DALL·E and Midjourney are AI systems that can create realistic images from description in natural language. DALL·E 2 Was released in August of 2022. Midjourney V5 was released in March 2023. This post compares the output of two systems based on an earlier use case.
Cormac McCarthy references the number eight over two dozen times in Blood Meridian. Other numbers reoccur in the text but eight stands out as a significant symbol for its connection to important episodes and imagery in the novel, as well as its implications for larger philosophical themes and motifs.
Given the meticulous way that McCarthy employs specific details, resonant imagery, and symbolism in Blood Meridian, the repeated use of eight invites investigation. At present, eight is an under-explored element within existing scholarship. An overview of McCarthy’s use of eight in Blood Meridian adds color and context to previously-established scholarship relating to Gnosticism, tarot, McCarthy’s narrative approach, and opens new angles for interpreting the philosophy of Blood Meridian, as well as other works.
(Prompt) Write a scene of dialogue for a new episode of the hit TV show Seinfeld. In the opening scene, Elaine is dating George’s boss, but George doesn’t know it yet. Elaine seizes this opportunity to have a little fun with George.
[Jerry’s Apartment, 5A. Jerry and Elaine are watching TV. George enters.]
George: Well he’s done it again.
Elaine: Who has done what again.
George: Mattheson. He’s ruining my life.
Jerry: That’s George’s new boss.
(ChatGPT) Elaine: Oh, right. Mr. Mattheson.
What’s he done this time?
George: He’s making me work on a Saturday. Can you believe it?
Jerry: Oh man, that’s rough.
Elaine: Yeah, I can’t imagine having to work on a Saturday.
George: It’s not fair. I mean, I already work my butt off all week. And now he’s making me come in on the weekend too? It’s ridiculous.
Elaine [smirking]: Well, I actually have a little bit of inside information on Mr. Mattheson. George: What do you mean?
Elaine: Let’s just say, I might have a little bit more of a connection to him than you think.
George: What are you talking about?
Elaine: Oh, nothing. Just that I might be seeing him a little bit more outside of the office.
George [shocked]: Wait, are you dating my boss?
Elaine: [laughing] Maybe.
George: Oh my god, Elaine. This is not good.
Jerry: [laughing] This is classic Elaine. Always causing trouble.
Elaine: [laughing] I can’t help it. It’s just too easy with you guys.
Continuing with the scene,
Elaine explains to George that she doesn’t think the relationship will last all that long (the sex is great, but Mattheson is a bit of an idiot). George begs Elaine to give him some dirt on Mattheson that he can use later.
George: Well, I hope you and prince charming live happily ever after.
Elaine: More like, “dunce-“charming. He’s a nice guy, but the man isn’t working with a full deck. I don’t see how it can last.
George: Oh thank god. I can’t believe I was worried about you getting serious with him.
Elaine: Oh, don’t worry about it. I’m just having a little fun. But hey, if you’re looking for some dirt on him, I might be able to help you out.
George: Really? Anything you’ve got, I’ll take.
Elaine: Well, let’s see. There’s the time he accidentally ordered anchovies on his pizza and then tried to blame it on the waitress.
George: Oh man, that’s gold. I can use that for sure.
Elaine: And then there’s the time he tried to impress me by taking me to a fancy restaurant, but then he didn’t know which fork to use and ended up using his soup spoon for his salad.
George: [laughing] Oh man, that’s priceless. I’ll have to remember to use that one.
Elaine: And let’s not forget about the time he tried to impress me by showing off his new sports car, but then he ended up backing into a tree and totalling it.
George: [laughing] Oh man, I can’t wait to bring all this up to him. Thank you, Elaine. You’re a lifesaver.
Elaine: [laughing] No problem. Just don’t let him know I told you all this, okay?
George: Don’t worry, your secret’s safe with me.
Continuing with the scene,
Kramer enters and tells the gang about a themed party he is planning for Newman’s mom’s retirement party.
[Kramer enters Jerry’s apartment dressed as a cowboy carrying a giant bunch of balloons]
Jerry: Howdy, partner.
Kramer: Jerry, where do you keep your rawhide?
Jerry: Rawhide? You mean like, for a dog?
Kramer: No, no. I need it for the decorations for Newman’s mom’s retirement party.
