Jumbled in Letters with Cassidy Williams

Hello! Can you tell us who you are and what you do?

Heya! I’m Cassidy Williams and I’m the creator of Jumblie—a word-guessing game with multiple four exact words hidden within a long jumble of letters! Right now, I’m currently working on multiple projects with different groups and in varying roles, but all of my work has always been involved with development and software engineering.

I co-founded the companies Cosynd and Contenda, and aside from Jumblie, I’ve also created the game Thirteen Potions beforehand. Aside from executive and leadership positions, I do a lot of advisory roles, as well as co-host the Stack Overflow Podcast!

Jumblie is a word game where you have to guess exactly four words that are hidden among the jumble of letters. These four words are always between four to nine characters long, with a color-coded system where blue is the longest word and red is the shortest. So if you’re looking for a new word game to add to your routine, I’m sure Jumblie can give you the challenge and fun you’re looking for!

What’s your favorite game to play?

Civilization VI! I've sunk so many hours into that game. It is so dang fun. And plus, because it's turn-based, I can play a turn and then, like, answer an email and then go back and I feel very productive while playing it. 

And then board games. I really like Agricola, and that one is kind of like a land management game where you're trying to build out your farm, feed your family and beat other players to certain resources and stuff. It's really fun.

Some people would think it’s similar to Settlers of Catan, but I think Agricola is a bit more strategic because you can have different improvements and occupations in every game. I've never played a game that's similar to another one because it is really randomized in a good way and balanced in a good way. And I feel like Catan is a good gateway drug to Agricola.

Can you tell us a bit about the early points of your life?

I interned at General Mills a long time ago, and I lived right next to headquarters. That was the best summer! I always gush that if I ever did move to Minneapolis, I would work for General Mills again in a heartbeat, purely because I had so much free food. I've been spoiled for cereal prices for the rest of my life. And on the way from, like, the parking garage into the rest of the building, there was a free room where they always just had snacks, whether it was like Nature Valley bars or Muddy Buddies or something. And I was at my largest at the end of that summer, because it was so much food. They had a grocery store, a doctor's office, a daycare… all of it was in-house. It was amazing.

It was so good. Like, the gas station on site, I saved so much money on gas. It has so many nice perks that I feel like people don't realize how nice that office is. And, yeah, I have some friends who still work there from back when I worked there, and I don't blame them. It's amazing.

Heading to Iowa

I went to Iowa State University. Go Cyclones! I was the school mascot there, so that's how much I liked it. I was wearing a big bird suit and I started to realize, “Oh, dang, did I have a furry spell?” I didn't mean to. I just thought it'd be fun to put on the costume. And this was truly my freshman year of college. I was just like, “Oh, cool. It'd be fun to wear the bird once. I wonder if they ever need help. Like, I don't know, waving in a bird costume?” And then they're like, wait, how tall are you? Our side just graduated, and thus I was the mascot for all four years of college.

So they have different costumes where, like, there’s this athletic person that actually goes to the games. I was not athletic. You have to be able to do a flip in the costume for that one. But I did all of the non-sport events!  And so I kicked off races, Winterfest, etcetera. And the biggest one was probably just the Iowa State Fair, where I was amongst the butter cow and, like, waving to all the kids and taking pictures and stuff. I passed out in the costume because it was so dang hot. It was quite the journey being the college mascot.

A knack for comedy and entertainment

I think I've just always been a clown, whether I realized it or not. I have always liked entertaining people, but it was more just like having a good time and stumbling upon various things that have led me to where I am today. I’ve made lots of funny clips and videos, especially regarding coding and programming and stuff, and I even have a lot of wigs to put to good use in many of them.

I admit I slowed down making them when I gave birth because it's kind of hard to take the time to make and edit a video when you have a baby rolling around, but I definitely made a lot.

