#madscarypod

Prompt:

"Create an image of two young women standing apart in a barn; POV is from someone standing across from them in the doorway; the barn is wooden and furnished like the interior of an old saloon; there are bottles of liquor and glasses of champagne and beer on the bar counter behind them; parts of the barn are on fire behind them. One of the women (standing on the left) has long wavy brown hair pinned back in a 20's style; she is wearing a pink sleeveless v-neck silk dress and pearl earrings, there are blood stains at the hem of her dress; she is holding a wooden stake in one hand and an oil lamp in the other; she has nervous look on her face and a smear of blood on her cheek; she has a light skin tone. The other woman (standing on the right) has long black hair parted on the side and styled like old Hollywood; she is wearing small gold drop earrings; she is wearing a beige silk flapper style dress with blood stains near the neck; she is holding a flaming wooden torch in one hand and a wooden stake in the other; she has a worried look on her face; there are blood stains on her arm and some blood splatter on her forehead; she has tan skin tone. The style of the image is like a detailed graphic novel; realistic lighting; the image is square."

I was given the following warning after asking to include blood stains, or an antique revolver in the hands of one of the girls: "I couldn't generate that image because the request violates our content policies. If you’d like, I can help adjust the prompt or create a new one—just let me know how you'd like to proceed!"

Prompt: "line drawing of ghostly hands typing on a computer keyboard, writing a blog post." (Generated Feb 2024; Dall-E)

When we began this research project last summer, we wanted to try out any publicly available (and usually free) AI tools to help us with the various aspects of conducting an experiment, producing a podcast, and writing a horror screenplay. One we ended up using often are text-to-image tools to create cover images for our episodes, social media posts, and more.

Our first use case - as you'll remember from one of our first blog posts - was to create our Mad Scary Podcast logo and image. Some takeaways there was that text-to-image tools (at that time, ~9 months ago) struggled with generating human features (like hands and faces), symbols (words, letters, numbers - even crucifixes on gravestones), and contextualizing certain figurative descriptors (like "eerie" or"spooky"). While some of the results were interesting or to our liking, we were new at writing prompts (specifically prompts for image generations vs text generation) and didn't actually have a clear idea of what we wanted.

Prompt: "full moon at night over new york city, spooky, in the style of a graphic novel." (Generated Jun 2023; Canva Magic Studio)

Eight months later, we're still struggling with some of the same things, but have learned (surprisingly) that less is more. Giving ChatGPT or Dall-E a more detailed prompt, rife with notes on style or tone may work sometimes, but I've noticed that they often ignore details that appear later in the prompt text. I suspect this is because the later details may "contradict" the earlier ones, or may seem redundant. For example, if you included "realistic lighting" and then "high contrast" later in the prompt - perhaps the LLM disregards the second lighting cue since it already has one? Or, maybe it's more simple than that; it doesn't understand what the ask is, or the punctuation used might be off.

Below are 2 images produced (this month, March 2024) from the same prompt, generated in Canva Magic Studio (available with a Canva pro subscription):

The prompt for both was the following: "2 ghost girls, sitting at a desk writing on paper or typing on a computer; one girl has long brown hair, one girl has medium length black hair; ambient lighting, realistic lighting."

Hello all. Rina and I share our picks for Spooky Season. These are not in rank order, ya'll.

