The studio’s answer is short: Yes, but very little. That same day, the game gets 366 negative reviews on Steam. The day before, it had one. The next day, Big Walk wins the Critics’ Choice award at the Golden Joystick Awards in London.

Some say players were told the truth too late. Others call it an overreaction to a ten-line bug fix. I’m in the second camp. But first, let’s look at what happened over those three days.

It started with a jury statement

Big Walk is a multiplayer adventure game published by Panic and released on August 4, 2026. It’s on PC, Mac, PS5 and Switch 2, and it had been sitting at a high rating on Steam since launch.

The debate didn’t start with the game. It started with an awards jury. On November 9, 19 jurors from the Independent Games Festival (IGF) announced they wouldn’t vote for any game that used generative AI in any form. They were speaking for themselves, not for the festival, and the statement didn’t mention Big Walk. But on social media, the conversation quickly attached itself to the game.

House House said it had originally written its AI disclosure for its IGF submission. The next day, the same disclosure went up on the game’s FAQ page. Nobody caught the studio red-handed. The studio wrote it down itself.

What the studio said

According to House House, AI was used in the final stretch of development to work through difficult technical problems and bugs. The team consulted chatbots and, in some cases, used code snippets from their answers as a starting point before modifying them. By the studio’s own estimate, this touched well under 0.1% of the game’s code. Generative AI wasn’t used anywhere else, including visuals, audio or text.

The example the studio gave fits that picture. For a shadow bug coming from Unity, they took a chatbot’s suggestion, reworked it heavily, and the final fix came to about 10 lines. That’s one example, not the total usage. The studio didn’t explain how it arrived at the 0.1% figure, and no source has independently verified it. Still, nothing the player sees, hears or reads is part of this debate.

The word “AI” throws a lot into one bag here. The systems that control enemies are called AI. So are the algorithms that build maps at random. What’s being debated is generative AI, the kind that produces text, images or code. And even that varies: a programmer asking a chatbot for help while tracking down a bug is one thing. A texture the player sees on screen being drawn by AI is another.

Why this wasn’t on the store page

Steam asks games that use AI to disclose it on their store page. But in January 2026, Valve clarified the form: the disclosure covers content players directly see and hear. Behind-the-scenes tools like code assistants are out of scope. That’s why Big Walk’s Steam page has no AI note.

So someone buying the game couldn’t have learned this from the store. I don’t see that as the player’s failing. The platform sets the rules and the studio makes the disclosure. Players digging into how every single game was made is neither possible nor their job.

What the numbers say

According to Steam’s daily review data, the game got 94 positive reviews and 1 negative review on November 9. On November 10, positive reviews rose to 426 and negative reviews to 366. On November 11, there were 652 positive and 130 negative reviews.

On Steam, waves like this, which arrive fast and usually react to an event more than to the game itself, are called “review bombing.” There’s no evidence that the wave against Big Walk was an organized campaign. But one detail stands out: as negative reviews went up, so did positive ones. People defending the game were writing in the same days.

As of November 12, Big Walk’s overall rating on Steam is still “Very Positive.” 94% of its 31,281 English reviews are positive. Over the last 30 days, that share has dropped to 87%.

What people are objecting to

Not all the negative reviews say the same thing. In the ones we read, some object to AI itself. Others object to the timing: the game came out in August, the disclosure came three months later, and Steam’s refund window had long since closed. Some players also write that they liked the game but won’t support it anymore.

The timing complaint is fair. Everyone who bought the game should have known this from the start, and the studio itself admitted that not having an internal AI policy was its own failure. House House said the disclosure wasn’t an endorsement of AI tools, apologized, and promised a more thorough policy.

Choosing not to support a game is everyone’s right, too. What I find hard to understand is something else: a game you enjoyed and loved playing getting torn apart over a ten-line bug fix that left no trace on screen. There’s no record showing the same people loved the game first and then changed their minds. Steam doesn’t show that. But that scenario, turning your back on something you love overnight, doesn’t strike me as a healthy reflex.

When a backlash makes sense

Not every backlash is the same. In December 2025, the Indie Game Awards took back the Game of the Year and Debut Game awards it had given to Clair Obscur: Expedition 33. There, placeholder textures made with generative AI had made it into the release version, right where players could see them. The studio had also confirmed in its submission that it didn’t use AI. In Big Walk’s case, according to the studio, the usage was somewhere players can’t see, and it was spelled out in the IGF submission. There was talk on social media that Big Walk had also been disqualified from an award. There’s no official announcement of that.

Then there’s Domina. In 2022, Dolphin Barn’s gladiator management game Domina was pulled from Steam after the developer posted an announcement targeting a trans streamer and the trans community. Valve said the announcement contained insults targeting another person. The issue there wasn’t how the game was made. It was the developer’s posts attacking people.

That’s exactly where I draw the line. If a developer disrespects people, a community’s values or a sensitive subject, a backlash is understandable. Piling on a game you love because it got some invisible help in the code is a long way from that line. All three cases get described with the same words, “backlash” and “boycott,” but one is an invisible bug fix, one is undisclosed content you can see on screen, and one is behavior aimed directly at people.

What players think

There’s no player survey specifically about coding help. For the general picture, there are two data points. According to a Circana survey from December 2025 in the US, a little over a quarter of players say they’d be less likely to buy a game if they learned generative AI was used in its art, dialogue, text or music; most are neutral or undecided. In a Quantic Foundry survey from October to December 2025, made up mostly of core gamers, 85% held a negative view. Both surveys ask about visible content. There’s no separate data on invisible usage like Big Walk’s.

Players shouldn’t have to carry this

I don’t think these tools can be kept out of game development. That doesn’t make them good. The consent issue in training data and the value of programmers’ work are real debates. But the place to have those debates isn’t punishing a beloved game from a studio that openly disclosed its own usage. It’s platform disclosure rules and studio policies.