Making videos

I actually first started making videos way, way long ago, like early 2000s, and back then, video editing was just so cumbersome at the time and took so long. As high school, college and life, in general, picked up, I kind of just put it aside and I was like, “You know, videos are fun, but video editing takes forever.”

And then TikTok rolled around and then back in, like, 2019, I kind of just started playing around with it, and I was like, “Wait, does this mean I can make videos again?” And it's so easy to just edit and hit submit then it's done! And so then I started making a ton, and then the pandemic happened and I had nothing else to do. And so I was making a ton more videos, and I still like making them on occasion, but, yeah, I have more limited time these days to do so.

Living in Spain

Spain was… that was an amazing rewiring, honestly. I've actually given a conference talk about how much it rewired my brain because it's such a great country. I love it. And it was really good for really getting immersed in Spanish and speaking better and everything. But before Spain, I was so stressed just trying to do everything and wasn't sleeping well. I literally almost went blind in one eye because I was sleeping so poorly and just trying to do everything.

And then I moved there, and this beautiful land of olive oil and siestas, I love it. And I got to make so many good friends. This was back in 2012. Oh, my gosh, it was 12 years ago. So many friends made there… I'm still friends with today! I just got to know them super well, got to know the culture super well and really just traveled the country and loved it. 

My husband and I still talk about what if we moved to Spain one day. And, yeah, it kind of rewired my brain to just be more chill, honestly. I remember when I came back, my mom was like, “This is the most zen I've seen you since you were five. Well, what's up?” And was a really great experience living there. I loved it.

It's so beautiful and everyone is just so kind and nice. I actually visited this past fall and we brought the baby. She was only six months old, and oh, my gosh, they treat us like royalty when you have a baby there. People were making room for us to sit at their own tables at restaurants and giving her bread. And we were able to just really just walk through the country in a place where they really care about kids.

Not that the US doesn't, but definitely, relatively speaking, it feels like they don't. It was a harsh reality coming back home because they were just so kind to us and let us go to the front of every line and just park everywhere and everything. And even though she was very small at the time, like, “Oh, I can't wait to bring her back!” Just because it is such a nice country with really great food.

SEO of her name

There is a Scooby Doo character named Cassidy Williams who I did not know until far too late in my life. Who is she and why did I keep this name? She's ruining it. That is a joke! A joke! But anyway, yeah, I guess SEO is not something I'm an expert at in any way, shape or form. I just know generally, “Hey, you should have certain keywords and have the right og images and descriptions and stuff.” 

And so I've always tried to do that on most things. And then I was very annoyed one day when I saw that I was still being overtaken by Scooby Doo. So I wrote a blog post where it was just my name and a bunch of keywords. So it could be like, you know, this will at least help bump up my website a little bit more. So that way if someone's looking for me, they can find me.

Can you tell us more about Cosynd?

Cosynd is one that I almost wouldn't even call the most recent, more like the longest. I started that with a really awesome group of people back in, I think it was 2015-2016-ish. It started kind of out of the blue with a very funny coincidence, when a mentor of mine founded a firm that no longer exists called Women In Mobile, where it was investing in women, building mobile applications and mobile startups, or Women Innovate Mobile. That's what it was.

But then there was another very large one called Women In Music, and the WIM domain name was a major conflict there. And so my mentor and this person who was on the board of Women In Music, they were just like, “Hey, we want the domain name.” And then they realized just how big both of their organizations were and became friends. 

And so my mentor was like, “You should talk to this person because this was kind of an accident, but she's really cool.” And then as I was talking with her, she was talking about how so much of the music industry is still done via paper, like writing split sheet contracts, where if, like, all three of us wrote a song together, we could say that, okay, each of us owns 20%, and then the publisher owns the rest or something like that. Stuff like that is still done via paper a lot to this day. Or saying, like, okay, I want this music in my movie. You have to get so many things signed off via really slow processes and legal processes that take forever, and also just copyright registrations in general are still very slow. And if you make a mistake on your application, you have to pay the fee again and start over.

