I’ll be honest. I was feeling pretty happy with myself. Yes, 2020 had dealt me (and almost everyone I knew) a handsome serving of lemons, but I felt confident that the proverbial lemonade I was able to muster would stand me in good stead in the months and years ahead. After all, I was consuming 34% less late-night A&W takeout per month, running at least 19% more each week, and slowly (but surely) increasing my understanding of the automation and AI worlds. All in all, I was in a relatively happy place despite the omnipresent “C-word”—that was until I heard about GPT-3!
Wait, it does what?
“What’s GPT-3?” I hear you ask. Great question. So great a question that if you had asked me it a month ago, I would have hazarded a guess that it was either a character from Star Wars (sorry, not sorry) or maybe a cute nickname for a baseball player! One thing is for certain; I would not have known that it would have had the potential to question my entire existence as a content strategist.
In brief, GPT-3 is OpenAI’s (a non-profit artificial intelligence research company backed by Elon Musk, and others) latest language prediction model and, in theory, it has the potential to create everything from young adult fiction to sweetly designed code as effectively as a human could. Wow! Needless to say, its arrival made a splash in certain quarters of Techie Twitter.
Given the hype that the launch of GPT-3 has garnered, I did begin to ponder what the arrival of GPT-3 would mean for copywriters, UX designers, and content strategists like myself. Would we be made obsolete by a clunky-named piece of AI? If so, should I cancel my subscription to The New Yorker, and finally stop trying to convince myself that I actually like black coffee? As appealing as it was to try and unpack these questions in more detail, I ultimately decided that what I really needed to do was face my fears head-on.
So, what did I find out?
But do you really get it?
When assessing GPT-3 it is important to acknowledge what it is and what it is not. At its core, GPT-3 is an extremely sophisticated text predictor that can generate detailed and nuanced content based on statistically plausible responses to the input it has been given. Perhaps the most impressive facet of GPT-3 is its size. The model has over 175 billion language parameters to pull from. To put that figure into perspective, its predecessor model GPT-2 had 1.5 billion parameters. However, despite this size, it is not without its faults.
"Given the hype that the launch of GPT-3 has garnered, I did begin to ponder what the arrival of GPT-3 would mean for copywriters, UX designers, and content strategists like myself. Would we be made obsolete by a clunky-named piece of AI?"
One vital flaw is that GPT-3 struggles to fully comprehend the language it creates. In other words (sorry!), it cannot reason abstractly and, as such, has a common-sense shortcoming when it comes to understanding the semantics of the terms it is using. This simple test by Kevin Lacker exemplifies this point quite succinctly.
On top of these shortcomings, OpenAI’s most recent iteration has also encountered some pretty serious teething problems in terms of a series of in-built sexist and racist biases that definitely need to be ironed out before it can look itself in the mirror. This is a very serious concern and one that needs to be addressed before it can integrate ethically into both the tech world and wider society.
So, now that we have a clearer idea about what GPT-3 can (and can’t) do, where does that leave the modern-day content strategist?
Content then marketing
As someone with over a decade of experience writing content to pay the rent, I have seen my fair share of fads (hello, listicles!) come and go. However, one thing that remains constant is the need for engaging and insightful content, regardless of the media form it takes.
In my role as Content Strategist for Wrk, I find myself spending more and more of my free time searching for the sweet spot between tech and content. These journeys often result in the discovery of new podcasts or webinars. On a recent voyage, I came across a content marketing episode from Inside Intercom that sought to debunk some of the most commonly-held myths around content marketing.
In essence, the Intercom team emphasized the importance of a content-first approach to marketing, and this point really resonated with me. They advocated eschewing the typical marketing copybook for a more authentic approach to content development that included a clear and coherent approach to content design, emphasizing brand relevance at all times, and not being afraid to have an opinion.
Frankly, I found this authentic approach (one which is far removed from the alphabet soup of catchy marketing acronyms) to be extremely insightful, and I have even adopted some of these principles into Wrk’s content strategy.
From a bigger picture perspective, I feel there is a lot for content folks to take from this innovative approach. In fact, by digging deep and delving into what really defines your brand, you are not only increasing the likelihood of creating a strategy that stands out from the crowd, you are also illustrating your very own invaluable set of skills to your team—skills that GPT-3 does not have the nuance or sophistication to replicate.
Use the tool, don’t let it use you
Like almost every aspect of automation, GPT-3 has the potential to be a game-changer in the content marketing sphere. That said, it also has a number of flaws and limitations (OpenAI even admits that), so instead of cowering from GPT-3, why not embrace it and let it help us become better writers and content developers in the process. After all, most of us are already using Grammarly, Hemingway, and other content automation tools that help us do our job more efficiently and effectively every single day.
"In fact, by digging deep and delving into what really defines your brand, you are not only increasing the likelihood of creating a strategy that stands out from the crowd, you are also illustrating your very own invaluable set of skills to your team—skills that GPT-3 does not have the nuance or sophistication to replicate."
The reality is that working with automation is something that many craftspeople (across a range of disciplines from graphic design to web dev) have had to grow accustomed to in recent years as tools surface to help their respective professions grow and evolve. But, in general, the vast majority of these tools have been a value-add to the field they’re operating in, and almost never result in replacing a skilled worker.
Finally, if history has anything taught us anything it is that the power of a skilled, passionate, and ambitious professional is rarely bettered.
Now, where did I put that sourdough recipe?
Until next time,
Featured image Photo Credit: Ernest Porzi