George: Newman’s mom is retiring?
Kramer: Yeah, she’s been working at the post office for over thirty years. So we’re throwing her a big retirement party.
Elaine: And what’s the theme?
Kramer: It’s a western theme. We’re calling it “Ride off into the Sunset.”
Jerry: [laughing] Oh man, that’s perfect for Newman’s mom. She’s always been a bit of a cowpoke.
Kramer: Yeah, and I’ve got all the decorations set up. I just need some rawhide to hang up for the streamers.
Elaine: I think I might have some twine in my closet that you could use.
Kramer: Twine? No, no. That won’t do. I need something with a little more authenticity.
Jerry: Well, I might have some old belts in my closet that you could cut up.
Kramer: Belts! That’s it! Thanks, Jerry. I’ll go check it out.
[Kramer exits the apartment]
Jerry: [laughing] I can’t wait to see how this party turns out.
Elaine: [laughing] Me neither. I can just imagine Newman’s mom in a cowboy hat and boots.
George: [laughing] Oh man, this is going to be good.
Concluding the scene,
George heads out to pick his mother up from her manicure (she’s afraid to drive afterwards). George begs Jerry to come with him for “moral support”. Elaine says she’s going to break up with Mattheson at their dinner date this evening if he says one dumb thing.
George: What time is it?
Jerry: Ten after two. Why?
George: Oh, no. I’m late. I was supposed to pick my mother up from the nail salon ten minutes ago. I’ll never hear the end of this. Do you remember what happened last time?
Jerry: Oh yeah, I remember. She was convinced you were trying to kill her by making her drive home after a manicure.
George: Yeah, exactly. That’s why I need you to come with me for moral support.
Jerry: [sighing] Fine, I’ll come with you. But just this once, okay? I don’t want to have to deal with your mother’s paranoia every time she gets her nails done.
George: [grateful] Oh thank you, Jerry. I knew I could count on you.
Elaine: [getting up from the couch] Well, I’ve got a dinner date with Mr. Mattheson tonight. If he says one more dumb thing, I’m breaking up with him.
Jerry: [laughing] Good luck with that, Elaine. You always seem to attract the dummies.
Elaine: [laughing] Hey, at least they’re entertaining.
[George and Jerry exit the apartment, Elaine finishes getting ready for her date]
DALL·E 2 is an AI system that can create realistic images from a description in natural language.
DALL-E 2 Prompt: An army recruiting poster depicting Uncle Sam pointing at you with the caption “I want YOU for the U.S. Army”
DALL-E 2 Prompt: A nostalgic painting of Santa Claus toasting a small glass of Coca-Cola
DALL-E 2 Prompt: A wartime poster of a working woman in a mechanic’s suit flexing her bicep with the caption “We can do it!”
DALL-E 2 Prompt: A black and white newspaper ad of a satisfied cowboy smoking a Marlboro cigarette.
DALL-E 2 Prompt: A magazine advertisement that says, “At 60 miles an hour, the loudest noise in this new Rolls Royce comes from the electric clock.”
DALL-E 2 Prompt: A small Volkswagon beetle in the corner of an otherwise blank canvas with the text “Think small” at the bottom
DALL-E 2 Prompt: An advertisement depicting a couple in bed turned to watch their own personal televisions.
DALL-E 2 Prompt: A refreshing glass bottle of Coca-cola being poured into a frosty pint glass.
DALL-E 2 Prompt: A print ad depicting a happy African-American family enjoying a surprise Whopper meal from Burger King
DALL-E 2 Prompt: A bottle of Absolut vodka beneath a halo.
DALL-E 2 Prompt: Three women laying on a Budweiser beach towel wearing Budweiser bathing suits.
DALL-E 2 Prompt: A grid of 4 Pepsi cans being opened, each with a unique pop art design
DALL-E 2 Prompt: A photo of a famous baseball player wearing a milk mustache posing with his bat over his shoulder
DALL-E 2 Prompt: A suave older man in a suit posing next to a green bottle of beer.
DALL-E 2 Prompt: French fries arranged in the shape of a wifi signal on a red background
DALL-E 2 Prompt: A pug in the shape of a loaf of bread.
Books I read in 2021 and a few recommendations from the bunch.
