Episode 13
Choice Paralysis in Unreal Engine: When Too Many AI Tools Kill Productivity
Today, we're diving into the wild world of creativity and productivity—or, as I like to call it, my chaotic journey of trying to get stuff done. I’m out here wandering through the woods, reflecting on how this podcast has turned into my accountability buddy while I grapple with the daunting challenges of using Unreal Engine. You know, it's a real love-hate relationship, especially when I find myself stuck in a loop of frustration instead of making progress. But hey, I've discovered that chatting with AI can be a game changer—think of it as having a coding partner who never sleeps and is always ready to brainstorm! So, as I navigate this blend of digital creation and real-life distractions (like family gatherings and the occasional existential crisis), join me in figuring out if I should stick with Unreal or embrace the simpler joys of, say, picking up a camera and just hitting record. A stroll through the forest sets the stage for an introspective dive into the world of game development and the trials we face in the digital realm. While wandering among the trees, I reflect on my journey, the challenges of working with Unreal Engine, and the ups and downs of creative accountability. The podcast serves as my personal diary, a space where I spill my thoughts on navigating the labyrinth of projects, bugs, and the occasional existential crisis. It's not just about the code; it’s about the journey of self-discovery wrapped up in deadlines and digital frustrations. I also explore the role of AI in my workflow, which has become a unique companion on this path, especially during those long hours stuck in front of the screen. It’s a conversation about finding motivation in unexpected places, and how sometimes, you just need a sounding board—even if it’s a chatbot. As I wander, I ponder the importance of community in this solo endeavor, realizing that while I may feel like a lone wolf in the coding wilderness, I’m actually part of a larger ecosystem of creators. The episode is a blend of personal anecdotes, technical musings, and a sprinkle of humor, all tied together by my love for creating in Unreal Engine and the hope of finding clarity in the chaos of creativity.
Takeaways:
- We’re diving deep into the chaotic world of bugs and obstacles while learning Unreal Engine, and honestly, it feels like my own tech diary on a wild adventure.
- This podcast serves as my accountability partner, keeping me motivated while I juggle projects and the reality of being a solo dev in a complex ecosystem.
- Walking through the forest while recording this episode is both a refreshing change and a reminder that stepping outside can spark new ideas in our creative processes.
- Unreal Engine is a beast that demands teamwork and resources; sometimes, it feels like I'm wrestling with a giant while trying to make my dreams come true.
- I’m realizing that using AI tools has made my workflow smoother, even if I'm still figuring out how to harness them effectively in my projects.
- It's all about finding the right balance—between using Unreal for its powerful capabilities and knowing when to step back and embrace simpler tools for quicker results.
Companies mentioned in this episode:
- EPIC
- Unreal Engine
- Chat GPT
- Claude
- Gemini
- Anti Gravity
- Marvelous Designer
- Godot
- Seat Dance
- Nano Banana
Transcript
Hey, welcome.
Speaker A:I'm walking through the forest, which is such an unusual place to record this particular podcast.
Speaker A:This podcast is all about sitting in front of computer, digging deep into projects, bugs, obstacles, challenges, new tools, older tools, and just making sense out of it.
Speaker A:It's my diary, it's my devlog, a journal of my path through learning and trying to get something done.
Speaker A:In Unreal Engine in particular, I cannot explain how much appreciation I have for this podcast because it really keeps me accountable, it keeps me motivated, liable, and really keeps me grounded around what I am doing.
Speaker A:It's not so easy to walk away because I know I not only opened a project and I'm logging into EPIC and Unreal Engine just to mess things around.
Speaker A:I need to tell you about it.
Speaker A:It's really an interesting ecosystem where I'm no longer in a vacuum.
Speaker A:You know, I'm no longer in this void of being alone in front of a computer.
Speaker A:Now, reality of work has changed a lot lately because I have those AI assistance chats in particular, so I get to talk to my AI about the project, and that helps a lot.
Speaker A:If you like what you're hearing right now, but you don't want to do a podcast out, recommend that you set up a chat, let it be Chat GPT or Claude or Gemini, and delegate it to be your coding buddy.
Speaker A:Maybe not on a code and development level, but as a partner, as an accountability partner.
Speaker A:I'm doing that.
Speaker A:Exactly.
Speaker A:But I'm also using this podcast as such device, and I've never been more motivated than now.
Speaker A:Well, anyway, right now I'm having a stroll.
Speaker A:I'm looking for a shelter from the wind on this Easter morning.
Speaker A:Yes, today is Easter and I'm visiting my parents and my sister and her family.