At a time when people complain that stores are filling up with rushed, AI-stuffed games, finding a game that genuinely makes you happy, takes your mind off things and makes you say “this is exactly what I was looking for” is no small thing. Once you’ve found a game like that, I don’t think where ten lines of code came from should change most players’ experience. And as more work gets made with AI, I think work made by hand, with real effort, will stand out even more. That’s an expectation; there’s no data showing it yet.

Big Walk’s Steam rating is still “Very Positive,” and even during the three harshest days of the debate, positive reviews outnumbered negative ones. As I see it, there’s no reason for players to carry this stress around all the time.

Piling On a Game You Love Over 0.1%: What the Big Walk Backlash Really Shows was originally published in HumanAI on Medium, where people are continuing the conversation by highlighting and responding to this story.

China has been making increasingly ambitious moves in the gaming industry. Over the past few years, we’ve seen major action RPGs, Soulslike experiments, high-budget productions, and free-to-play games emerge from Chinese studios. Black Myth: Wukong, Wuchang: Fallen Feathers, and now Phantom Blade Zero have made it clear that Chinese developers are becoming serious competitors in a field long associated with Japanese studios.

And technically, there’s plenty to admire.

Phantom Blade Zero looks incredible. Its sword animations are remarkably fluid, its character designs are elaborate, and some of its boss encounters resemble carefully choreographed sequences from a martial arts film. Watching these fights unfold, it’s easy to understand why people immediately began comparing the game to Sekiro: Shadows Die Twice.

But I have a problem with the idea that Phantom Blade Zero has somehow surpassed Sekiro simply because it looks more impressive or delivers more spectacular combat.

To me, that comparison misses much of what makes Sekiro exceptional.

Sekiro was never just about sword fighting. And FromSoftware’s greatest strength has never been its ability to produce the most technically impressive visuals.

There’s a particular weight to its games, something that goes far beyond graphics, mythology, environmental design, or even difficulty. It’s present in the way characters are introduced, the tension that builds before a confrontation, the consequences of failure, and the feeling that every significant encounter carries something beyond the fight itself.

Phantom Blade Zero may share some of Sekiro’s combat mechanics, but I don’t believe it approaches them with the same philosophy.

The Weight That Graphics Cannot Reproduce

When people talk about atmosphere in Soulslike games, the discussion often revolves around familiar elements: dark environments, mysterious lore, ruined castles, grotesque creatures, and imposing architecture.

All of these contribute to the experience, but they don’t fully explain what makes FromSoftware’s games feel the way they do.

Take Sekiro’s confrontation with Genichiro Ashina.

By the time you face him on the rooftop of Ashina Castle, he’s already an established threat. He’s directly connected to Kuro’s fate, and your previous encounter with him ended in defeat. The fight isn’t simply another obstacle between you and the next area. It’s an opportunity to confront someone who once completely overwhelmed you.

What makes the encounter memorable isn’t just the character’s design or the environment surrounding him.

It’s the way Sekiro’s combat system supports that confrontation.

You’re constantly pressured to stay close, attack, deflect, and respond to your opponent’s movements. Hesitating gives the enemy time to recover. Attacking recklessly leaves you exposed. Winning requires you to understand the rhythm of the duel rather than simply overwhelm your opponent with damage.

The posture system plays a major role in this. Breaking an enemy’s stance requires sustained pressure, while your own mistakes can quickly turn the fight against you.

The result is a kind of controlled aggression where the player remains vulnerable even while attacking.

And then there’s the Great Shinobi Owl encounter.

Fighting the man who raised you, after being forced to choose between loyalty to him and protecting Kuro, gives the confrontation a personal significance that no amount of visual detail could create on its own.

These fights are memorable because the mechanics, characters, and narrative context reinforce one another.

That’s the weight I’m talking about.

Sekiro doesn’t need every sword swing to produce an elaborate visual effect. It doesn’t need every boss to perform a spectacular sequence of impossible movements. A series of precisely timed sword clashes can create more tension than an entire screen filled with effects.

This is where I find the comparison with Phantom Blade Zero increasingly difficult.

Both games feature demanding combat, but I don’t think they’re equally interested in making that combat feel the same way.

When a Boss Fight Becomes a Performance

One of the clearest examples of Phantom Blade Zero’s different philosophy is the Lion Dance Grandmaster encounter.

The fight begins with traditional Chinese lion dance drums, drawing on a recognizable cultural performance. But during its second phase, the music shifts toward rock as the encounter becomes increasingly elaborate.

This wasn’t an accidental choice.

Director Soulframe Liang has explained that the transition was intentional, reflecting the game’s broader attempt to combine traditional Chinese martial arts with modern influences. The developers call this approach kungfupunk.

I understand what they’re trying to achieve.

But understanding a creative decision doesn’t necessarily mean I find it effective.

Personally, the moment a boss fight begins to resemble an elaborate performance accompanied by electric guitars, some of the tension I associate with a deadly confrontation starts to disappear.

The spectacle becomes the focus.

The fight begins to feel less like a desperate struggle against a dangerous opponent and more like an action sequence designed to impress the audience.

And while that might be exactly what Phantom Blade Zero is going for, it’s not what I look for in a game like Sekiro.

FromSoftware’s approach to boss encounters, particularly in Sekiro, is often much more restrained in its presentation. The sound of clashing swords, the pauses between attacks, the enemy’s movements, and the pressure of maintaining your posture are all essential to the encounter.

The player isn’t simply watching a beautifully choreographed fight unfold. They’re responsible for surviving it.

Of course, rock music doesn’t inherently make a boss fight less serious, and Sekiro isn’t devoid of spectacular or even absurd encounters. The Guardian Ape alone should make that obvious.

But there’s still a difference in how these games use spectacle.

Sekiro can be fantastical without losing the feeling that the player is engaged in a dangerous confrontation. Phantom Blade Zero often seems more interested in making that confrontation look extraordinary.

That distinction matters to me far more than the number of animations, visual effects, or instruments involved.

Japanese and Chinese Mythology Don’t Carry the Same Cultural Familiarity

There’s also a cultural dimension to this comparison that I think deserves more attention.