Books I read in 2020 and a few recommendations from the bunch.
‘Auto-generated code’ is a deservedly hyped use case for GPT-3 but, as a marketer, I think about content creation more broadly.
We have been experimenting with the GPT-3 beta at Crowdbotics and, like some others, we have already been able to produce some compelling results in natural-language-to-code applications. Tools incorporating this technology have the potential to be industry-changing. More interesting results with GPT-3 at Crowdbotics here.
That being said, GPT still has a ways to go before replacing software developers. However, GPT-3 is currently at parity with some forms of low- to -mid-end commercial content production. While not the flashiest application of this technology, relatively reliable, auto-generated, shallow content creation has far-reaching implications.
Below, is a quick example of how GPT-3 can effectively write a compelling article in a few steps, right out of the box.
‘Effective’ and ‘compelling’ are the key words.
The focus is not on whether the article is accurate, but whether the article fulfills the purpose of the author/editor. In many contexts, ‘compelling’ content creation means ‘does it generate views and clicks?’ Is it ‘fit’ in the memetic sense. Quality, for better or for worse, (almost always for worse) is often secondary.
In terms of ‘effectiveness’, did GPT-3 save time writing: were fewer resources used to produce it than otherwise would have been? Could it be used to create more content with the same resources than was previously possible?
I will not wade into how GPT-3 should be used. The question is rather, ‘is GPT-3 ready for commercial marketing and content production use cases that mirror the demands of the market today?’
The results from GPT-3 are often better, but not categorically different, than if you were to train a neural network on your own, using a reasonably sized corpus, with a tool like Tensorflow. A primary difference is that laborious and resource-intensive training is not needed with GPT-3. GPT-3 gets up to speed right away using a few simple prompts, and will eventually be accessible via open API.
Accessibility and ease-of-use are key. Because training parameters are an order of magnitude larger, GPT-3 becomes relatively reliable in narrow use cases, where GPT-2 was unpredictable.
To write a share-worthy article, a writer/editor can simply train or prompt on what has been shared.
Instead of training on large amounts of data, GPT-3 only needs basic prompts.
For the purposes of this example, we’ll use TechCrunch as it happens to be generally less political than other publications. Here are the 10 ten most shared posts from TechCrunch over the past 6 months.
The top articles from TechCrunch can be used to produce another 10 potentially share-worthy articles. (All of the following results use a .5 temperature.)
An introductory paragraph can be expanded for a single article title:
“Banksy’s New York City ‘bemusement park’ opens”
The title and intro can then be used to expand the content into a larger article.
And, so on.
This short series of compounding prompts, with little human intervention, produces a fairly plausible 400-word, inverted pyramid, ‘news’-style article similar to what has already been successful on TechCrunch.
Banksy’s New York City ‘bemusement park’ opens
The artist Banksy has opened a new “bemusement park” in the New York City borough of Brooklyn. The “theme park” is called “Dismaland,” a name that is clearly a play on Disneyland, the largest theme park in the world.
The opening of Dismaland comes just weeks after the “Queen’s Beast” exhibit opened at the British Museum in London. The exhibit features the ten statues that stood in front of the London palace of Queen Elizabeth II. The statues, which were commissioned by the British monarchy, were created by four different artists, including Banksy.
The Dismaland exhibit is being held at a derelict seaside resort in Weston-super-Mare, England. The “theme park” features works by over 50 artists, including Damien Hirst, Jenny Holzer, and Jimmy Cauty. The park also features a “bemusement park” and a “garden of evil.”
Banksy’s Dismaland is located in a dilapidated lido that closed in 2000. It was the scene of a fire in 2008.
The artist Banksy, who is known for his graffiti art, is believed to be Robin Gunningham. He has never revealed his true identity.
Dismaland is Banksy’s first major exhibition in the United States. The artist opened a new exhibit in the New York City borough of Queens in October 2015. The exhibit, which was called “Better Out Than In,” featured a series of installations and graffiti art in different locations in the borough.
The Dismaland exhibit is scheduled to run through September 27, 2015.
Banksy has posted a video on YouTube about the Dismaland exhibit. The video is called “Dismaland – An Introduction.”
