Generative AI: Reshaping Our World (Part 1 of 5)

Welcome to the first installment of our five-part series on Generative AI.

In this series, we’ll explore the transformative power of this technology, its implications for various industries, and how businesses can prepare for the AI-driven future.

Generative AI is a technology redefining our understanding of intelligence, neural networks, and machine learning.

It’s reshaping our perception of possibility and randomness, creating never-ending brand-new stochastic paths for future generations.

To put this technological leap into perspective, let’s consider a historical analogy:

“Fire is one of the oldest technologies of humankind,” writes archaeologist Silje Bentsen in her article “Fire Use,” dating it back to two million years ago.

Fast forward to Gobekli Tepe (Potbelly Hill), an archaeological site in Turkey believed to be among the first human settlements, dating back 11,500 years.

This represents a technological leap of 1,988,500 years, from cooking to settling. How long does such a leap take today?

As we ponder this, let’s dive deeper into generative AI, infusing personal experience with actuarial judgment and business perspective.

What is Generative AI?

At its core, generative AI refers to algorithms capable of creating new content across various mediums – text, audio, code, images, and videos.

Think of it as a digital Renaissance artist proficient in multiple forms of creative expression, much like how Leonardo da Vinci excelled in painting, drawing, and scientific exploration.

I think it’s no wonder OpenAI named one of their models ‘DaVinci’—a nod to the polymath’s unparalleled versatility and innovation.

These AI models, with OpenAI’s ChatGPT and Anthropic’s Claude 3.5 Sonnet being two prominent examples, are built on a foundation called transformer architecture.

They’re trained on massive datasets, essentially digesting oceans of text and information from across the internet, books, articles, and any other form of data that can be digitized.

The Inner Workings

The secret sauce of generative AI lies in its training process.

These models learn patterns and relationships within the data, enabling them to generate new content that follows these learned patterns.

A classic example is akin to reading every cookbook ever written and then creating entirely new recipes.

Here, the text is recipes, and the context is cooking.

Humans tell the machine in machine’s language to find patterns between different cuisines and generate new recipe ideas based on the recipes all human generations have collectively developed.

The scale of this training is staggering!

For instance, OpenAI’s GPT-3 was trained on approximately 45 terabytes of text data – equivalent to about 174 million average-sized books.

Looking Ahead

In the next part of our series, we’ll explore the capabilities and limitations of generative AI, including its potential for factual inaccuracies, biases, and resource intensity.

We’ll also delve into a real-world application in the pharmaceutical industry, inspired by personal medical experiences.

Stay tuned for Part 2, where we’ll continue our journey into the fascinating world of generative AI.

---

This article is part of a 5-part series on Generative AI. For a complete list of references used throughout this series, please visit this post.