Speaker A:We're gonna have a big gathering today around 12 people and six animals.
Speaker A:It's a big gathering for my family standards.
Speaker A:I know there are bigger families, but also smaller families.
Speaker A:They live in a countryside, and that's where I am right now.
Speaker A:I went for the first walk this year into the forest, and you may know me as someone who just stays in her computer all the time and never goes out, but that's only partially true.
Speaker A:Honestly, I don't have enough opportunities to go into a place like this.
Speaker A:I live in Berlin and there are forests and lakes around, but I don't really have enough opportunities to go there, and I doubt myself right now.
Speaker A:I don't think it's a valid excuse.
Speaker A:I think there are ways to go there.
Speaker A:I just need to motivate myself more.
Speaker A:Well, now days are getting longer, so it's actually easier to do this to at least, you know, go and use public transport.
Speaker A:Before, days were shorter, so there was less hours in the day to actually go into the forest and come back while it's still bright.
Speaker A:Anyway, I'm appreciating this walk because I needed a break.
Speaker A:Even though I'm on my holiday break.
Speaker A:I took my laptop and a whole bunch of projects with me and I'm grateful for that because I had the most horrible trip over to Poland.
Speaker A:I don't have a car currently and I rely on either a bus or a train or a plane.
Speaker A:Although the best connection, it seems, is through a train connection.
Speaker A:I didn't take a train this time.
Speaker A:I took a bus for the first time.
Speaker A:It was not only longer than the train trip, but also there was a delay, so I spent close to eight hours on the bus.
Speaker A:Now, yeah, it's as horrible as it sounds, but luckily enough I had plenty of room and I had my laptop and I got a lot done.
Speaker A:Really a lot.
Speaker A:Now, I didn't work with Unreal Engine, but I worked around my projects.
Speaker A:I was working my website, on my portfolio.
Speaker A:As you may know, I became unemployed with the beginning of this month.
Speaker A:I had a garden leave for last month and now I'm officially on my own.
Speaker A:I'm unemployed, which is great in a way that I can focus on my projects and I can, you know, pursue my own dreams and aspirations.
Speaker A:And as reality hits deep, I cannot honestly say that my Unreal Engine will ever elevate me to to the status of being self employed.
Speaker A:Working on Unreal Engine, I just as hard as I try, I just seem like I cannot get anything actually done in Unreal.
Speaker A:I'm going deep into those projects.
Speaker A:I'm having lots of fun, but it stays in the realm of hobby of something I do for fun.
Speaker A:Two years ago, it seems I was here starting my project at a time when I've used assets from a sample city, you know, the Matrix demo from Unreal.
Speaker A:And at the time I didn't know Unreal as much as I do know it now.
Speaker A:And I just used those assets and I created a new map and I created my own map using those assets and Traffic and all that.
Speaker A:It looked impressive, man.
Speaker A:It looked really great.
Speaker A:Now I showed it to my family and they were very impressed.
Speaker A:I don't think they expressed it well enough because I got a sense that they don't really understand what I'm doing.
Speaker A:But in some ways it was impressive because they're asking me about this project.
Speaker A:Even today, my nephew, my dad, they're asking, so what's happening with this project of yours?
Speaker A:And it was a moment of reflection for me where I asked myself, yeah, what's up with that?
Speaker A:What's up with that?
Speaker A:And when I compare my Open World Metaverse project that I started so long time ago, I'm getting a little bit frustrated, you know, and I don't really know what it is like, am I still not good enough to finish it?
Speaker A:Or is Unreal truly a teamwork device environment?
Speaker A:Maybe you.
Speaker A:I mean, there are countless examples of people doing things by themselves in Unreal, but arguably they do a much smaller projects and I knew it from the start.
Speaker A:I went for way too complicated project to pursue on my own.
Speaker A:But after two years in, I don't see the end of it, you know.
Speaker A:And in the meantime, especially recently, I've.
Speaker A:I've done a lot of smaller things with smaller tools that actually can be called competed, you know.
Speaker A:So I've been vibe coding lately.
Speaker A:Yeah, I've been using Gemini Cli, Courser, Claude Cod, of course, Anti Gravity to create different apps, different tools to accomplish tasks.
Speaker A:One of my favorite is a podcast production app.
Speaker A:I basically am able to drop audio file into that tool and it generates everything.
Speaker A:It uploads my episode, it titles it, it writes descriptions, but also evaluates my episode.
Speaker A:And it's.
Speaker A:My mind is blown.
Speaker A:I have this software I custom built for myself and under the hood.