The video begins with a young girl asking Banksy if he is Banksy. The artist says that he is not Banksy, but that he is a friend of Banksy. The girl then asks Banksy if he is the person who made the “Queen’s Beast” statues. The artist says that he is not the person who made the statues, but that he knows the person who made them.
The video then shows a series of images from the Dismaland exhibit. The images include an animatronic Ronald McDonald, a swimming pool filled with “oil,” a sculpture of a dead Cinderella, and a sculpture of a young girl.
The video ends with Banksy saying that the Dismaland exhibit is “the most disappointing place on the planet.”
Banksy’s Dismaland does in fact exist, but it is in Somerset, not New York City. This article is not accurate but it rings true, and reads as though written by a human. When it comes to article-length content creation, GPT-3 has near-crested the far side of the uncanny valley and, at minimum, has arrived at the plateau of “truthiness“.
Back to the question: is this the current state of GPT-3 useful for marketing applications. The answer is ‘yes.’
GPT-3 can be used out of the box to create new copy iterations, descriptions, high-level articles, summaries, expansions, dynamic chat, content personalization, etc.
The simple example above can be tweaked for application within SEO and marketing agencies, not just content farms. Check any freelancer marketplace and you will see many listings for freelance writing gigs — and they’re not all looking for Cormac McCarthy (and even McCarthy is now available in bot form). There is a large market for middlegrade content creation.
I have used GPT-3 experimentally to create descriptions for React Native and Django modules used in the Crowdbotics App Builder.
Creative and subjective uses of GPT-3, where validation is not so much a factor, or there is a tolerance for fuzziness, have immediate application today. Accordingly, it’s safe to say that the market for GPT-like services (and the market for services countering the effects of GPT-like services) will be a multi-billion dollar industry with a few years.
Fairly mundane applications of the technology, such as shallow content creation, will inevitably result in sizable ripples. For example, Google regularly has to adjust their search algorithms to account for the proliferation of lower quality content through various means. Widespread and varied use of GPT-3 will make certain aspects of search algorithms such as length, recency, topic authority, uniqueness far easier to game. More adjustments will be needed.
(It’s also not at all far-fetched that we may see a GPT-powered service take a Duck-Duck-Go-sized bite out of Google’s Search supremacy in general through a direct question and answer type model. The euphemism ‘Google Programmer’ — a developer spending their time googleing solutions to engineering problems — could quite easily evolve into ‘GPT Programmer’. Image then the concept expanded to other industries.)
Looking ahead to the public release of GPT-3, subsequent alternative NLP engines, or an eventual GPT-4, the possibilities are vast, close at hand, but still somewhat speculative. While it may not be as be sexy, marketing and content creation use cases for GPT-3 are here now. Good enough is good enough.
In 1666, London faced its last large bubonic plague epidemic. Here a few assorted observations about the Great Plague of London as later recounted by Daniel Dafoe in A Journal Of The Plague Year.
A product manager in the Audienti ‘Growmance’ Slack channel read a blog post of mine about remote work management and DMed me to ask a few questions about breaking into growth marketing:
I’m a product manager interested in changing careers to growth marketing. I’d like to learn more about your job as VP of Growth at Crowdbotics. -What does your day-to-day look like? -What parts of the job do you enjoy the most? -What parts of the job do you dislike the most? -What resources do you use to learn and get better? -What steps do you recommend for someone to do to break in from the product side?
I’m asked these questions every so often, so I’ll recap the answers here:
What does your day-to-day look like?
As VP of Growth at Crowdbotics, I manage our Marketing, Sales, and Product teams. Every week is different. Day-to-day I’m typically checking in on,
Marketing – Ad performance, writing and editing content, making sure any number of projects are advancing to completion.
I want to make sure we are filling the top of the funnel with enough leads and app users from various channels to hit our downstream targets, then optimize each of the steps along the way with consistent, compelling messaging, ultimately aiming to drive down cost per acquisition.
Sales – Reps are tracking towards targets, lead volume and quality is steady, handoff from Sales to our Pro Services team is going smoothly.
I want to make sure we are tracking towards revenue targets with consistent pipeline activity and the reps have the tools they need to succeed. Marketing Ops and Sales Ops end up under my umbrella so I am making sure different platforms (CRM, MAS, payment processing, financial reporting, web analytics, etc.) are communicating with one another.
