Transcription
Why are there all these copyrighting A.I. tools popping up all of the sudden, and why are some of them dirt cheap while others cost thousands of dollars per year? Why don’t I have as many subscribers as WP crafter? I have way more hair. As with most things, the answers come back to you. Elon Musk, of course. Now, back in 2015, Elon Musk and a few of his billionaire buddies started to open a research lab that had the goal of promoting and developing friendly A.I. in a way that benefits humanity as a whole.
What could go wrong? Well, where the potential death spiral begins is in twenty eighteen. When Musk left the board of Open AI over a potential conflict of interest with Tesla’s own AI efforts, then in twenty nineteen open AI transitioned to being a nonprofit, to being a for profit. Then they quickly signed a billion dollar deal with Microsoft man that got out of control quickly. But here’s where the part that is relevant to you comes in. Back in June of twenty twenty, while the entire world was shut down open, I announced three, a language model that’s trained on literally trillions of words that it got from the encyclopedia.
No, the Internet. Could you find a worse source? People most importantly, open Nusii also released the API, which is literally the connector that feeds dozens of these A.I. copywriting tools that you keep seeing all over the place. Now, we’re going to talk a lot more about the API because that’s the real heart of the matter. But first, I want to emphasize what a big deal that three really is. Now what is it? So it’s it’s aimed at natural language, answering of questions so you can ask it questions in natural language and it just answers them.
And it also can translate between different languages and it can improvise text that’s actually coherent. It really is the stuff of sci fi dreams. It’s so impressive that before three, the largest language model was Microsoft’s own Turing and LG. Now Turing and LG was no slouch. It was capable of 17 billion machine learning parameters. Any time you get into the tens of billions of things, that’s a lot to count, right? Well, three is capable of one hundred and seventy five billion machine learning parameters.
So literally over ten times an improvement over the previous champion, Microsoft. Of course, Microsoft made this deal with GP2. So you got to wonder what’s going on there. But what’s even more important is that Turing came out in February of twenty twenty and he came out in June of twenty twenty. So just a few months later, a 10x increase in technology is pure alien stuff. Thanks, Elon. The output of three is so good that it’s honestly difficult to tell that it was created by a computer.
But what’s interesting is that the three output also struggles with stuff that humans find very easy, things like factual knowledge. But what it’s really good at is writing creatively and persuasively skills that only a small percentage of adults actually have developed and maintain. There is no doubt in my mind that three is here for good, the same way that drum machines, synthesizers and pitch correction are never going to be erased from the planet, no matter how much us old guys complain.
So for better or for worse, you can expect to be sold some stuff by A.I. sooner than later. You probably already have, honestly. All right. So back to the open A.I. timeline. Well, Microsoft struck another deal with open eye in September of twenty twenty. This time it gave them an exclusive license of GTP 3s underlying code. That’s right. Microsoft gets the actual underlying code while everyone else is stuck with the API. So what is the API?
Well, you’ve probably heard the term API before. It stands for Application Programing Interface. It’s basically software that allows two applications to talk to each other. So it’s basically like Jarvis connects to the API, which connects to GTP three. So all of the applications built with the API should basically be the same. Right. They’re just connecting to an API and spitting out data from a service. Right. So should you just buy the cheapest one and move on?
Well, it’s not quite that easy to start opening eyes. API is made up of several different engines. They’re called Da Vinci, Tury Babbage and ADA. Each of these engines has its own specialty, but Da Vinci is generally the most capable. So that sparks the question that if you made a product that used the API to connect AGB three, why wouldn’t you just use The Da Vinci engine? Well, obviously it’s because each engine has its own pricing.
So let’s say as a consumer, I purchased a thirty nine dollars copywriting product that I bought at the local LTI yard. The more that I use that product, the. More API requests that product is going to need to make. So they’re definitely going to need to optimize those requests and choose the less expensive engines whenever possible. And no matter what overtime, if I continue to use that product, I won’t be a customer of theirs. I’ll actually be a liability.
I’ll be costing them more money the longer I stay active with the product. So basically, they’re going to need to either sell me more things, create more expensive plans, or get me to leave the business so that I don’t bankrupt them. And if I don’t and there’s a lot of people like me, eventually they’ll go out of business if they’re not able to generate some monthly recurring revenue to pay their API bills. So for this simple reason alone, a lifetime deal for an A.I. copywriting tool is probably not the wisest investment, in my opinion.
So should we just look for a monthly tool that only uses the Winchester bombarding developers saying, hey, what engines are you connecting to? Well, it’s not that simple either. These products aren’t going to just tell you what they’re up to. You behind the scenes, how they’re utilizing the API is their secret sauce. It’s literally what gives each product its own personality when it really comes down to is the developers who are writing the software that connects up to the API, that talks to GTP three.
How well are they able to navigate the API to be able to pull the best results out of three and then make it enjoyable to use make it so that it’s user friendly for the average person like you or I to interact with the software? Now, if you’re into this tech and you see the immense potential in increasing your creative output with computer assisted writing, you’re going to want to subscribe to this channel right now, because this is just the first video in a multiple part series where I’m going to go deep into the top three writing tools.
I’m going to be creating dedicated reviews to compare and contrast the top tools from Jarvis, which is formerly Conversion.ai, Copy.ai, Copysmith and Writesonic. The criteria that I’ll be using when I review these products is looking at how useful they are to create long form blog content or how useful they are to create long form stills content. I also want to see how easy it is to create shorter content, things like social media content. But most of all, we want to check to make sure that it’s passing any sort of plagiarism test.
Is this content coming back as original? Then I’ll be looking at other things like features and integrations, the number of templates they have available to them, and then just overall value, because there is such a wide disparity in terms of the cheapest tool being, you know, just a one time fee of under 50 bucks to other tools, literally charging thousands of dollars per year. So like the lumberjack whose ax was replaced by a chainsaw, the music producer whose tape machine was replaced with pro tools and auto tune, or just the cashier who is replaced by the self checkout lane three is already changing the Web for better or worse.
So I think you owe it to yourself to at least have an eye on how the shape of the Internet is changing, even if you decide not to use it at all. I’ll see you in the next video.