Speaker A:Most likely it might be a mess, but so far it's quite dependable.
Speaker A:I still can do things with it, I can rely on it.
Speaker A:And it's a software that I did for myself and it's completed, I can use it, you know what I mean?
Speaker A:And it took me a week to finish it.
Speaker A:Now I don't get to feel that when I work in Unreal Engine.
Speaker A:I don't know, it's just there is a huge gap between my output in other tools in other projects.
Speaker A:I've been trying to pivot to movie making in Unreal, and even that is so complex as a workflow.
Speaker A:Now I'm not quite complaining because it fascinates me.
Speaker A:You know, I really enjoy how much control we've got in Unreal, but at the same time it's so complex and so power hungry that I'm not happy with the result.
Speaker A:I'm not publishing those movies.
Speaker A:They're not so good technically, you know.
Speaker A:And I've been ranting before how, you know, Unreal gives you all of this stuff for free, but in order to get something done.
Speaker A:You kind of need to invest both time and money.
Speaker A:The biggest caveat I spoke before was metahuman outfits.
Speaker A:I'm not doing any outfits, even though I bought the license for Marvelous Designer because that thing on its own is also a rabbit hole that I don't want to get into too deep because it takes me away from the big picture.
Speaker A:You know, does it make sense?
Speaker A:Does it resonate with you?
Speaker A:Because it feels like not quite a unique frustration.
Speaker A:So, yeah, I give an example, specific example.
Speaker A:All right.
Speaker A:I try to make a video using audio to face workflow.
Speaker A:So I skipped the step where I do a mocap and I just recorded audio just like this one.
Speaker A:And I've used this workflow where Unreal Engine does lip sync of metahuman.
Speaker A:And it, you know, makes the face move accordingly to the audio.
Speaker A:So I made a video like that.
Speaker A:The rendering, the processing, even on my beefy PC, it took so long and result was so mediocre, I really lost the interest.
Speaker A:I was really disappointed.
Speaker A:And I realized at that time, well, why don't I just take a camera and record myself if that's what I want to achieve?
Speaker A:In the end, you know, what I want to do is to make a movie.
Speaker A:And I want to communicate the message.
Speaker A:That's my goal.
Speaker A:And I want to use Unreal Engine because in some ways, I feel like this will give me augmentation of reality.
Speaker A:I need.
Speaker A:I can do like a vlog in Unreal Engine.
Speaker A:And I don't need to, you know, do all the effort of going out or of looking for a good backdrop and weather and light.
Speaker A:Meanwhile, that would be still easier than doing this in Unreal Engine.
Speaker A:Now I'm doubting myself as I say it, because maybe there is a better way of doing it.
Speaker A:I don't know yet.
Speaker A:But from what I've experienced, anything that I do in Unreal Engine is just heavy, power hungry, resource hungry.
Speaker A:And there is no way around it.
Speaker A:Now, things work as a little test.
Speaker A:I made whole bunch of little clips that look great, you know, but it's just metahumans standing still, doing some simple animation, looped.
Speaker A:But when I'm thinking about creating whole composition of shots with different animations and different mocaps and different metahumans in a scene with dynamic backdrop.
Speaker A:You know, this gets to the point where my project is huge and rendering takes not forever, but every step of the way is frustrating.
Speaker A:Now, I know it's possible to achieve something, but you need to scale down your expectations.
Speaker A:You need to embrace what works and go with that, but that was not my goal.
Speaker A:I wanted to use Unreal Engine to accomplish what I want.
Speaker A:It's not like my goal was to use Unreal Engine and accomplish something, but maybe that's what I should do.
Speaker A:You know what I mean?
Speaker A:My goal is to make content.
Speaker A:My goal is not to use Unreal Engine at all cost.
Speaker A:Now, for the last seven days, I was hyper, hyper productive with AI tools.
Speaker A:And I thought to myself, okay, if Unreal Engine is so cumbersome, let's see what other tools we've got.
Speaker A:And obviously video generation comes to mind because, well, it's there.
Speaker A:It's becoming very popular.
Speaker A:There are those demos that look impressive.
Speaker A:Seat Dance 2 is blowing mind of many people and it's making movie industry really confused and angry and frustrated.
Speaker A:I looked into that stuff and I quickly realized that using those tools is very expensive.
Speaker A:Seat Dance is released semi officially and I budgeted creation of a short movie for about 4 to €500, hoping that the output will be coherent, that visuals will be consistent, and the whole thing will look nice.
Speaker A:Now I've seen some impressive creations using Seat Dance, but what frustrates me is that AI when it creates a footage, it always has this flavor of paranoia.