So anyway, it was just a very, very manual process. And so as she was talking to me about what about this as a problem, I was just like, well, what if we built this? What if we did something like that? And she was just like, “This is what I want to do. But I don't know anybody technical.” And thus Cosynd was born. 

There's a group of us now where we made this product that now kind of runs on its own, and we're building some more features that I can't talk about currently, but it's very exciting where new copyright rules come out. We can build it into the platform and then publications can use it for copyright. 

And it's been something where it's still kind of a side project for all of us, where some folks work on it full time. Most of us are on it part-time as we do other things. But that's kind of the story of Cosynd and it's been really, really awesome to work on.

How about Contenda, what’s the story there?

Oh, man. It's still very bittersweet for me because we're open-sourcing and shutting that one all down, and the team is splitting up and my CEO is trying to see what things she can salvage from it right now, and so it'd be cool if she can build something. But anyway, Contenda was one that kind of came out of the blue as well.

For Contenda, I have a Patreon group where I originally made it so people could get resume reviews and interview prep and stuff just to kind of help scale and protect my time a little bit better. And it turned into just a really great Discord community where now to this day, I think I've missed like 50 messages since this podcast started. It's full of really active people just talking about tech in general, what they're building, games that they're playing, stuff like that. 

My CEO at the time was in there and she noticed that I was mailing stickers individually to people just as like a thank you for being a part of it and stuff like that. This was back in 2020 or so I think it was right around pandemic times. And she was like, “Hey, if I built something that would help you automate sending stickers, would you use it?” And I said, absolutely, because I was like handwriting envelopes and stuff, and I didn't know of any services at the time.

Then a couple of friends ended up building this platform for mailing stickers. And then they were like, “Wait, this is kind of cool, actually.” As they scaled it and messed with it, it turned into a really good platform for retaining subscribers. And they ended up helping Ludwig, the Twitch streamer, break a world record for the most number of subscribers on Twitch because he was mailing stickers to everyone with the platform. And that's kind of how Contenda started.

And from there they were like, “Okay, stickers don't scale. What if we built different creator tools and pivoted a bunch of different times?” Eventually, I joined and we were building AI things before the AI hype wave started. So like before ChatGPT and stuff. And we were building tools where you could take media like this podcast, for example, and then turn it into a blog post or a Twitter thread or highlight clips or things like that.

And so we had built this tool and then ChatGPT came out and we kind of tried to refine it, saying, “Oh, but we can help with technical accuracy with these custom models. So you could say ‘install react in a conference talk’, and it would actually have the code samples for installing that.” But anyway, that was going well enough and then the tides started to turn economy-wise.

Unfortunately, all of our paying customers got laid off. And that just happens where a lot of our customers were dev advocacy teams or groups that were specifically focusing on technical content creation. And not only were layoffs happening there, but then also ChatGPT4 came out and our edge kind of started to go away as everybody was kind of making a “turn a video or a YouTube into a transcript and a blog post and summarize it”. 

And so we started the pivoting process again. Our latest software that I particularly loved was Brainstory, where we had a few mini pivots in there, but then we started leaning into Brainstory; where it took a conversation that you were having with the browser and turned it into a really good summary where you could say, “Okay, I want to talk about podcast idea that I have.” And what it would do is instead of generating like, “Okay, this is the podcast that you should do:” it would just ask you questions. 

It used Socratic questioning where you could say, “Okay, I have a podcast idea, and it would say, “Okay, who's your target audience?” And as you talked it through, it'd help you brainstorm ideas. And so I ended up using it for conference talks, meeting prep, blog posts and more. And that was kind of the main use case where a lot of people were using it to write essays or to write content, or to, again, plan for meetings, and you could get feedback through it and stuff. And it was going just well enough that we really tried to hold on and make it work, but we realized that we had a really nice shiny nail or really nice shiny hammer, but couldn't figure out which nail we wanted to hit. And none of the nails that we tried to hit were making enough money to sustain it. 

