I’m not going to explain what GPT-3 stands for, because it doesn’t matter. What matters is what it is and what it can do. And what it can do might surprise you.
First of all, if you haven’t heard, GPT-3 is a machine learning model that can write text. Text that looks like a human wrote it. Almost out of thin air.
It’s a cutting-edge, gigantic machine learning model that’s been trained on an enormous amount of data, which makes it extremely flexible and versatile. It’s not perfect, but the writing it produces is often surprising because of the quality and creativeness that comes out.
In fact, some of this podcast was written by GPT-3. I’ll let you decide which parts are me and which parts are the AI. In all honesty, the AI probably wrote the better text.
But how good is GPT-3, really? Does it live up to the hype? Maybe yes, and maybe no. It depends on the hype you’re listening to. But one thing is for sure—it’s certainly not something to be ignored.
As CEOs, we’re always looking for force multipliers. Ways we can get better results using less resources. In the music world, a good example would be synthesizers and pitch correction tools. They help you make more music, faster, that sounds better.
GPT-3 is that same concept, but for writing. And writing is powerful.
It’s in ad copy, articles, blog posts, product descriptions, podcasts, emails, social media posts, and on the back of Old Spice body wash bottles. It’s one of the highest leverage activities you can do for your business. If you can write well, you can persuade well. But writing well is difficult and time-consuming.
Enter GPT-3. It won’t write a perfect 500-word essay for you. It’s not quite that good. Leave that to GPT-4 (maybe). But it will certainly make your writers more productive.
Since OpenAI, the organization responsible for creating GPT-3, opened up the API last year so products could be built using its GPT-3’s capabilities, the start-up scene hasn’t disappointed. We’ve seen everything from apps that can write basic code for developers, to a website plugin that will automatically A/B test your pages and suggest changes.
But looking across the landscape of tools that have been built, you notice something striking. The vast majority of them deal with one single problem—generating unique, useful copy, quickly.
These apps are called things like Copysmith, Copy.ai, Snazzy, Drafter, Ryter, Writesonic, Shortly, and Headlime (both of which just got bought by another one called Conversion AI, which then subsequently changed their name to Jarvis). And more keep popping up.
Clearly, there is demand. In fact, one of the frontrunners (Jarvis), has already amassed 30,000 users and seems to be iterating their product at a blazing rate (whether through their development team or their seemingly active acquisition team).
So what does this mean for you? What does this mean for your business? What can these tools actually do, and where do they leave something to be desired?
First of all, they can solve the copywriter’s dilemma. These days, you need quality copy, but you also need it fast. Every minute lost waiting for your writer is a minute you’re not selling. I know what I’d rather have—more sales. And so do these tools. By automating most of the copywriting process, they allow your writers to focus on strategy and content creation rather than repetition and formatting.
The real opportunity is in the ease of use. GPT-3 doesn’t replace your writers. It augments them—making writing faster and more accessible without compromising your results. The last thing you want to do is compromise the quality of your content because copy is an exponential business multiplier. The more you can apply it, the more business you’ll get. Despite advances in technology, one thing that hasn’t changed is that people still talk about products and services—and what they say matters.
Whether we like it or not, the copy game on the internet is changing, quickly. You can argue that AI writing isn’t human writing, that it should be used to create content, and that it’s going to increase the amount of spam everyone has to deal with (except now you may not know the difference between spam and what a non-spammy human wrote).
But the fact is, AI writing is here to stay, and the internet itself is going to shift as the technology grows and gets better. Today it’s a force multiplier. Tomorrow, perhaps it will replace human writers altogether. It’s hard to say.
Some welcome this revolution, some run from it. Whatever your stance is, you at least need to be prepared for the consequences of it.
And in the end, what matters is that the content that we produce is factual, helpful, and useful to human readers. If it can help them answer a question, find something they need, or take useful action, then the content did its job, whether or not it was partially or mostly written by an AI.
So you have to ask yourself in this case, do the ends justify the means? Or do the means justify the ends? And if that doesn’t make any sense to you, you can ask GPT-3 to explain it.