Speaker A:I don't know how to better describe it, but basically when I'm watching even highly sophisticated video like a short, not a clip, but a short, well thought out video, I'm witnessing a lot of craziness.
Speaker A:You know, shots might be nicely choreographed, but after a while I'm realizing that AI video model doesn't understand gravity, dynamics, and especially it doesn't understand the time it takes for something to happen.
Speaker A:It feels like every motion takes equal amount of time whether something is moving or falling.
Speaker A:It seems to be all at the same rate.
Speaker A:And therefore it's.
Speaker A:It's a very subtle, minor thing that makes the whole video flow incorrectly.
Speaker A:And then quite, always, quite often, there is this sense that there's.
Speaker A:Yeah, there's some kind of weird sense of paranoia.
Speaker A:And there are of course very little mistakes that you could just cannot avoid.
Speaker A:You know, little things where direction of objects change randomly, there are some anatomical discrepancies.
Speaker A:So you need to live with that.
Speaker A:Even though you've paid a lot of money to get the video done, you cannot avoid those little mistakes.
Speaker A:And yeah, in a year from now, this statement will be completely irrelevant because of course all of this changes.
Speaker A:But maybe not.
Speaker A:Maybe this will be always like that.
Speaker A:Who knows, maybe it'll take 10 years.
Speaker A:And what do we do until then?
Speaker A:That's the question.
Speaker A:What do we do until AI gets cheap and dependable.
Speaker A:So after spending a week downloading local AI generated generation models.
Speaker A:Yeah, I've downloaded a lot of models to render video and results were actually quite nice.
Speaker A:But then I discovered a limitation where clips I was able to generate were quite short and I couldn't get them any longer.
Speaker A:And that created a new obstacle.
Speaker A:You know, the length of video.
Speaker A:It just seemed to be impossible to render forever.
Speaker A:Yeah, so just a lot of limitations wherever I look.
Speaker A:And it seems like the only unlimited tool, truly unlimited tool, is a camera.
Speaker A: t I want to make is videos in: Speaker A:Not, not some kind of augmentated augmented workflow that involves AI or 3D graphic tool.
Speaker A:I mean, yes, but I need to assess what takes more time and effort and what's actually doable for one person.
Speaker A:So I'm thinking instead of spending €400 for and buy AI credits or instead of spending 400 hours in Unreal Engine and money also like I need to buy credits to do mocap as well.
Speaker A:Right.
Speaker A:Like all of this cost money.
Speaker A:I'm thinking instead of spending money on that, I'll just get a new camera that I will use to capture stuff with.
Speaker A:That might be my next stunt.
Speaker A:Making videos with camera.
Speaker A:Who would have thought?
Speaker A:Yes, well, don't take it as turning my back on Unreal Engine, but I'm seriously slick.
Speaker A:I'm seriously considering the end game, but I think answer is kind of clear.
Speaker A:There is no end game.
Speaker A:All of this stuff is changing and it's pointless to be offended.
Speaker A:You know, it's not like I should say Unreal is rubbish, Unreal is free.
Speaker A:But you need to have a team and beefy, beefy beefy hardware to actually get something done.
Speaker A:Because this might be, this may change very soon.
Speaker A:I've been seeing that there are some public releases of new features on GitHub, where apparently Epic is putting a lot of effort to push some AI features into the next release of Unreal.
Speaker A:That's really exciting.
Speaker A:I'm talking AI assistant, AI writing blueprints and scripts within Unreal.
Speaker A:If that's the case, I may reconsider working on my open world game.
Speaker A:Why not?
Speaker A:I just want to get things done, you know, And I'm thinking I really need to become a tool agnostic.
Speaker A:I need to be free and move freely between those tools.
Speaker A:What I will tell you now will seriously surprise you considering you're tuning in here for Unreal Engine this morning.
Speaker A:I Installed Godot.
Speaker A:Godot is open source game engine that weighs less than 50 megabytes.
Speaker A:And my intent is to integrate it with AI coding tool and I want to create an app with that.
Speaker A:What may surprise you even more is that a few days ago I completed my interactive game demo, Vibe coding it and I'm really proud of that tool.
Speaker A:It's completely standalone, it uses local AI to power an NPC shop clerk that you can talk to.
Speaker A:And it runs, yeah, like I said, locally, it's free and it has voice component.
Speaker A:You can talk to it, you can type with it, it has memory and it can tell stories.
Speaker A:It's different than your regular AI chatbot.
Speaker A:It's actually a virtual personality you can interact with.
