresistible allure about prediction, a sense of controllingr the uncontrollable, of peering into the mists of uncertainty to discern a shape, a pattern, an outcome. And that's precisely what Machine Learning gives us. It's the reason why predictive technologies are causing such a seismic shift, toppling the old world order in favor of the new.

We've been living with predictive AI for some time now, perhaps without even realizing it. It's there every time your browser magically completes your search query or when Spotify or Netflix suggest a song or a movie that matches your taste to uncanny precision. Behind the scenes, predictive AI works tirelessly in fraud detection systems, sniffing out anomalies and threats before they materialize. And let's not forget the self-driving cars that deftly navigate our roads, their course plotted out in real-time by predictive algorithms.

One prime example of predictive technology is ChatGPT, developed by OpenAI. It operates on a simple input/output basis. You provide the input—the information you possess—and it fills in the output—the information you don't. It's the same principle at work when your browser autofills the search box, but there's something radical about ChatGPT: it's a Large Language Model.

ChatGPT has been trained on an unimaginably vast corpus of the world's written communication. Imagine a dataset as big as the internet itself, and you'll have a sense of its scope. It's the scale of its training data that gives ChatGPT such power and versatility, taking us to places even its designers never anticipated. It's not far-fetched to imagine a future where language models like ChatGPT will be able to diagnose illnesses more accurately than doctors.

The sudden mainstream popularity of chat-based AI like ChatGPT can be attributed to one key factor: cost. OpenAI has managed to keep the cost of its predictive technology impressively low. In fact, the pricing on their API access dropped a staggering 96% this year alone. This cost reduction opens the door for widespread application of predictive technology, and according to the laws of economics, it's almost inevitable.

Predictions create value. There is massive demand for accurate predictions. Therefore, cheaper predictions will inevitably lead to more predictions. It's as simple as that.

We've already seen this principle at work. We've reframed writing, coding, images, and music as prediction problems. The challenge now is to identify other disciplines and problems that can be similarly reframed. It's a shift in perspective that promises to unlock untold potential, and the biggest opportunities likely haven't even been identified yet.

So, as we stand on the precipice of this new era, it's worth reflecting on the transformative power of prediction. With Machine Learning leading the charge, we're not just predicting the future—we're shaping it.