And so that's kind of what led to us making a very sad decision to open-source it and shut it down. But at the same time, I'm really happy that we are open-sourcing it so it can live on in some way. And I'd love to kind of recreate my own version of it because it's a tool that I've really enjoyed working with.

What was Thirteen Potions?

Thirteen Potions was a game that I kind of made for a hackathon. And I thought, you know, I've always wanted to make a little game like this where it was a little character running around with arrow keys and stuff. And I really enjoy arcade-style games. And there's this online hackathon that happens every year called js13kGames, and it's all about making a 13-kilobyte game with JavaScript and going wild. That's pretty much the prompt.

Also, if you are interested in this game jam, I think this year is its 13th year and it's going to be in September. Well, that’s when they, I think, release or announce the newest challenge. But anyway, so I saw that the hackathon was starting and I was like, “You know, I've always followed it because it's been interesting, so I'll try submitting something.” Thus Thirteen Potions was born.

I was originally thinking, “Okay, I found this free tile pack on the internet. What if I made a little maze?” And as I was kind of fiddling around with Tiled, the map editor, I was like, “Okay, you know what? Instead of a maze, what if I tried to actually make it where there's a goal, instead of just going from point A to point B?” And so that's kind of how I came up with the concept of Thirteen Potions, where it started with a drawing, and then I was like, “Okay, let's make it a proper goal.” 

And it was a very fun learning experience where I made it so that way you could have, like, randomized ghosts that slow you down. I learned about how you could do speed changes based on collisions or color changes, or a timer that's always constantly going as you're running that sort of thing. And it was really, really fun to build. 

It was kind of stressful because I had to do it under the deadline. And then sadly, it was disqualified because I didn't realize that the game engine that you use counted towards the size. And so I was like, “Oh, yeah, I'm in the size limit. I'm perfect.” Yay. And then I realized that Phaser is a very, very large game dev framework. And so, the game was disqualified, but it was very fun to build, and so it's worth it. And you can play it on itch.io.

On game creation

You know, I've always wanted to make more games, and I just haven't really taken the time to wherever a notebook of game ideas. I love making games, and I think games are just fun in general. But also game studios are very scary to me because you always hear about just how intense it is at crunch time and all that. And so I've always wanted to just make fun, little casual games where you could play a little bit and then move on with your life.

And I have a whole mega list of them. And I think, especially, when my daughter was born, I was just like, you know, I want to make some of these games in those little in-between times when she's napping, when I have these brief moments that I have to myself, I want to fill it with things that I really enjoy and things that I want to build. And so that's kind of just how it started, where I've wanted to make games but hadn't really done it. And then I started, and now I've made a handful of little silly ones where Jumblie is probably the “biggest one” because there are so many regular players of it. But I have always wanted to just make something where it is fun, kind of low effort, might teach you something and is a good time.

Can you tell us how Jumblie came to be?

Jumblie, believe it or not, was also initially made for a competition, and then it was fully not related to the competition. I made it separately after that, where a friend of mine runs a little kind of a game show. Think great British Bake Off. But for developers, his name's Jason Lengstorf, and you can find his content, and it's really fun. But anyway, he had this challenge where it said, “Make a leaderboard. Make something that has a leaderboard of some kind.” And so I actually pulled out Brainstory, the tool I was talking about earlier, and I started talking through it, saying, “Okay, I need to build something with a leaderboard. I'd like it to be some kind of fun game, but I don't know what.” And as I started talking out ideas, I ended up fully neglecting the leaderboard part of it and just wanting to make a fun game. And that's kind of how Jumblie was born.

I like word games, like Connections and the Wordles and Crosswords and stuff. And so I wanted to make something like that, some kind of daily word game. And I ended up coming up with Jumblie where I thought, “Okay, if you made it like a word search and you were trying to just find words within a pile of letters, that could be kind of cool.” Each word is used once or each letter is used once, kind of like Connections, but it's actual words. And unlike Connections, it makes for more creativity, in my opinion, because you can make so many different words and try to think about it.