Speaker A:So I Vibe coded this and it, it just works now.
Speaker A:I wanted ui, so I created this very simple pixel art top down game interface.
Speaker A:And it was kind of cute, you know, it was nice.
Speaker A:But then I thought to myself, what if I do it isometric?
Speaker A:What if I do like a real walkable space?
Speaker A:And I, you know what, I did that, it was a matter of one prompt and I did it.
Speaker A:But then I realized Now I need 3D models and AI is not good with that.
Speaker A:You know what I did?
Speaker A:I generated isometric game visual with Nano Banana and I tried to convert that into a mesh.
Speaker A:It just failed.
Speaker A:It just didn't get it.
Speaker A:But what I did instead, I said to Claude, hey, this is going to be the game.
Speaker A:Let's pretend it's 3D and let's just block those areas where I don't want my player to walk over, you know.
Speaker A:So basically when my player gets close to the counter, don't let it go over that picture, make it stop before the counter, make the shop clerk sit behind the counter.
Speaker A:And you know, basically it gives you impression of interactive 3D space, but it's actually flat image and there are, and there are players, there are, you know, NPCs and stuff limited with range of movement on top of that image.
Speaker A:So you get an impression that these are 3D objects.
Speaker A:And I know this, this is already, it's not a new thing in game design.
Speaker A:But I just did it and you know, I just, I skipped so many steps and I get results that I can live with.
Speaker A:I can truly live with this result.
Speaker A:It's very simple.
Speaker A:It has some you imperfections, but I will optimize that later.
Speaker A:I'll probably regenerate the whole scene again and do it with more attention to detail.
Speaker A:But my point is I got it done in like in just A couple of days.
Speaker A:And I just cannot done.
Speaker A:I cannot think, I cannot get things done in Unreal, but I can get done everywhere else.
Speaker A:So my question is, should I, Should I just.
Speaker A:What do I do?
Speaker A:My podcast is called Unreal Engine Diaries.
Speaker A:What do I do?
Speaker A:So what comes to mind is I want to maybe, maybe rename this podcast into Unreal Engines because, because we, we are dealing with those unreal tools and unreal platforms, platforms that are unreal tools that are unreal.
Speaker A:And maybe I will just rebrand my podcast to be that.
Speaker A:And I will claim myself as a tool agnostic.
Speaker A:And I, I just want to get things done and I'm using those unreal devices or devices that are unreal and software that's truly unreal.
Speaker A:And I really hope that Unreal Engine from Epic can deliver in terms of workflow and it can support us creators with limited resources.
Speaker A:I was never motivated to do mediocre game to make a simple game where you collect points and you may die because something eats you.
Speaker A:I just kind of see no point in doing this.
Speaker A:And it seems like doing anything done bigger than this, it's just simply impossible.
Speaker A:All right, so thank you for listening guys.
Speaker A:That's, that's truly just me unwinding and dumping on you thoughts that I was holding back.
Speaker A:Maybe I'll tell you more about specifically about my AI endeavors with AI video generation.
Speaker A:Or maybe I'll just move on to something else in the future.
Speaker A:But that's all for now.
Speaker A:I'm going now to see my sister and help her out with preparations for Easter.
Speaker A:So yeah, I'm completely taken by whole other set of the responsibilities.
Speaker A:I can hear when this probably disturbing audio here as well.
Speaker A:So thank you for listening.
Speaker A:It's a nice half an hour episode.
Speaker A:I hope you enjoyed it.
Speaker A:Keep on listening and hey, email me with your thoughts.
Speaker A:Email address is in show notes and I believe also in the podcast description.
Speaker A:Let me know what you think about what I said.
Speaker A:Share your story.
Speaker A:I'm always looking forward to connect with people who have similar thoughts or even contradicting thoughts.
Speaker A:You know, I want to learn from you as well.
Speaker A:And if you want to be guest on the podcast, you can as well.
Speaker A:It's all up to you.
Speaker A:I'm here for you.
Speaker A:I'm here for myself, probably mainly for myself.
Speaker A:I really enjoy seeing how this podcast is growing.
Speaker A:I don't want to be very specific, but the growth is exponential and the traffic seemed to be real.
Speaker A:I mean you cannot really trust downloads that much.
Speaker A:But what I'm seeing is quite impressive.
Speaker A:And quite believable.
Speaker A:So I'm really glad to see this podcast is growing.
Speaker A:I consider it to be.
Speaker A:Yeah, one of those things that work, one of those things that I consider successful.
Speaker A:So thank you for helping me.
Speaker A:Stay with it and I'll talk to you soon.