Japanese popular culture has spent decades establishing itself internationally. Anime, manga, Japanese cinema, and video games have created an enormous audience familiar with Japanese aesthetics, storytelling traditions, character archetypes, and historical imagery.

And that familiarity has continued to grow.

For many Western players, elements associated with samurai, shinobi, Japanese folklore, and Sengoku-era imagery are already part of a recognizable cultural vocabulary.

You don’t necessarily need to understand every historical or religious reference to feel some familiarity with the setting.

Sekiro benefits from that cultural background, even though the game isn’t a direct adaptation of Japanese mythology.

China, meanwhile, has a different history of international cultural exposure.

Chinese folklore, wuxia, and martial arts traditions have certainly reached global audiences. Films such as Crouching Tiger, Hidden Dragon demonstrate that these traditions can have enormous international appeal.

But I don’t think their familiarity among many Western gaming audiences can simply be assumed to match that of Japanese popular culture.

Japanese games, animation, and associated aesthetics have been part of international gaming culture for generations. Chinese fantasy and wuxia traditions have not necessarily developed the same level of recognition among those audiences.

That doesn’t make one tradition artistically superior to the other.

It does, however, influence how audiences approach unfamiliar fictional worlds.

When I see Sekiro’s interpretation of feudal Japan, its architecture, costumes, character relationships, and supernatural elements fit into an aesthetic tradition I’ve encountered repeatedly through games and other media.

Phantom Blade Zero draws from a different cultural language.

Its world is influenced by wuxia, the traditions of martial arts cinema, and a deliberately exaggerated combination of historical and modern imagery.

Those differences aren’t superficial.

They influence how characters move, how confrontations are staged, how dramatic moments are presented, and what kind of emotional response the game tries to create.

China’s growing success in high-budget gaming doesn’t automatically mean its productions will reproduce the particular appeal Japanese games have developed over decades.

And I don’t think Phantom Blade Zero should be treated as though it has achieved the same cultural and artistic effect as Sekiro merely because both games feature swords and East Asian settings.

Even the Sound of a Language Changes the Atmosphere

Voice acting is another part of this discussion that I don’t think should be overlooked.

Language has its own rhythm, cadence, and sonic character. These qualities affect how dialogue feels, especially in games where atmosphere depends heavily on how characters speak.

Personally, I’ve always found Japanese particularly compelling in dramatic game dialogue.

There’s something about its rhythm, the delivery of certain expressions, and the way pauses and vocal emphasis work together that I find especially effective in serious confrontations.

The Chinese dialogue I’ve heard in Phantom Blade Zero doesn’t produce quite the same response in me.

This is partly a question of phonetics and prosody. Mandarin Chinese and Japanese have different sound systems, tonal patterns, and rhythmic structures. Those differences are audible, and they contribute to how I experience a performance.

I also recognize that familiarity matters. Years of exposure to Japanese games, anime, and cinema can shape what sounds natural, dramatic, or emotionally convincing to a listener.

But that doesn’t make the aesthetic experience any less relevant.

The sound of a language is part of how we experience fiction.

Consider Tolkien’s Elvish languages. Quenya and Sindarin weren’t simply collections of invented words. Their sounds, structures, and historical influences were carefully developed as part of Tolkien’s fictional world.

For many readers and listeners, the way those languages sound contributes to the atmosphere of Middle-earth.

I see a similar relationship between voice acting and atmosphere in games.

A language’s rhythm can affect whether a character sounds intimidating, tragic, mysterious, or emotionally distant to a particular listener.

That doesn’t establish an objective hierarchy between Japanese and Chinese. It means I don’t experience the two languages identically, especially when they’re being used to create very different dramatic worlds.

In Sekiro, the Japanese dialogue feels closely connected to the atmosphere I associate with its characters and setting.

Phantom Blade Zero’s Chinese performances belong to another artistic tradition, and they don’t necessarily create the same emotional effect for me.

I don’t think that difference should be ignored simply because the games come from neighboring regions of Asia.

Spectacular Movement Isn’t the Same as Meaningful Combat

Another aspect of Phantom Blade Zero that distances it from Sekiro, at least in my eyes, is its visual approach to physical movement.

Some of its encounters feature characters suspended in the air, elaborate acrobatics, extremely fast combinations, and movements that seem almost entirely detached from ordinary physical limitations.

The Fallen Hope encounter is a particularly striking example. Its suspended figure and attacks involving red strings create a style of combat that feels closer to a highly stylized martial arts performance than a conventional sword duel.

There’s nothing inherently wrong with this approach.

Wuxia cinema has a long history of exaggerated movement and wire-assisted choreography. Not every impossible leap or airborne attack should be dismissed as an imitation of anime.

Still, the visual language can resemble the kind of exaggerated confrontations commonly associated with anime action sequences.

And that aesthetic doesn’t appeal to me in quite the same way as Sekiro’s combat.

Sekiro isn’t realistic either. You can grapple between rooftops, leap over enemies, perform airborne attacks, and fight supernatural creatures.

But its combat still revolves around a relatively focused relationship between the player, the opponent, and the timing of their attacks.

Even when the game becomes fantastical, that underlying structure remains central.

Phantom Blade Zero appears much more willing to turn combat into an elaborate sequence of movements, weapon changes, and cinematic flourishes.

Its combat mechanics are not merely decorative. Reviews of the finished game describe genuine depth in its deflection systems, weapon management, and boss encounters. I wouldn’t dismiss those achievements.

But mechanical depth and artistic identity aren’t interchangeable.

A game can demand considerable skill while presenting that skill through a style I find less compelling.

And for players who particularly appreciate FromSoftware’s restrained presentation, punishing mistakes, and tense duels, Phantom Blade Zero’s emphasis on acrobatic spectacle may not satisfy the same expectations.

The ability to execute dozens of extraordinary movements doesn’t automatically make a sword fight more memorable.

Sometimes a single well-timed deflection can accomplish more.

Phantom Blade Zero Doesn’t Need to Be Another Sekiro

I don’t want to dismiss what Chinese studios have achieved.