Millie’s List The Stuff (1985) written and directed by Larry Cohen This movie is chock full of super fun satire. Way ahead of its time - it digs into the stronghold of viral marketing; our obsession with food and our want to fill the void w/a sweet treat, corporations selling their souls to make a buck. Well done Mr. Cohen. Also, if you have Michael Moriarity in your movie it’s gonna be a banger. 1 million corns up. Elvira Mistress of the Dark (1988) Directed by James Signorelli Written by Cassandra Peterson (Elvira, herself), John Paragon, and Sam Egan This movie is right up there with Beetlejuice for me but doesn’t get the same love. It has a lot of the same funny scary energy. Some really great jump scares, gooey monsters, and family drama. Not to mention Elvira is iconic - strong, beautiful and hilarious with some classic ditz. Gotta love her. Also 1 million corns up. Buffy the Vampire Slayer (1992) Directed by Fran Rubel Kuzui Written by Joss Whedon This movie was life changing. Not only was it all the spooky things I love but I was introduced to one of the greatest female superheroes ever. I love the show but this movie has a special place in my heart. I love the early 90’s campiness of it. The love and reverence paid to the valley girl archetype and there’s some really classic Paul Ruebens post-Pee Wee that’s just the best. Rautger Hauer is both menacing and hilarious. All the corns. Babadook (2014); written and directed by Jennifer Kent Ok, like legit I haven’t been this scared since The Ring. I couldn’t sleep for days/maybe weeks after this movie. If you’re afraid of the dark - which I am - this is going to hit you like nothing else. Also, it’s also a “real” movie w/proper themes. It’s a wonderful play on the parent child relationship, especially parents w/neurodiverse kiddos. Amazing. All the corns, as well. The Color out of Space (2019); story by HG Wells directed by Richard Stanley You’ve already heard me talk about my love of this movie. It’s just great. Really gross. Really irreverent. Very scary. It has a lot to say and I’m here for it. You can never go wrong with a Nicholas Cage film. He sells the whole thing. All the corns facing up.

Rina’s List I was initially thinking of the quintessential fall horror movies that remind you of Halloween and the harvest and things like that. For that vibe, my list looked like this: Halloween, Sleepy Hollow, Pumpkinhead, Children of the Corn, Wicker Man. The Pale Blue Eye (2022) would be an honorable mention here (but ultimately a letdown). But I didn’t really feel like my heart was in that list. Although I fucking love Sleepy Hollow, and I definitely watched the Pumpkinhead series too much as a kid (shoutout to it ALWAYS being on on SciFi channel), what I really like to watch when the leaves start to turn and the days start to get shorter are movies that have a classic spooky vibe. What I mean by that is like movies that are foggy, and the character is basically alone. Maybe there’s a cool old creaky mansion (no longer in its prime/spendor), or a remote cabin. There are definitely scenes of the protagonist walking through a crunchy yard, or sparse, foggy woods.

With that in mind, here's my list of (fairly recent?) movie recs for the remainder of the Spooky Szn, in no particular order: The Witch (2015) - Robert Eggers’ full-length writer/directorial debut (sticking to our writer/director theme accidentally?) Also known as the V-Vitch. Maybe the ultimate harvest season, New England, spooky tale? It’s got it all: the unforgiving, remote wilderness; crops failing mysteriously; heavy religious overtones; creepy (and annoying) twins; virginal teen daughter; “is mommy crazy?”; goat. “Similar but different” bonus pick: Stoker (2013); also coming-of-age, teen girl story? The Awakening (2011) This movie is moody as hell! Fog; beautiful manicured grounds; cronchy piles of leaves; spooky boarding school; hollywood hunk Dominic West taking a bath; creepy lil kids; ominous warning from an old lady character; ghost hunting! “Similar but different” bonus pick: The Others (2001); also a story about a woman with a husband who went to (a different) war and wants to protect children? Don’t Be Afraid of the Dark (2010) I like this version over the 1973 original – it's a whole different vibe. This one is more, “parents not listening to the kid,” and the original is more, “is mommy crazy?” This one is set at a beautiful, old, Rhode Island mansion. Spooky groundskeeper; precocious little girl; new stepmom vying for approval; preoccupied dad; things going missing; things going bump in the night. “Similar but different” bonus pick: Silent House (2011); also a story where a woman is fixing up an old house to sell it? Deerskin (2019) Guy buys a deerskin jacket and goes off the rails. It was way too much money and he threw in a camcorder. “Similar but different” bonus pick: Creep Crimson Peak (2015) Gorgeous movie, saturated with color. The costumes, the house, the mysterious family secret. A little bit more of a late autumn/early winter movie but it fits the mood.“Similar but different” bonus picks: The Haunting (controversial hot/bad take: I prefer the 1999 over the 1963); also a movie set at a beautiful, haunted house with a dark and dangerous past?

