An Epic Human-Machine Write-Off

If you didn’t catch the latest Internet fad captivating our short attention spans, let me fill you in. On November 30, the non-profit OpenAI released ChatGPT, a quirky AI-powered chatbot that will change the concept of homework writing forever. 

ChatGPT is a large language model trained on the mountains of information available online. It comes from the same company behind DALL-E, an image generator that made the tech-media headlines this summer. It’s also the next iteration of the text generator GPT-3. 

After signing up for ChatGPT, users can ask the AI system to answer questions or generate content based on “prompts.” From birthday wishes to school essays, the thing can even write usable code. But what it most definitely can do is the same thing that I do, write marketing content. 

Gasp! Here comes a low-flying panic attack. 

I could get all philosophical as many authors have, but I’m just going to focus on the question every normal self-centered human being working at a laptop right now is covertly asking themselves: Is AI a better writer than I am, and will it take my job? 

On first glimpse, my reaction was PFFT. Unless somebody broke all ten of my fingers, I could be hungry, underslept, cold, and hungover and still beat the shit out of that “robot writer.” Then I decided to conduct an experiment. 

Let’s have the AI take its best shot at writing about the things I’ve already written, a couple of this year’s Infinum newsletters. An epic human-machine write-off. 

Source: dall e

Competitor #1: Mia, 38-year-old human female, content writer with a degree in journalism and English.

Competitor #2: ChatGPT, chatbot based on the large language model GPT-3.5 trained by OpenAI

Exhibit 1: February 2022, Make Me Stream

Writing about how streaming platforms influence our relationship with video and music content, I used the example of Disney’s Encanto and the incredible popularity of the hit song from the animated picture, We Don’t Talk About Bruno. Here’s what I wrote:

Streaming has this characteristic that is especially favorable for children’s content. You can play it again and again. And again. 

“Bruno” was played 100 million times on Spotify by January 20, while the entire soundtrack had 494 million all-time plays globally by early February. YouTube count is 186 million at the time of writing this. I won’t even go into TikTok’s trending videos because I’m exercising my right as a millennial not to.

I prompted ChatGPT with “write me an introduction to a column about the popularity of streaming platforms, journalistic style, funny.” Here’s what it wrote:

Welcome to the streaming revolution! It seems like every day, another streaming platform pops up, offering endless hours of content for us to bing-watch until our eyes go square. But with so many options out there, it can be tough to keep up with the latest and greatest. That’s where we come in. In this column, we’ll be exploring the ever-changing world of streaming, from the biggest players like Netflix and Hulu to the niche platforms you may not have heard of. We’ll take a closer look at what sets each platform apart, what kind of content they offer, and which ones are worth your time (and your money). So grab some popcorn, find a comfy spot on the couch, and let’s dive into the world of streaming!

What’s bing-watch, ChatGPT? Ok, if I don’t get too nitty-gritty about spelling mistakes, this could work with some minor editing. It could be more specific, but I couldn’t ask it about Encanto because its knowledge ends with 2021. It’s not half bad, but I’ll say I’m better.

Mia : ChatGPT 1:0

Exhibit 2: May 2022, #neverstop branding

Infinum did a major brand revamp and a new website launch, and I did a special on tech companies rebranding over the years.

Some of the companies that produce the devices and the software we spend far too much of our waking time glued to, have been around for a very long time. You can’t be around for a very long time and always look the same. Just ask Madonna.

Here’s what ChatGPT had to say on the subject:

Tech companies are always looking to make their brands stand out in a crowded market, and one way they do this is by changing their visual identities. Over the years, we’ve seen some major transformations, like when Microsoft decided to ditch its old, boring logo in favor of a new, hip one that looked like it was designed by a high schooler in their graphic design class. Meanwhile, Apple has gone through more logo changes than Taylor Swift has had boyfriends. And let’s not forget Google, whose logo looks like it was drawn by a toddler with a box of crayons. All in all, it’s clear that tech companies are willing to try just about anything to stay on trend.

This is very good, but unfortunately, my editor is a huge Taylor Swift fan and definitely wouldn’t appreciate the sexist reference. It’s also not entirely accurate because Apple did make many logo changes, but they were so minor that their logo has practically stayed the same for 40 years. This is a problem with ChatGPT called hallucination; it makes things look very plausible while, in fact, it’s all bluff. 

Exhibit 3: July 2022, Sense of Impending Doom

Catching the zeitgeist, I wrote about the flood of bad news that struck us this summer and how every crisis is an opportunity for growth.

Did you know that the sense of impending doom is an actual medical symptom?

Seriously, if you feel like something bad is about to happen, it may be a sign of an oncoming heart attack. However, it may also be a sign of being in a state of perpetual moderate panic because so many bad things did in fact happen recently. 

What the actual F, world?

I gave ChatGPT very literal instructions. “Slightly funny introduction to a column about the sense of impending doom felt because of actual bad things happening in the world.”