Phantom Blade Zero is an ambitious production with a distinctive visual identity, an elaborate combat system, and clear creative influences. Its use of wuxia traditions and modern martial arts cinema gives it a character of its own.

But that’s also why I find the constant comparisons with Sekiro so unconvincing.

The two games can certainly be compared mechanically. We can examine their deflection systems, boss design, difficulty, animation quality, or responsiveness.

What I reject is the assumption that improvements in technical presentation or more spectacular choreography automatically make Phantom Blade Zero a superior version of Sekiro.

Because Sekiro was never trying to be the most spectacular martial arts film you could play.

Its appeal lies in a particular relationship between mechanical discipline, dramatic tension, and storytelling.

The weight of a sword strike matters because of what the game asks you to risk. A boss matters because of the journey that brought you to that encounter. A victory matters because you were forced to understand the opponent rather than simply overpower them.

That combination is what makes Sekiro so memorable to me.

Phantom Blade Zero approaches combat with a different set of priorities. It embraces elaborate choreography, modern musical influences, extravagant movements, and a cinematic style rooted in Chinese martial arts traditions.

And while I can appreciate the technical achievement behind that vision, I don’t find it inherently more compelling.

Better graphics don’t necessarily create a stronger atmosphere. More elaborate animations don’t automatically make a fight more meaningful. And a louder, more spectacular boss encounter doesn’t always leave a deeper impression.

I would much rather face an opponent whose every movement forces me to think than watch an encounter become increasingly elaborate simply for the sake of spectacle.

For me, Sekiro remains the more convincing expression of what makes this kind of combat special.

Not because Japanese culture is inherently superior to Chinese culture, or because every Soulslike should follow FromSoftware’s formula, but because Sekiro brings its mechanics, atmosphere, sound, and dramatic context together in a way that Phantom Blade Zero’s more cinematic approach doesn’t replace.

Phantom Blade Zero may offer a more spectacular sword fight. But spectacle alone isn’t what made Sekiro exceptional, and it certainly isn’t enough to surpass it.

Both games were widely acclaimed when they came out. Each had its own approach to design and left a mark on the gaming world. But someone playing them for the first time today might enjoy Arkham Asylum far more, while struggling to understand why Ocarina of Time is held in such high regard.

Another player might feel that Zelda still stands apart from similar games. So what does it really take for a game to be considered a classic?
Is it the impact it had at release, the experience it offers years later, or the players who keep its memory alive?

To understand that, we need to go back to 1998.

At the time, game developers were grappling with the challenges of moving from two-dimensional worlds into three-dimensional environments. Characters no longer moved only left, right, up, or down. Players had to manage the camera, keep track of enemies, and interact with their surroundings. Many of the control conventions we take for granted today had yet to be established.

Released for the Nintendo 64 in 1998, Nintendo’s Ocarina of Time brought the Zelda series into a three-dimensional world for the first time. Its Z-targeting system let players lock onto an enemy or character, making it easier to manage the camera and their movements. Moving around an opponent, dodging attacks, or talking to a character became easier to understand.

And it wasn’t just about combat. As players acquired new tools, they could reach previously inaccessible areas, solve environmental puzzles, and use melodies they had learned to affect how the world worked. The game built a structure that tied exploration to progression.

But it would be wrong to suggest that Ocarina of Time was the first to introduce all these ideas. Mega Man Legends, released in 1997, also featured a system for locking onto enemies. Nintendo’s contribution was to bring together ideas already seen in different games within its own adventure design. That’s why it’s important to distinguish historical innovation from being the first to invent a particular mechanic.

Now let’s move to 2009, and Batman: Arkham Asylum.

This time, the developers faced a different challenge. How could controlling one of the most recognizable characters in comics actually make players feel like Batman?

Rocksteady didn’t focus solely on the character’s fighting abilities. Players could confront enemies directly, hide in the shadows and take them down one by one, examine their surroundings using Detective Vision, and use different gadgets to access new areas. The combat system, known as Freeflow, linked attacks and counterattacks into fluid sequences of movement.

In a 2009 interview, game director Sefton Hill explained that the combat system went through three major design iterations before reaching its final form. The goal was to avoid making actions that should come easily to Batman unnecessarily difficult for the player.

The same interview revealed a lesser-known detail. Hill explicitly named the Zelda and Metroid series as major inspirations for Arkham Asylum. He described how they had drawn on those games, particularly when it came to using different tools together and creating a consistent, cohesive game world.

That doesn’t mean Arkham Asylum was directly copied from Ocarina of Time. Hill was referring to the broader design philosophy of the Zelda series, not one particular entry. Still, his comments show how games from different eras can be connected through their design ideas.

But what do these games’ review scores actually tell us?

The Nintendo 64 version of Ocarina of Time holds a Metacritic score of 99, based on 22 critic reviews. The Xbox 360 version of Arkham Asylum has a score of 92, based on 78 critic reviews. Both games received exceptionally high scores from critics. But those figures, drawn from different years, platforms, and numbers of reviews, can’t definitively establish which game is better today.

Even games that were celebrated in their day had aspects that drew criticism. In the original release of Ocarina of Time, particularly in the Water Temple, players had to repeatedly open the menu to switch certain pieces of equipment. This was one of the issues the developers specifically addressed in the Nintendo 3DS remake released in 2011.

In Arkham Asylum, meanwhile, some critics felt that Detective Vision, which highlighted enemies and interactive objects, provided too much guidance. One complaint was that it encouraged players to use this mode constantly instead of viewing the world normally.

So a game being praised at release doesn’t necessarily mean every one of its mechanics will be received the same way years later. How, then, do games maintain their reputations over time?

In 2022, Ocarina of Time was inducted into the World Video Game Hall of Fame by The Strong National Museum of Play. In 2024, it also won a Game Informer readers’ tournament featuring 64 games and more than 200,000 votes. These examples show that, decades after its release, the game continues to be recognized by institutions and gaming communities alike.

Arkham Asylum’s story didn’t end in 2009, either. It was re-released in 2016 as part of the Return to Arkham collection. In 2023, it arrived on Nintendo Switch through the Arkham Trilogy collection. Arkham Shadow, set in the same series universe, was released as a virtual reality game in 2024.