Honorable mention: Not a Movie → The Haunting of Hill House Dishonorable mention: Sucked and was bad → The Lodge We asked ChatGPT to participate, as well. Here's what it said:

Then we were like - hit us w/a bonus.

Nailed it. Honestly, these are all just good. For any season. So, happy watching!

In a previous post, I talked about how we used Generative AI tools to create a logo we liked. All that remained to get our podcast started was a good name and spooky theme music.

For ideas, we turned to OpenAI’s ChatGPT, which has accumulated over 100 million users* in the past few months and has a free-to-use version. I also tried out WriteSonic, which gives you 10,000 free “words” to use with their ChatSonic bot. ChatSonic also claims to be powered by Google Search and can generate real-time content, where this version of ChatGPT was trained only on data up to the year 2021.

I’ll start with ChatSonic. At first, I fed it a simpler prompt just to see how it would respond:

As we mentioned on the intro episode of the pod, we didn’t think “Ghoulfriends” was that bad! But there was a problem: we didn’t prompt it to output names of podcasts that don’t already exist. So, it didn’t tell us that there were already a number of podcasts with some variation of “Ghoulfriends” in their names. It also made a bunch of assumptions, for example - that we’re looking to come off as “lighthearted and fun.” Did it associate these descriptions with “women” or “millennials”? There might be data from its training that suggests that these groups want to be perceived this way. Or did it assume that these traits would be beneficial for any podcast, like “appealing to a wide audience”? This seemed to ignore part of the prompt that we did give it, that this was a podcast about horror movies, which implies a niche audience.

As for “Terror & Tequila” I have no idea where it was going. First of all, it sounds like a podcast about cocktail pairings for scary movies. Also, “…represent the opposing viewpoints of the hosts” suggests that if you hate horror movies, you must love… fun? Is comedy the opposite of horror? It’s clear that ChatSonic has never seen a horror movie. The last part is bizarro too: “the name suggests a party atmosphere.” Like, why?? Millennial women looooooove to party and do tequila shots. I gave it some generic feedback:

Not great output, and the explanations were shorter as my prompts got shorter. Also, sarcastic apology? It was all like “soooooo sorry my answers weren't cool enough for you” and called me a bitch, pretty much. Whatever! And “Nightmare Avenue” basically only considered the “horror” aspect and ignored all of the other details of each prompt. Lazy! I pointed this out:

Less passive-aggressive apology, at least! Alot of millennial pop-culture “girly” tropes here, like calling each other “queen” and being in a “girl gang.” I’m least offended by “Horror Honeys” because maybe it thinks we’re sweet. “Blood Babes” sounds like it's probably some kind of sex kink or fetish. It sure loves alliteration too. We really fucked up by not going with “Blood Babe Final Gore Girl Gang of NYC,” honestly.

Ok, so, with ChatGPT I input longer, more detailed prompts for the most part:

Ah yes, that’s what “Sex and the City” was missing! I’m glad it concluded that “horror movies are all about screams” - now we can stop our research and write a script where everyone screams at each other and then goes out for cosmos. As the kids say, it’s giving “humorous and sarcastic tone”. Maybe it needs more info? Let’s throw chaos into the mix and see what it thinks “sassy” means:

There you have it - they both came up with the same name. ChatSonic arrived at it from “cool or edgy”, ChatGPT from “sassy and sarcastic,” but they both got to "Scream Queens of NYC." I thought it was interesting that ChatGPT hallucinated that we went to high school in Queens - it responded so matter-of-factly! “Horror High: Bronx Tales” has a “catchy ring to it”?? Maybe we can use it for when we tell our inevitable “funny or ironic stories” on the pod. 

