GPT and AI for “Dummies”

Do you know what the acronym GPT stands for? Doesn’t it have something to do with artificial intelligence and whatever chatbots are supposed to do? And, for that matter, what is artificial intelligence (AI) — and what isn’t it? Also, while we’re asking all these questions, if we’re so afraid AI and GPT chatbots are going to take over humanity and make us extinct, why do we keep building so many hyperscale data centers all over the place? This makes it sound like we’re building our own digital gallows.

Depending on the exact definition you might use for the term Artificial Intelligence, we very possibly have been living with AI for a long, long time. So, what’s the big deal all of a sudden? A liberal definition of AI might include many of mankind’s earliest inventions like mechanical clocks, the abacus and the light bulb. But a more operative AI definition for today’s world would only include computers, networks and software that AI developers say have the capacity to “think.”

If you don’t like any of these definitions, here’s one from Google’s AI agent, “Gemini”: “Artificial intelligence refers to computer systems that perform complex tasks typically requiring human-like reasoning, decision making, creating, or perception, often without significant human oversight and with the ability to learn from experience and improve performance.” (That’s a machine talking, just so you know.)

I don’t think any of these definitions are very helpful when we’re trying to figure out what kind of a future we will be living in, and for how long. How does it help to know GPT stands for Generative Pretrained Transformer? Or that GPT is the latest advancement in Large Language Models (LLM) that feed trillions of bits of data to networked neural processing computer chips and logic boards in all those giant data centers that keep getting built.

Even after “inventing” all that, aren’t we still just a bunch of apes that taught ourselves how to walk upright and do embarrassing things?

Share Moral Dilemmas: First Drafts & Commentaries

The Dawn of AI

The first labs dedicated to studying and advancing Artificial Intelligence were opened in 1956 including at Dartmouth College, IBM and a few other private corporations. This followed the challenge we now call the “Turing Test,” published in 1950 by Alan Turing, the mathematician, inventor of the Enigma Machine and a pioneering computer scientist. He challenged researchers to determine if a machine could “think” or be “intelligent,” and to write down the steps on how to prove it. A version of the Turing Test is broadly used today on the Internet as the CAPTCHA test to determine whether a user is a human or a bot. (CAPTCHA stands for Completely Automated Public Turing test to tell Computers and Humans Apart. It rhymes with “gotcha!”)

In just the last few years, GPT, Google’s Gemini, Meta’s Galactica, OpenAI’s ChatGPT4 and Anthropic’s Claude chatbot, among many other high-profile versions of AI assistants, have brought us into a new age of computer convergence. The inventors of these thinking machines are now confessing they can’t keep up with all the rapid advances in AI and artificial reality. So, where does that leave the rest of us?

The advent of generative pretrained transformers (GPTs) is reinventing all the former architecture of the Internet, web search engines, basic coding, gaming systems, virtual assistants — and even autonomous vehicles that are now on our streets and highways and soon headed to the Moon and Mars.

The World Wide Web that linked “thinking” computers together via the Internet is only 37 years old, younger than many of our children and subject to just as much commercial faddism and corporate slavery.

“Web 2.0” happened in 1999-2000 and propelled the growth of user-generated content on the Internet and brought us Facebook, YouTube and Twitter and took away most of the remaining remnants of our personal privacy. That’s when we first became “friends” with Siri and Alexa, if you remember.

Now we have Claude and all the other chatbots that are taking over the daily conversations on the Internet, including our personal web searches and social media feeds. These “generative AI” assistants are now controlling the context and focus of our digital universe we once dialed up across slow-whirring modems with our AOL accounts.

We’ve just studied the structure of how LLMs and GPTs are put together to “create” deep thinking, human-like reasoning and to “transform” a string of spoon-fed data into large matrixes of instructions and tasks that create complex algorithms, new apps, college essays, Jeopardy game answers and kitchen cooking recipes. But, do you really want to know anything else about Generative Pretrained Transformers? We thought not.

Moving on to a low orbit future

“The internet is an open and global computer network that connects billions of users and devices without a centralized governance mechanism,” the introduction to a recent Congressional report on the internet’s architecture reads. “The emergence and advancement of artificial intelligence (AI) is transforming the internet and poses new questions (for Congress and others) to consider.” The internet includes a flotilla of low orbiting communication satellites, thousands of miles of undersea cables, rooftop dishes, copper wires, fiber-optic ribbons, acre-sized data centers — and your own router, smartphone and personal computer. Some of the internet’s hard-wired core and nerve centers are now more than 70 years old.

The biggest change since AOL dial-up days is the rampant expansion of “cloud computing.” Cloud computing allows individual internet users to gain access to seemingly infinite server time and data storage. The “Cloud” allows 24/7 admittance to all the “networks of network” that make up the global internet and World Wide Web. The advent of cloud computing allows even the lowest level user to remotely access powerful data processing chips, data storage devices, all kinds of software (apps), infinite databases, analytics and all the chatbots mentioned above. These all come with “free” updates.

Don’t you think it’s ironic that when more people make more use of the “cloud,” that it requires using more and more real estate, electricity and water to build the hyper-sized data centers that actually don’t float in the sky, but are being built all around us — closer and closer to our own neighborhoods?

Right now, there are 2,594 large-scale data centers operating in the United States, with another 1,400 on the planning boards. This does not include thousands of more smaller and dedicated data centers. Amazon, Google and Microsoft own most of the bigger (hyperscale) data centers. All these internet-connected and cloud-dedicated data centers use almost five percent of the United State’s total electricity output. That total is expected to double by the year 2030.

All this energy gluttony and related environmental concerns is why there’s so much talk by Elon Musk and others to expand future data centers to the Moon or in low orbits around the Earth. This would bring a more fitting meaning to “cloud computing,” don’t you think? But there are a few outer space obstacles such as designing a cooling system and adding protections against gamma rays and micrometeorites.

Meanwhile, back here on Earth, we’re wondering what if GPT really stood for “God’s Practical Trick?” Either way, it looks like us humans may become the next dinosaurs.

— Rollie Atkinson

7-3-2026