Here, we can see two different processes at work. One game can find a place in museum collections and greatest-of-all-time lists because of its historical significance. Another can remain part of the conversation through sequels, re-releases, and its player community. These examples help us understand how well a game is remembered, but they don’t tell us which player will enjoy it more.

Nostalgia is part of the discussion, too. Returning to games from childhood years later can stir up different emotions. Research into video game nostalgia suggests that past experiences can shape the way people relate to games. But there isn’t enough evidence to say that everyone who enjoys an older game does so purely out of nostalgia. Someone encountering that same game for the first time as an adult brings a different experience to their assessment.

There’s another question worth asking: who writes video game history?

Stanford University’s Game Canon Project distinguishes between three areas when examining games’ lasting reputations: their place in popular culture, their critical recognition, and their significance in academic research. The project looks at different indicators, ranging from sales and awards to academic citations.

That’s because being played by millions of people, influencing other developers, and appearing on critics’ best-of lists are not the same thing. The games these lists include, the criteria used to compile them, and the audiences they address can all shape the results.

Someone playing Ocarina of Time for the first time today might find its controls outdated. Another might still be drawn to its dungeons and approach to exploration. Some players find Arkham Asylum’s combat more fluid, while others feel that certain systems are overly simplified.

Over the years, game technology, player habits, and our expectations have changed. But the mark a game left on the past and the experience it offers today don’t always move in the same direction.

When we call a game a classic, sometimes we’re talking about what it achieved in its own time, sometimes about its influence on later games, and sometimes simply about the fact that it hasn’t been forgotten. There is no universally agreed-upon standard for deciding which of these matters most.

And perhaps the question is still the same: when we judge a game, are we looking at the game we can play today, or at what it meant when it first came out?

October 7, 2026. Google announces Playground, an experimental platform where users can describe their game ideas in a few sentences, turn them into playable games, and share them with friends. And they don’t even need to know how to code.

On the same day, Unity introduces Spark, a new tool developed in partnership with Google. A day later, Roblox publishes the first results of its own AI-powered game creation system.

The three companies are developing different products, but they’re working toward much the same goal: reducing the technical knowledge and effort needed to turn an idea into a working game.

Meanwhile, tens of thousands of new games are being added to Steam every year. AI-assisted projects are becoming more common, older games are being adapted to different systems, and technical tasks that small teams once struggled to handle are becoming increasingly accessible.

On one side are people who have never made a game before but now have the opportunity to try. On the other are studios that have spent years working on their projects and are already struggling to attract players in a crowded market.

If game development really becomes accessible to everyone, what kind of competition will independent developers face? How much of game creation can AI actually take over? And how will digital storefronts deal with the growing volume of content?

First, let’s look at what these new tools can actually do. Because creating a game with a few sentences certainly sounds impressive. Whether that game is actually worth playing is another matter.

Is Describing a Game Idea All It Takes Now?

Google’s Playground turns game development into something that almost resembles a messaging experience.

Say you want to make a game about a character jumping between platforms while escaping dragons. You describe your idea in writing. The system generates the game logic, visuals, and necessary code. You can then try the game directly in your browser.

Want the character to jump higher? Change how the enemies behave or how the levels are designed? You can describe those adjustments in writing, too.

And Playground doesn’t just generate images that look like games. According to Google’s help center, users can download their creations as ZIP files containing HTML, JavaScript, CSS, and generated visual assets.

So, at least for supported games, the platform produces actual software files that can be examined and modified. That’s quite different from simply generating moving images that look like a game.

Of course, being able to download the code doesn’t guarantee that the game will run smoothly on another platform or that you have unlimited commercial rights to every visual asset it uses. Making games easier to create hasn’t magically solved technical compatibility or licensing issues.

For now, Playground is available to users aged 18 and older in the United States. Access to its game creation features is being rolled out gradually, depending on subscription status and available capacity.

Unity’s Spark project aims to go further.

Unity plans to combine its own game runtime, professionally created assets, and AI-powered development tools. The goal is to support more advanced 3D environments, game mechanics, and projects that multiple people can work on together.

There’s one small but important detail here: Spark isn’t publicly available yet. For now, all we have are the company’s announcements, promotional demonstrations, and planned features.

So it wouldn’t be fair to evaluate Playground and Spark as though they’re two products at the same stage of development.

Still, the approach behind both platforms tells us something important about how game development is changing.

Until a few years ago, AI tools were primarily expected to help professional developers work faster. Now companies are trying to make it possible for someone who has never even opened a game engine to create a game directly.

But are the games they produce actually any good?

A Game in Eight Minutes: Engineering Quality 98, Game Design 31

In a technical report published on October 8, Roblox explained how it evaluates the performance of Build, its AI-powered game creation tool.

Build allows users to create Roblox games through written instructions in everyday language.

During alpha testing in New Zealand, more than 100,000 people tried the tool, and approximately 9,000 games were published.

According to Roblox, 71% of the people who tried Build had never used Roblox Studio before.

That’s a significant result. When technical barriers come down, more people are willing to give game development a try.

But to understand the difference between making a game and designing a good one, we need to look at another example Roblox shared.

The company tested a 3D platformer called Skyforge Ascent.

The game featured moving obstacles, checkpoints, and a finish line that players needed to reach. The AI system had created all of this in approximately eight minutes.

Roblox’s own evaluation system gave the game the following scores:

Evaluation category;
Engineering quality 98/100
Visual quality 86/100
Playtest results 100/100
Game design 31/100

The game launched. The character moved. The obstacles and checkpoints worked. Every automated test passed successfully.

Pretty impressive for eight minutes of development.

Unfortunately, the game design evaluation revealed a lack of meaningful progression, variety, original gameplay ideas, and reasons to play again.

In other words, the system had successfully made the game work. It just hadn’t given players much reason to launch it a second time.

And that highlights two very different sides of game development.