I tried a few iterations where I gave it specific words to work with. For example, include "slice" in the name. But it was becoming increasingly evident that the pun aspect was the issue. Finally, I gave it this prompt for kicks:

Umm, yes - I will now refer to men in NYC as “Gotham Ghouls”. “Midnight Massacre” might be a good name for, like, a metal band - adding that one to my Notes app. It's interesting how there's a gender association with that phrase, though, it seems neutral to me. Probably because men be massacre-ing IRL, I guess. “Horror in the City” sounds like a dating podcast, amirite.

Anyway, overall, I preferred using ChatGPT. It provided more relevant answers after refining prompts, better ”understood” descriptive words, and the free version seems pretty robust. ChatSonic seemed to have limited access to data (or different types of data), gave lazy answers at times, and isn't free to use for further experimentation.

Although we didn’t go with any of their suggestions, both were good brainstorming and ideation tools. From this exercise, we realized that most horror puns were mad corny. We also realized that the complexities of our POV and the overall vibe we were aiming for were challenging to articulate, and certainly too tough for either program to predict since it doesn’t already exist in their training data. This inspired us to get down to the bare bones (ha!) of what we’re trying to achieve with our podcast: figure out if something is mad scary or not. And voila - we had our name.

Also, again, a takeaway here is to be as specific as possible with prompting, or try giving it examples to get closer to what you’re looking for. Or, go completely opposite and try a prompt out-of-left-field to see if the results are interesting. I found it great for spitballing and getting our creative juices flowing. Just remember that it can’t generate anything entirely “new”, has no taste, and hasn’t gone through all the necessary childhood and lifetime trauma to be actually funny. 

Welcome to the Mad Scary Podcast! Hi. Hello.

In this introductory episode, we forget to share the podcast name, discuss what drew us to the horror genre, explain our relationship to writing (& each other), and ask ChatGPT to recommend horror movies to watch and discuss in future episodes.

We make a pledge that our podcasts will not be over an hour (let's see how we do). Millie (me) rambles on about why we chose the genre - spooky stuff being the sweet spot for 80s/90s kids. I name-drop Buffy the Vampire Slayer (both the movie and show) as well as the influence that movies like Scream (which Rina incorrectly attributed to Matthew Williamson instead of Kevin Williamson, our bad!) and The Ring had on my developing brain. Rina waxes nostalgic about writing 30-page papers and working production as part of the theater club (nerd alert) at our all-girls Catholic High School. We chat about our early creative collaborations (drop a Euphoria spoiler?) and talk about our general goals for this hella meta pod.

We delve in on our healthy curiosity of Generative AI from the POV of two millennial women, ex-teachers, women in tech, and brunettes with bangs that love Ratatouille. Also, is Ratatouille a horror movie? Maybe.

We make our case for Generative AI in the Arts which is: it's not going back in the box, so give it a test run and try to find the benefits of this new tool that is very much in its infancy. It can't be "creative" without us, and even when we prompt it to be so, it's pretty meh. Check out Rina's post for more on how she used text-to-image AI to help us create our logo. With that being said, there is some major potential for it to support creative thinking and endeavors. Also, let it help you negotiate a lower rent. Shout out Sean!

Finally, we ask ChatGPT in real-time to help us with a short list of movies to watch for our horror movie research. Check out our prompt and the OG list below:

We got a pretty solid list - most of which are contemporary picks. We'll definitely need to make some adjustments. Also, umm, hello ChatGPT, Let the Right One In IS a vampire movie. Fix it!

After a fairly insincere apology, ChatGPT redeems itself with a pretty good pick. We take this as kismet and decide this is the first movie we'll be reviewing for our next episode. "Under the Shadow" (2016), directed by Babak Anvari, is currently streaming on Netflix. Check out the trailer... if you dare!