As we navigate the current state of the world, it’s hard not to feel like we’re living in the end times. Between the ongoing pandemic, the climate crisis, and the constant stream of bad news on our screens, it’s no wonder many of us are feeling a sense of impending doom. But fear not, dear reader! In this column, we will explore the reasons for this feeling and offer some tips on how to keep the doom at bay (or at least hold it at arm’s length). So grab your apocalypse survival kit and join us on this journey to the (hopefully not) end of the world.

Ok, so it understandably didn’t know about the war, inflation, and gas prices. It’s still in pandemic times. I love the last sentence, though. You win this one, chat robot.

Mia : ChatGPT 2:1

Exhibit 4: September 2022, Are You for Real?

I used the newly popular social media platform BeReal as a segue to comment on social media features in general. Here’s a snippet about “likes” on Facebook.

Did you think that “likes” are a Facebook invention? Wrong. Back in the early 00’s, there was a site called B3ta.com, a newsletter and a message board (these things were popular back then), a place to share jokes or images that would today be called memes. By 2005 the community grew so large that it was becoming difficult to navigate the front page, and lots of the “good stuff” would get lost. They needed a way to curate content et voilà – they came up with a button that said “I like this.” It wasn’t completely original, though. Even before B3ta, FilePile, and Digg used buttons to measure users’ approval. 

Facebook introduced the like button in 2009, but not before it bought FriendFeed, a service that had already had its own like button for several years.

Here’s the ChatGPT version:

Ah, the good old “like” button – a staple of the social media experience. But have you ever wondered where this little icon of approval came from? It all started back in the wild west days of the internet, when the cool kids were hanging out on MySpace. Yes, that MySpace. The site had a “thumbs up” button that users could click to indicate that they liked a post, and this feature quickly caught on with the other cool kids on the block. Fast forward to 2009, when Facebook introduced its own version of the “like” button. And the rest, as they say, is history. Nowadays, you can find likes on just about every social media platform, from Twitter to TikTok. So next time you double tap that heart icon on Instagram, just remember – you’re participating in a time-honored tradition of internet approval.

Again, it’s exhibiting the hallucination problem. MySpace didn’t even have the like button. Or definitely not in its glory days. But wow! That last sentence, a time-honored tradition of internet approval. Fire emoji reaction for this one, ChatGPT. 

Mia : ChatGPT 2:2

Enough with the competition, I stand humbled. Ok, I was both a competitor and the judge, but you catch the drift. This is one very capable bot. Still obviously a bot, but with mind-blowing writing skills.

The human take

We’ve had some fun with robochat, but it’s time to get serious. Chat GPT crossed over one million users in only five days, something that took Instagram two and a half months and Facebook ten months. 

While everyone is going ha-ha, look at the stupid mistake it made, we’re glossing over the fact that it’s sometimes wrong and prevailingly right. It’s also worth mentioning that it takes 3-5 seconds to write a couple of paragraphs, while I take… much longer than my team would like. And it doesn’t require health insurance, vacation time, team syncs, or coffee breaks. It doesn’t have to be perfect; it just has to be more efficient than you. 

And it will get better. This technology took years to build, but now that the groundwork has been put in, the acceleration will be incredible. GPT-4 is already in the works, and we can expect its release in the next couple of months. 

Mark my words, it’s a revolution. The last decade was marked by the shift to the mobile paradigm, and this one will be marked by AI. Once the technology is there, other companies can build their own tools on top of it. And I assure you, they will. By the end of next year, Forrester predicts that 10% of all Fortune 500 enterprises will use AI to generate content.

Here’s a portrait of me in five years, befittingly commissioned from DALL-E.

source: dall e

I jest, again. But before I get any more why-so-negative comments, I’m not being negative at all. I am truly and honestly amazed by this accomplishment. Changes are inevitable, and technology changes lives. The industrial revolution changed lives two hundred years ago. Maybe I’ll be begging on the street while AI is working the job it stole from me, but maybe I’ll be working half the time for twice as much. Maybe we’ll all be.

Tech vs. Product

Why are we talking about ChatGPT with such fervor? It’s actually not a new thing. Chatbots have been around since, believe it or not, 1966. The first version of this particular model by OpenAI, GPT, was released in 2018. So why is everyone talking about this now?

Because ChatGPT is an amazing product. While the technology has been around for a while, OpenAI decided to release the latest version of their model publicly, packed in a super simple user interface. When you read an article about some AI writing essays, you think it’s mildly interesting and scroll on to the next clickbait. But when you actually get to prompt it yourself and have it write a farewell card to a colleague you hate, you are amazed. 

The reason for this can be explained in two letters – UX. You can create ground-breaking, world-changing, industry-disrupting technology, but if you don’t know how to package it so that it appeals to people – good luck on the market and close the door on your way out.

This is an excerpt from our monthly-ish company newsletter. If you would like to see a story like this in your inbox, subscribe below.