When I initially made it, I was using my family group chat with my cousins to say, like, “Okay, can you try this?” And I didn't initially have a theme. It was originally just four random words in the English dictionary. That was not a great choice. One of my poor cousins, he was just like, “I've been at this for an hour, and I haven't found a single correct word!” And I went, “Okay, I need to make this easier.” After putting together huge word lists and scraping the internet and everything, I realized, that the English lexicon is big. And so that's when I came up with the concept of a theme. And the words could be associated with the theme, but that also makes for a harder puzzle creation process, because I can't just pick four random words. It has to be four random words that are affiliated together in some way. And so that's when I turned to AI to kind of create a word list.

And I learned a whole lot about LLMs in that process where, like, I knew it from my work, but at the same time, there were drawbacks in it, like, LLMs are not good at counting, for example. So when I said, okay, I want each word to be, like, four to seven letters long. It doesn't know the length of words. It just knows, generally, the vibe of the word that you want. And so I had to be a bit more loose on that and say, “Okay, generally give me words related to these themes.” And it was really decent for, like, the first ten suggestions, maybe the first 20. But then after that, it just started to be really, really obnoxious. Where one of them was like, “finance”, and the words were like, “macroeconomics”, “microeconomics”, “loan”. It was just doing bad words. And so I ended up using it to create a general set of, like, the first 50 or so puzzles that were okay.

I ended up having to tweak a bunch of them, and that gave me enough, like, “runway”, to be like, “Okay, I can come up with enough puzzles after that that this game can sustain itself.” And it got just popular enough! Sometimes we hit, like, a little over 1000 players a day. On low days, it's around 500 players a day, but it's just enough where I'm a puzzle editor now, and so I added the ability to suggest puzzles so people can submit some. I add most of the puzzles myself and try to do it far enough in advance that I can still enjoy the puzzle later when I play it.

And then I did ultimately make a leaderboard, but it's just a leaderboard with yourself, and you can see what your fastest times are and have some visibility settings and things like that. And so, anyway, that is generally how I made Jumblie, and it's been really fun to play and work with.

On different levels of game affinity for Jumblie players

It's interesting how different people are faster at other games or some themes more than others. I remember “architecture” was one and one of our friends who's an architect was just like, “I nailed that. This is amazing.” And then his wife was just like, “This is impossible, I don't… building?! That's it, that’s what I know. What other words are there?!” And so it's fun to see which puzzles make people tick more than others. And my mom texts me each day saying like, this was a good puzzle or no, this was not, this was not it.

On feedback

It's fun to get feedback from people who aren't in my usual tech sphere because so many times I'm building developer tools or things for fellow developers and stuff. But it's really fun now where friends of mine, just from other aspects of my life will be just like, “Oh, my grandma really liked the puzzle today.” And so being able to serve that kind of audience that I don't normally hit is really, really fun, just because I get to think about very different perspectives. And I wouldn't do one where it's like programming languages because even though that would be super easy for someone like me, someone like my own grandma would not know what to do about that.

On Jumblie being referred to as “pure, old-school JavaScript”

It is. Apparently, I like torture. I decided to use no frameworks or anything I fully like. It's like making a cake from scratch. I built it all purely just in the browser on the platform, vanilla HTML, CSS, JavaScript, no libraries or frameworks or anything, which in hindsight probably could have increased my dev time by a whole lot if I actually built it in a fancy way. But at the same time now, because of the way that it's built, I'm able to add features pretty easily because it's just saying, “Oh, that's just that one button. I know how to edit that button, and I don't have to edit a server code anywhere or database anywhere.”

And also, if you were to download Jumblie just, like, the entire web app onto your phone or a device, and then suddenly be lost in an ocean with no internet, it would still work. And so the entire game will still function for the number of puzzles that have been loaded into your browser at any given time. But you could be on a deserted island and play Jumblie just fine until you run out of battery.