Making a character jump between platforms is a technical task. Deciding how that jump should feel, when the player should face a challenge, what order the obstacles should appear in, and why someone would want to continue after finishing a level is all part of game design.

Of course, Roblox’s scoring system isn’t an independent industry standard. We can’t draw definitive conclusions about the quality of all AI-generated games from a single example.

Still, even the company’s own example shows just how far apart technical achievement and game design quality can be.

TechRadar writer Eric Hal Schwartz encountered a similar situation when he tested Google Playground in a review published on October 9.

Schwartz created two games. One was a platformer featuring a dragon, while the other was an RPG satirizing bureaucracy.

Both became playable relatively quickly. However, problems with the controls, limited quest design, and shortcomings in the narrative showed that creating a more substantial game would require additional work.

Someone with no programming knowledge being able to create a playable RPG in a matter of minutes is certainly nothing to dismiss.

But if we decide that game development ends the moment the first working version appears, we’d also have to forget about testing, balancing, content development, and bug fixing.

Unfortunately, skipping those steps doesn’t make games any better.

Will AI Transform Indie and Small Games First?

Developing a massive open-world RPG and building a small browser-based platformer don’t involve the same technical requirements.

A large RPG needs character behavior systems, interconnected quests, save systems, performance optimization, progression mechanics, and numerous systems that must remain consistent across dozens of hours of gameplay.

A small puzzle or arcade game, meanwhile, may involve far fewer systems with much simpler relationships between them.

That’s why it’s reasonable to expect some of the earliest significant effects of AI-powered game creation to appear in short games, prototypes, and mini-games.

For example, someone could create a simple racing game, a short horror experience, or a small multiplayer game to enjoy with friends. Compared with larger projects, these games are better suited to rapid experimentation and making changes based on the results.

There are similar opportunities for independent developers.

A solo developer could use AI to help with tasks that would normally require outside assistance. Testing new ideas, preparing placeholder visuals, writing specific pieces of code, or developing early versions of gameplay mechanics could become easier.

But there’s an important distinction to make here: Developing a small game and developing an independent game aren’t the same thing.

An independent studio can create an extremely complex game. A major company can release a very simple mobile game.

So there isn’t enough evidence yet to claim that AI will take over the entire indie gaming market. How quickly different genres change will depend on how reliable the tools become and how developers use them.

That said, expanding the technical capabilities of small teams could have significant consequences for competition.

If certain tasks that once took a developer months can now be completed in less time, experimenting with more ideas within the same period becomes possible.

That could allow original ideas that previously couldn’t be developed to finally become reality. It could also make producing new versions of familiar games that look remarkably similar to one another much cheaper.

So the same technology could encourage creative experimentation while also increasing the number of games that are practically identical.

The gaming market already had plenty of choices. Apparently, making even more of them is about to get easier, too.

Steam Was Already Crowded. Will AI Make Things Even Worse?

It’s hard to argue that Steam is suffering from a shortage of new games.

According to SteamDB records, 1,712 games were released on the platform in 2014. By 2025, that number had reached 21,256.

Annual game releases on Steam

Comparison of 2014 and 2025. Source: SteamDB

2014 > 1,712

2025 > 21,256

Annual game releases increased roughly twelvefold in eleven years.

And much of that growth happened before today’s generative AI tools became widely available.

Accessible game engines like Unity and Unreal Engine, simpler development tools such as GameMaker and RPG Maker, marketplaces selling ready-made assets, and digital distribution platforms had already been reducing the cost of making games for years.

AI may be another stage in a long-running trend rather than the beginning of an entirely new one.

But one possible consequence of falling production costs is that games requiring little effort, repeating existing ideas, and receiving insufficient testing become easier to publish.

The increasingly common term AI slop is generally used to describe this kind of low-quality, mass-produced AI content.

Still, putting every game that uses AI assistance into the same category would be misleading.

A developer using AI to find bugs in their code isn’t following the same production process as someone who generates all their game’s visuals with a few prompts and publishes them with barely any further work.

The 2026 Game Developers Conference survey also shows just how complicated the industry’s relationship with AI has become.

In a survey of more than 2,300 industry professionals, 36% of respondents said they use generative AI in their work.

But that doesn’t mean they’re all using it to create games directly. The reported uses included brainstorming, everyday tasks, coding assistance, and prototyping.

In the same survey, 52% of respondents believed AI was having a negative impact on the gaming industry, while only 7% viewed its impact positively.

So while some industry professionals are using the technology, that doesn’t necessarily mean they’re happy with it or optimistic about where the industry is heading.

These developments create two possible scenarios for smaller developers.

In the first, ideas that couldn’t be realized because of limited budgets become possible. Experimental mechanics, unusual stories, and commercially risky projects have more opportunities to get made.

In the second, falling production costs make it much easier to upload similar-looking games to digital storefronts. Small studios that already struggle to get noticed suddenly have even more competitors to deal with.

It’s too early to say which outcome will dominate.

But there’s one thing that isn’t likely to change: Adding more than 20,000 games to Steam every year doesn’t magically give players another 20,000 hours in their day.

Developers can produce more games, but players still have the same amount of time.

Are New Ports and Mods Part of This Shift Too?

Alongside new game creation tools, projects that modify existing games or bring them to different systems have also been attracting attention recently.

Examples include systems attempting to bring Elden Ring’s combat mechanics into other games, experiments connecting Minecraft’s gameplay rules to different worlds, ports adapting older games to modern platforms, and AI-assisted coding projects.

From the outside, these developments can make it seem as though the technical boundaries between game engines are gradually disappearing.

But the methods behind these projects are quite different.

Some developers recreate a game’s behavior in another programming language. Others transfer data between two separate games running on the same computer. Developers working on ports adapt older game code to run on different hardware or operating systems.

Technically, these aren’t the same processes.

And not all these projects involve AI. Modding, reverse engineering, and game porting existed long before generative AI.

Examining how games work and adapting them to other systems aren’t newly discovered techniques. We simply have new tools that can make those processes faster.

And that’s exactly where AI-powered coding tools could make a difference.