When we decided to start this podcast back in April, Millie and I didn’t really know much about the process. Millie had been a guest on a podcast(link) in the past, and I have a couple of friends who run their own - that’s about it. What we did know was, a) that we wanted to spend a year researching and writing a horror screenplay, and b) we wanted to experiment with generative AI tools like ChatGPT to help us along the way. If you’re not sure what generative AI is, here’s a primer to get the gist of it.

We typed “how to start a podcast” into Google and hoped for the best. Aside from a device to record with, most results said that you just need 3 things to be able to publish your first episode: a name that represents you, an image or logo, and theme music. Our tired, screen-addled brains were grateful that we asked the AI overlords to help get us started.

So, how did these freaky internet bots help us create our Logo? Many AI text-to-image generators are free, or have some free functionality, while others (like OpenAI’s Dall-E) require you to purchase credits. I went with Picsart and DeepAI’s free tools, and Canva’s version that’s available with my existing paid subscription. Full disclosure, this was my first time using any of these programs so my first attempt at writing prompts for them was pretty weak.

Some of my starting prompts were super basic, like these:

Inside of a New York City subway car

Spooky New York City skyline

Full moon over the Brooklyn Bridge

To which I selected one of the available style choices (“Playful, “Filmic,” “Ink Printing”), yielding results like this:

These weren’t really what I was looking for, so I started to refine them by adding info about the style, amount of detail, type of light, positioning, etc. For example:

“Full moon at night over the Brooklyn Bridge, centered, in the style of a children’s book illustration, spooky” (below, left) or “Inside of an empty dirty New York City subway car at night, realistic lighting, medium detail” (below, right):

Most of these were warped or alien in some way. Like, any text you specified to be included came out as gibberish language, and anything to do with human - or, in our case, robot - forms were jacked up. Here are some of the robot rejects:

The prompt for both was: “one metal robot hand reaching up out of the dirt in a cemetery at night during a full moon, centered, high detail, realistic lighting.” As you can see, there’s a lot of weird stuff going on here. Extra fingers, half finger, half hand, hand off-center, moon off-center, tiny moon, weird double-crosses on graves - and that’s with some pretty specific prompting!

Canva’s tool was easiest to use because I could refine the prompt quickly after each batch was generated, and see the change until I got something that I liked that was usable. I think the prompt for these two was something like, “Inside of a spooky New York City subway train car in the style of a retro anime” and “New York City subway train car passing by in the style of a comic book”:

I messed with the colors, saturation, filters, and some lil ghost buddies. Then layered the text, added various effects to the fonts, etc until they looked like this: 

There are still some things I’d want to change about it, and it’s obviously derivative of the base “Retro Anime” style it was pulling from. But the point was to try and do some things that were out of our element with the help of generative AI and I think this was a good example of collaboration. Here’s one of the robot ones that also came out alright:

Honestly, this took a LOT of time to prompt and re-prompt, select, and edit. Likely way more time and effort than it would have taken a professional Graphic Designer to ideate concepts, get approval, and do edits for something like this. 

Some takeaways:

Text-to-AI generators are fun and easy to use, but yield better results with detailed prompts

The generated images are, by design, derivative of other human artist creations in some way

A great way to get inspiration, or develop a raw concept

You can make the output yours with lots of editing and a human touch

Hi! 

Welcome to Mad Scary Podcast, a year-long podcast experiment and research project documenting our first-time horror movie screenwriting process. On the pod, we dig into movies that explore scary themes and tropes that we love as inspo for an eventual script. To make matters worse, we're throwing in a heathy curiosity for and discussion of the current moment of Generative AI, from the POV of two millennial women (Rina & Millie), ex-teachers, and best friends from the Bronx.

Our intro episode is out -- learn a bit about us, what we're working on, why we're doing this, and how we intend to use Generative AI systems (like ChatGPT) to help us along the way.

We'll be using this blog as a companion to the podcast to document anything from episode recaps, to references, to digging in on a specific topic or horror trope.