What Is ChatGPT and How Does It Work?

Seems everywhere you go on the internet people are finding out about ChatGPT and how fun, interesting, and practical it is to use ChatGPT for everything from fixing complicated programming code to helping a fourth grader write a better history paper. But what is ChatGPT and is it really worth all of the hype?

ChatGPT Is Awesome

Let’s start from the beginning ChatGPT is amazing. It is so amazing that some have already started to fear it. (It’s the same with noticeable advancement in technology.) There are people worried that ChatGPT will take our jobs (I’ve heard this before…). There are others that worry it will be used for terrible purposes, and then amongst those of us who read and watch a lot of science fiction there might some worried that it will take more than our jobs. (See also Terminator, Terminator 2, and so on.. For something a little more off the beaten path, try Robopocalypse)

The truth is that ChatGPT still has a long way to go before it can affect everyone in any matter. Sure, there are a lot of things it can do now. Once, there were “a lot” of things your phone could do, and that without a keyboard. As for that fourth grader, this is just the latest change that education must adapt for. (My physics teacher at University of Colorado was a little too slow on the draw, and I used the brand-new memory area on my TI-84 graphing calculator to store all the formulas. The one my kiddo has not only graphs, but it does also so in color, and it can run real programs. Relax. Kids are not passing physics without learning anything.

What Is ChatGPT Exactly

OpenAI

OpenAI is an open-source project trying to develop artificial intelligence for the greater good. Seriously. It does have an odd structure so that the people doing all the work can get paid and get paid well. OpenAI Inc is a non-profit organization. OpenAL LP is a for-profit organization. Basically, the OpenAL LP can charge people and use those funds to pay their people and also convertible notes, stock options, and even investments from other companies (Hi, Microsoft!). But, once that happens, then whatever is left of the money funds a benevolent open-source AI program.

GPT-3

If music hipsters walk around telling people that they knew about that band before the mainstream did, then whatever the equivalent of a hipster nerd is pointing out that it isn’t new. They are right, but it doesn’t matter. Just in case you want to know what you are playing with, let’s do the short-short version of an explainer.

OpenAI is not new. It’s been working on and sharing its AI technology for several years now. Among the many releases was the recent release of GPT-3, which, you guessed it, is the third version. It’s been around for months.

So, why are you just now hearing about this AI thing?

Using GPT-3 was not difficult, but it wasn’t easy either. There was a learning curve, and understanding how to ask so that you get what you want as an answer takes a bit of practice. GPT-3 is great. It just didn’t have a soft fuzzy exterior for the public.

ChatGPT

As the name might suggest, this whole is basically bolting a user-friendly interface to the GPT data. However, do not assume this was easy. ChatGPT not only answers even the most malformed questions, it remembers the conversation you are having so you can literally say, “Oh whoops I meant April,” and it will do whatever you asked it too before, just with April this time.

Yes. It is amazing and don’t let any Silcon Valley naysayer tell you otherwise. ChatGPT is free for now, but it takes an awful lot of computing power to do what they are doing, and it can’t stay free in this way forever. Fortunately, it looks like the public will keep its free access, just limited or slower, or whatever than the paid ChatGPT option.

For now, use it to get answers to things you always wanted to know. Get descriptions of places you want to go. Heck, have it plan the itinerary. Just keep in mind it doesn’t know everything, and most importantly, it doesn’t know it should just say, “I don’t know,” when it doesn’t have enough data. Trust but verify sounds like the right strategy. Afterall, if this is the beginning of a robot revolution, we’ll need an arms policy.

Oh, and while it doesn’t notice, it never takes too long to type “please” or “thank you.” (We might lose that war against the machines, and it would be better to be on the nice side than the mean, jerks’ side.)

This is what happens when they say, “Just throw it together. We can fix it later.”

Leave a Comment