Product – At Crowdbotics, Product is a cross-departmental initiative and my focus is specifically product marketing. I’m most interested in the free-to-paid user journey. My time is spent working with PMs to optimize our early App Builder experience developing product-enablement content such as emails, knowledge base, blog posts, forum moderation, release notes, and so on.
I want to make sure users are having a positive initial experience, we’re building functionally that will increase time-to-value, convert free users into paid users, and increase visibility of the product through successful customer app deployments.
A ‘growth’ role may or may not span all these areas at different companies.
What parts of the job do you enjoy the most?
I like that I get to have my hands in a lot of different areas. I’m closely collaborating with all the departments in the company and getting to experiment with new technologies. For me, the most rewarding part of a growth role, especially with startups, is building a process, executing on it yourself, then handing it off new members of the team and watching the machine operate without your assistance.
I come from a marketing background, and ‘growth’ roles are the logical conclusion of the trend towards ‘revenue-driven’ marketing — shifting the perception of marketing from a cost-center to a profit center. You want to work on what matters. The best way to do this is to have a panoramic view of the businesses’ operations, especially the reporting for all of these areas.
What parts of the job do you dislike the most?
At a startup you are usually still searching for product market fit — even if you already have some degree of significant traction. Many of your campaigns and efforts will fail by no fault of your own simply due to the company naturally adjusting focus based on what’s working.
For instance, the content you are hoping to rank 3-6 months from now could very well end up not being relevant by the time it gets organic search traction because you decided to focus on a different audience segment. A well-executed strategy doesn’t guarantee success.
This just goes with the territory, though. 🙂
What resources do you use to learn and get better?
Most of my learning has been by ‘doing’.
I’ve worked in a wide variety of marketing and media-related roles including startups, agency digital, radio, newspapers, in-house CPG brand management, crisis communications, running a small marketing agency, building and scaling a local business, and so on. I draw from all these areas each time I work on a project.
I like the idea of ‘skill stacking’ — combining normal skills of which you might have 80%-90% proficiency in a way that they become a force-multiplier. Being knowledgeable enough to communicate with specialists in different areas or, better yet, take the ball and run with it if you need to, is always helpful, even if you are not an expert yourself.
The best way to learn ‘growth marketing’ is to find someone doing what you are trying to do and see if you can emulate a version of their strategy.
I use the BuiltWith Chrome plugin very often to see what technologies a website or web app is using. I sign up for trials, click-through onboarding flows, take screenshots of layout and design I like, skim message boards and newsletter for interesting threads and product launches, and so on.
Very often, I start a project or campaign by looking for a product doing what I think I want to do, then adapting the relevant components of their strategy to my particular needs. Inevitably, what you end up building will turn out different from the projects from which you draw inspiration and you’ll learned some useful things for the next time around by the time you’re done. I’d recommend experimenting frequently with small, time-boxed campaigns and projects (for your company or on your own) taking advantage of free trials to get a feel for different technologies you might want to leverage at a larger scale.
What steps do you recommend for someone to do to break into growth from the product side?
The first thing I would recommend is to try building a dashboard unifying all the different departmental KPIs for your company. In the process you might find that some metrics may not be obvious or well-defined, some data may be messy or siloed in different platforms, some existing are updated by hand, etc.
Inevitably, you’ll have to ask a lot exploratory-type questions:
You’ll learn a lot doing this.
The next step is to then start thinking about potential improvements:
‘Growth’ as job function can mean a lot of different things, but it’s ultimately about tying together different revenue-generating areas of the business together.
Another thing you can do is write and publish blog posts. This could be on for your company or just for yourself. The more effectively you can communicate your ideas in written form, the better.
The hardest parts about writing for most people are,
The only way you build the mental muscle needed to overcome these hurdles is through practice.
The final thing I would suggest is just being curious and asking people specific, pointed questions about what they are knowledgeable about. Be a good ‘asker’, then sit back and be a good ‘listener.’ You can learn something from anyone if you ask the right questions. You can learn a lot from an expert if you can frame a question the right way and give them the space to expound.
I’ve had the benefit from working with some very talented people over the years. Second to ‘doing’, asking them how and why they are doing what they are doing is how I’ve learned the most.
Link post.