“Jumblie” on Netflix?

That was really convenient timing, honestly, where there's this show on Netflix called Devil's Plan, and it's really fun. It's basically just a bunch of smart people in a room doing puzzles, and then people get kicked out if they don't do the puzzles right and that sort of thing, and lose points. And there's a big prize at the end, all that.

And one of the games, it's a Korean show, but they had all of these English letters, and they had just a pile of blocks with letters, and the game was a lot like Jumblie, where the big pie in the sky man was just like, “The theme is ‘animals’.” And they had all these blocks of letters and had to pull out all of the words that could be animals, and they had to pull out four animal words and spell them out.

So it was basically a physical version of Jumbly. And it was very cool because suddenly I was getting all of these mentions on Reddit for, like, the Devil's Plan subreddit, saying, like, “Hey, if you really liked that game, this one is an online version of it.” And so suddenly, just traffic spikes from that end of the world. All of these people in Korea started playing it, and it was very funny seeing the comments where people who English was their first language were just like, “Dang, they did this as their second language.” This is pretty hard with English as my first language anyway. But that's kind of where it got its first spike, and it's been really fun to see the traffic from there.

Next step for Jumblie

I'm trying to figure that out myself because I love the game, don't get me wrong. But being a puzzle editor is hard work. And so I'm trying to figure out, is there a way that I can automate this a bit more? Once again, I added ways to like, suggest a puzzle now, and that kind of helps. However, AI is not reliable enough yet to be able to generate puzzles. And so it is very, very human. And though I enjoy it, I'm trying to figure out what should I do in that regard. Should I, like, hire a puzzle editor? I don't make any money on Jumblie, so that would just be negative dollars unless I change something up. Should I contact publishers? I don't know. So I admit I'm kind of in a wishy-washy place where I can't even give you a clear answer because I'm trying to figure that out myself.

Did Jumblie ever bring you pleasant surprises career-wise?

With how Devil’s Plan impacted traffic, it's really fun to kind of tap into a lot of these communities and learn more. One of the games that I play often that I honestly probably should have brought up before is Go, also known as Baduk in Korea. And it's the black and white pieces on the grid. And, like, AlphaGo was a big thing a few years ago, that was the AI playing it. One of the competitors on Devil's Plan is a professional Go player that I actually follow on YouTube, and I was able to talk to her. And so the little things like that, we were just like, oh, dang, this is awesome!

If you could, what would’ve you done differently with Jumblie?

Right now, my primary problem is puzzle editing. I think it would have been interesting to add another layer of hints or something where I could maybe do, like, the same puzzle again, but it has a different kind of hint and that might be interesting. I've been thinking about, “Oh, what if I made a hint where, like, easy mode colors all of the letters in the words that they're supposed to be or something like that? Or if I added some degree of challenge where there's like a stray letter that is added. But first of all, I feel like the existing players might be just like, “I don't know how I feel about it.” I'm thinking about that, but something that would make it so I could reuse puzzles a little bit more would probably be nice for me.

On the difficulty of puzzle editing

I was actually talking to Josh Wardle recently, who made Wordle, and we were both lamenting about being puzzle editors. When he originally just got tons of words, there were a lot of five-letter words in the English dictionary. And he ended up making a whole separate app where it was just people saying, I know this word, I don't know this word, back and forth until he had a good enough dictionary of words that he could populate his game with it. And he was saying, “There are ways to automate. Word games are different compared to a lot of logic games which are very math-based and you can generate things more. And finding that balance is definitely challenging, but it's an interesting challenge because it's different for every game.

Where should we go to learn more about you and your projects?

You can find me easily. You could just google my name. But once again, I'm competing with Scooby-Doo here. So Cassidoo is my username on most things, and cassidoo.co is my website. And there you can find my newsletter, which is a weekly newsletter where I do practice interview questions, web news of the week and I talk about the various projects I work on.