Examining a large, outdated codebase, rewriting specific systems, or investigating bugs across different platforms can take smaller teams a considerable amount of time.

AI agents taking on some of that work could make it easier to develop projects that previously weren’t considered worth the effort.

But removing technical obstacles doesn’t automatically eliminate other problems.

For example, being able to run a game on another platform doesn’t give someone the right to distribute its original content.

Similarly, taking inspiration from a game’s mechanics isn’t legally the same as using its copyrighted characters, visuals, or music without permission.

The requirements for making code work are different from the requirements for legally publishing a game.

So as game development becomes easier, the distinction between what developers can technically achieve and what they’re actually allowed to publish may become increasingly important.

Could Digital Game Stores Introduce New Rules?

If the number of AI-generated games increases, questions about which content digital platforms should accept, how they should evaluate it, and how it should be presented to players could become more prominent.

And we don’t need to wait for hypothetical future legislation to see the first examples. Some platforms already have rules in place.

Valve requires developers submitting games to Steam to disclose generative AI content that players may encounter.

The Steamworks Content Survey distinguishes between two types of AI use.

The first covers visual, audio, or narrative content generated with AI during development and included in the finished game.

The second involves content generated by AI in real time while the game is running.

Developers using real-time AI generation must also explain what safeguards they’ve implemented to prevent the system from producing illegal content.

Valve doesn’t treat coding tools used solely to improve developer productivity in the same way as AI-generated content that players directly experience.

Google Playground also has systems for reviewing public games before they’re published and allowing users to report content.

These aren’t new laws introduced by governments. They’re publishing and moderation policies established by the platforms themselves.

As the volume of content grows, issues such as publishing similar games as spam, intellectual property violations, misleading promotional materials, and moderating automatically generated content could lead to additional measures.

But coming up with a single rule to address all these problems isn’t easy.

Overly strict policies could also restrict small developers who only use AI as an assistive tool or want to experiment with new ideas.

Rules that are too lenient, meanwhile, could make it easier for low-quality content, misleading storefront pages, and imitation projects to spread.

There’s also an important distinction between a poorly designed game and one that actually violates a law or platform policy.

A game being boring isn’t, by itself, a legal reason to ban it. Which makes the seemingly convenient solution of removing every game with a sufficiently low design score from the store a little more complicated.

That’s why it may matter whether future regulations focus simply on the use of AI or on the safety, ownership, and quality of published content.

For now, we can’t say that Google Playground or Unity Spark has directly caused any new legislation.

But in a market where more people can publish games in less time, the way digital storefronts manage their content will become increasingly important.

Why Could Getting Noticed Become Harder as Making Games Gets Easier?

Once a developer finishes a game, they face a completely different problem.

How are players even supposed to know it exists?

A game published on Steam can appear in various discovery sections of the store. But visibility depends on more than simply having a game available for sale.

According to Valve, sales, player interest and preferences, the store’s recommendation systems, and the rules governing different discovery sections all influence which games are shown to which users.

Google Playground’s Explore gallery also takes player ratings and play activity into account. Google additionally plans to highlight selected games across its own services.

This is where the difference between the cost of developing a game and the cost of reaching players becomes more apparent.

A developer might use AI to create their first prototype much faster than before. But that doesn’t mean the game will automatically attract players.

Preparing promotional materials, reaching the right audience, demonstrating gameplay clearly, collecting feedback, and building a community are still separate tasks that require effort.

And if production costs fall while more games enter the market, all that work may have to be done in an increasingly competitive environment.

For independent developers in particular, getting noticed could become an area that requires more resources alongside their development skills.

And we’re not just talking about advertising budgets.

Games recommended between friends, titles played by content creators, experiences discussed in communities, and loyal audiences built around particular genres all play a role in how games get discovered.

Showing people why a game is interesting could become just as important as actually finishing it.

Of course, that doesn’t mean only games with massive marketing budgets will succeed.

A small but original game can attract attention by meeting the needs of a particular group of players. Likewise, a major promotional campaign can’t guarantee success for a game that fails to keep players interested.

Game quality and visibility aren’t completely independent of each other. But they’re not the same thing, either.

So no matter how advanced production tools become, developers will still have one problem to solve: They need to show people not just that they’ve made a game, but why anyone might want to try it.

AI can create a prototype in minutes. But there’s still no comparable shortcut for attracting players’ attention and keeping them interested.

Can Google Really Replace Roblox?

Google’s Playground initiative has naturally drawn comparisons with Roblox. But evaluating the competition between the two platforms solely on the capabilities of their game creation tools would give us an incomplete picture.

For years, Roblox has brought game developers, players, groups of friends, and in-game economic systems together on the same platform.

It has an established ecosystem where developers can publish games, reach players, and earn revenue under certain conditions.

Google is trying to create a similar game creation and sharing experience with Playground, while also relying on its existing user network and partnership with Unity.

Google’s technical infrastructure and ability to reach large audiences through its various services could offer significant advantages.

But moving the player habits, friendships, and communities Roblox has built over the years to a new platform might take more than simply offering a more advanced game creation tool.

Roblox is already investing in generative AI, too.

Build’s early results show that the company wants to bring people who have never developed games before into its ecosystem.

So the competition between Google and Roblox involves much more than which company’s AI can generate a game faster.

The real difference may emerge after someone creates a game.

Can the developer invite friends? Will players come back? Can the project be supported and developed over time? And if the game becomes successful, can its creator earn money from it?

Playground’s existing community features attempt to address some of these questions.

However, Google hasn’t yet built an ecosystem comparable to Roblox in terms of user scale, creator economy, and years of established social infrastructure.

And it’s still unclear how successfully Google and Unity will deliver all the features they’ve announced.

Creating a game in eight minutes might be impressive. But replacing a platform where people have spent years hanging out with friends, building communities, and spending money could take a little longer.

At least for now, we haven’t seen an example of that being solved with a few lines of text, either.

So What’s Actually Changing in Gaming?

Game development has never been just about writing code.

Turning an idea into a working game requires programming, visual design, audio, storytelling, testing, and various technical systems to work together.

