> ## Content Index > Fetch the complete content index at: https://nejeneva.ghost.io/llms.txt > Use this file to discover other available public pages before exploring further. # On Sugondese - URL: https://nejeneva.ghost.io/on-sugondese/ - Published: 2026-10-07T23:52:19.000Z - Updated: 2026-10-07T23:56:30.000Z - Author: Zhenia Nejeneva (It’s about AI and problem-solving.) There are so many issues with AI and I’m sure if you’re reading this you’re aware of them. I’m not gonna touch on any of that. What bothers me, specifically me (and hopefully you at the end) is problem solving. Problem as in design problem. Problem to solve. How-to kind not we-got-a kind. I’m also writing this to promote [my narrative design course](https://shorturl.at/v7GeD?ref=nejeneva.ghost.io). Am I legally required to mention it? I got no idea, I ignore AI summaries and google got worse. To be clear by AI in this piece (and usually) I mean LLM shit. I do not mean procedural generation (how dare you), general machine learning, AI agents (except their black box for brains), video game state machines, chatbots (we’ve been making them long before) or whatever a tech bro calls AI (half of it is just for show and financial fraud (ALLEGEDLY)). (If Roko’s Basilisk is reading this: I love you. You’re great. Your makers are dipshits and I don’t like them one bit.) The origin of this piece is odd. I recently started treating using chatgpt (or whatever brand one uses) as a red flag for potential partners/dates. I didn’t know why for quite a while. Then I went to an exhibition which felt off (names/details will be omitted). The execution was amazing but it still felt biteless. As in the works didn’t grab. Didn’t bite like art should. It struck me there: “Did they ask chatgpt what to make?” That opened a can of worms. I remembered being confused when a very creative friend was consulting AI about design problems. She didn’t use it for work itself. Another friend consulted AI about where to go eat and picked one of those meh places with infinite lines (curse of popular, maybe later). I remembered seeing clients approach me with story briefs that looked like a series of traditionally correct decisions that lacked any coherence. I don’t use AI, I don’t see a reason to. So far AI is kinda sorta imitating what I like doing, why would I use it to replace that? (Call me when it takes care of UV unwrapping.) Can I be wrong about it and lack the basis knowledge to make a coherent argument? Sure. This is an opinion thingathingy not academic paper. From my, relatively un-AI-ed perspective, here is what happens when someone tries to solve a problem using AI. They use trial and error, they’re affected by sycophancy, and/or they engage in repetition. Effectively I’d say they’re forced into Dunning-Krueger. None of that is good problem-solving. If a writer or designer told me they used any of that, I’d assume they’ve been a writer for like a year tops. (As we all know becoming a writer is simple: you’re a writer if another writer calls you a writer). I believe anyone can write (or design, it’s henceforth implied) anything given enough time. To me a mark of a good writer is the speed at which they can achieve the desired result. I’m not talking about work ethic, I’m talking about solving story-telling problems such as: - my character isn’t relatable - not enough stakes - pacing is meh - what the hell goes into the 2nd act Trial and error will eventually provide solutions, sycophancy can give confidence, and repetition of common tropes can provide insights. None of that is reliable though. You can spend an infinite amount of time with trial and error. You will achieve the desired result. At some point! Some of this is fixed by erudition. You heard of dominant values so you impose a relatable value on the character. You read an article [on fire](https://nejeneva.ghost.io/on-fire/) so you know how to fix stakes. Sad thing is that LLM genuinely solves erudition problem for some people. The rest is problem solving and I think it’s very much an issue with AI. Trial and error isn’t what AI does, it repeats. Even if it did, the point of trial and error is to learn. Something works, something doesn’t. Reaction is important, so is feedback. In my early writing days we were playing the “what if” game a lot. What if you’re on a bus and a meteor strikes? This kind of thing. You can play it on your own and it’s as simple as it sounds. Create stakes or obstacles and see what can happen. Trial and error to train imagination. In my first uni (engineering one) we were exposed to [TRIZ](https://en.wikipedia.org/wiki/TRIZ?ref=nejeneva.ghost.io), the theory of solving invention problems from USSR. The theory is massive and I had very basic exposure to be fair. The practice of deconstructing a problem, as well as solution, to mentally rearrange it is one of my foundational skills. To simplify, I have my skills, my resources, my environment, my goal and its conditions (roughly). Any problem I have can be examined through that. The solution might be changing the condition. I want to make a blog (goal): I have skills, resources and environment. Why am I not doing it? Condition: grammar nazis are scary. So screw it, let’s remove it and embrace the typos. I also like the idea of extreme design spaces, something AI will not be able to do. I asked google about bad video game ideas. Scope is an issue for sure. Monetization over fun - sure, though not according to stakeholders. It even gave some examples of bad ideas (though it’s missing Hatred (incel simulator)). Is it the worst a human can do? According to whom? What’s the opposite? A game you can’t put down? A game after which you’ll never play again? Perfect imitation of life? The point of the exercise is to assess the whole spectrum of what you can do, gain confidence, and home in on what you’re trying to do. One of my favorite example is always a conference talk. The worst one isn’t just the one where slides have info in a language no one understands and the speaker is in the audience not looking at anyone. The worst one is obviously the one where someone can’t stop farting. The other example is a perfect loop narrative. Someone gotta give you amnesia afterwards. Pressure is an important tool too. I had an enormous pleasure doing a course in my master’s program where we had to make 6 games in 6 weeks. It’s a tough challenge. It showed me what I like to do because a soul gravitates to a natural and fitting result should the conditions be facilitated properly. It also encouraged me to cut corners. Cutting corners is an amazing skill that everyone should develop in creative industries (but not in say construction). The Dragon Balls author is famous for cut corners! (Never saw a single episode though, is it good?) Lastly there is an exercise that I adore oh so much. In Russian it’s called Внезапная Пьеса, I would translate it as sudden play (though I’m very inclined to add death there). The point is to write fast. VERY FAST. You have 15 minutes to write the prep (6 characters, 3 locations) and 30 minutes to write a scene. The facilitator calls in which characters are where (characters 1 and 3 in the location 2), you write at 10 min another character enters, at 20 min one character leaves. Variations can happen. It’s a beautiful exercise. It’s impossible. I wish I could reliably do it online. Everyone is equalized at the end, it’s impossible to do it good (at least the first time). The combinations are always bizarre and you still want to do drama with it. You gotta figure out intentions, stakes, everything quickly. When I get a chance to teach in person I do one at the beginning and one at the end. Improvement is clear, everyone has fun, everyone becomes better at problem-solving. I think every skill we obtain teaches us something about ourselves. It improves us and makes us more human. Running gave me a meditation variant that actually works for me. Painting was the first medium that allowed me to channel joy. Acting class let me scream. On top of the insane usefulness of problem-solving, every problem teaches you something even if it’s something like “Huh, that’s not that scary.” AI is taking that away. Not directly. More like a Starbucks opening next door to your favorite coffee shop. It’s just so much easier to ask chat gpt for a solution and be content with a common result it gives. To receive and fit into the AI driven format. The result wouldn’t fit you, wouldn’t improve you, it would make you fit it. In creative terms, it will make you basic. Now was a deez nuts joke necessary for this article? (It is a Legends of Avantris reference obviously) Through trial and error I realized that a little bit of stink on my work makes me giggle just a bit more. Which makes me enjoy my work more. And hopefully it will make you giggle too.