AI is beginning to make it possible to complete some of these tasks faster and at a lower cost.

Solo developers could gain access to technical capabilities that were previously beyond their reach. Small teams could experiment with new ideas more easily. Efforts to preserve older games, adapt them to new platforms, and connect different systems could also benefit from these tools.

On the other hand, the same technologies could make it much easier to publish large numbers of similar games created with minimal effort.

All of these outcomes are possible. But we don’t yet have enough long-term data to determine which effects will be more significant.

What we can already observe is that technical barriers are beginning to fall in certain areas of game development.

Tools like Google Playground and Roblox Build show that turning an idea into something playable no longer requires years of technical training in every case.

But it’s still unclear how many games created with these tools will last, become commercially sustainable, or genuinely interest players.

As more games are added to digital storefronts, developers may find themselves facing more competitors. Platform publishing policies, recommendation systems, and content moderation will play an important role in shaping that competition.

For many years, one of the biggest obstacles facing people who wanted to make games was the lack of necessary technical knowledge and resources.

As some of those barriers disappear, more people will be able to enter the same market.

But players don’t have unlimited time, attention, or money.

So in the coming years, which games can be made may matter just as much as which ones get discovered, actually played, and keep players coming back.

We don’t yet know whether AI will lead to better games or simply add more content to already crowded digital storefronts.

And those two outcomes aren’t mutually exclusive. We could see far more original indie games and thousands of nearly identical releases at the same time.

For now, the distance between making a game and reaching players doesn’t seem to have gotten much shorter. What’s changing is the set of tools needed to cover the first part of that distance.

Turning a game idea into reality could become easier than ever.

Figuring out why someone would actually want to play it, though, is still a problem developers have to solve.

And now there may be a whole lot more people trying to answer that same question.

Making Games Is Getting Easier. But Who Will Play Them? was originally published in Babel on Medium, where people are continuing the conversation by highlighting and responding to this story.

Judging by the videos circulating on social media, the boundaries between game engines have practically become a thing of the past. Taking a mechanic we like from one game and dropping it into another now looks like something that can be done in minutes. Presumably, all those technical problems that development teams have spent years dealing with have somehow solved themselves along the way.

Of course, there’s one small detail about these videos: They usually end while everything is still working.

How much work went on behind the scenes, which features remain unfinished, or whether the project would even launch on another computer is often unclear. And despite looking similar on the surface, these projects rely on very different technologies.

Some involve rewriting game mechanics, while others establish connections between separate pieces of software. The source code is publicly available for some projects, while others can only be judged by the footage their developers have shared.

To understand these differences, we need to look at how the things we’re seeing in these videos actually work.

Was Elden Ring’s Combat System Really Ported to Skyrim in Five Minutes?

A developer known as Funny-Bones has been working on a project that recreates Elden Ring’s combat mechanics using the Rust programming language.

This goes beyond simply adding new animations to Skyrim. Rules governing attack timing, stamina consumption, dodging, target lock-on, and how weapons make contact with enemies are also being recreated.

Normally, these calculations operate within Elden Ring’s own game engine. Funny-Bones is bringing the combat behaviors together in a more independent system, potentially making it easier to adapt those mechanics to other games like Skyrim.

The developer extracts some values directly from Elden Ring’s game files. However, parrying, certain weapon skills, and physics calculations are not yet fully implemented.

On social media, though, the claim was that this system had been brought into Skyrim in just five minutes.

Five minutes is certainly impressive. Especially if we ignore the minor detail of having to recreate Elden Ring’s combat system in Rust beforehand.

That timeframe may refer to the initial process of connecting the previously developed system to Skyrim. Building the underlying framework, extracting game data, and preparing the mechanics involve considerably more work.

The source code published by Funny-Bones makes it possible to examine how the independent combat simulation works. However, confirming the full scope of the Skyrim integration and the five-minute claim would require more detailed technical records.

In short, there’s genuine engineering work behind this project. Managing to make just five minutes of it the headline is a separate achievement.

Are Minecraft and Red Dead Redemption 2 Really Being Combined?

A project called RedCraft aims to bring Minecraft mechanics into the world of Red Dead Redemption 2. But unlike the Elden Ring example, it takes a different approach.

In SkyCraft, the project RedCraft builds upon, Minecraft and Skyrim run as two separate games on the same computer.

A connection between them allows Skyrim’s terrain and collision data to be transferred into Minecraft. Minecraft then calculates movement, blocks, and certain interaction rules before the results are reflected back into Skyrim’s world.

In other words, Minecraft’s entire system isn’t being rewritten inside Skyrim. Both games continue using their own underlying technology while exchanging data with each other.

From the outside, it looks like you’re playing a single game. The computer, meanwhile, has to run two games at once rather than celebrate their beautiful union.

And the developer’s documentation does mention high resource consumption, along with camera inconsistencies, coordinate problems in interior areas, and interaction errors involving certain quest objects.

Installation files and technical resources are available for RedCraft’s RDR2 adaptation. However, independent testing of which mechanics work reliably in this version remains limited.

Rather than two games being completely merged, what’s happening here is communication between systems that continue running separately. The results may look similar, but the method is not the same as the one Funny-Bones is using.

One project recreates combat mechanics. The other connects the existing systems of two games. Naturally, they also face different technical challenges.

We Can Now Destroy Roads and Buildings in GTA San Andreas

SA Destruction, a mod developed for GTA San Andreas, is another example of a project focused on changing the rules of an existing game.

The mod allows explosions to create craters in the ground and break apart certain structures. And it doesn’t just add visual effects. Because the game’s collision data is also modified, players can walk, drive, or shoot through the resulting holes.

A road that would normally remain intact after an explosion can now change the space players are able to move through. The destruction affects gameplay rather than remaining purely visual.

Anyone hoping to completely flatten Los Santos will have to wait a little longer, though.

The mod works on specific outdoor objects. Since some buildings don’t have modeled interiors, destroying a wall doesn’t necessarily reveal a room you can explore behind it. Apparently, not every building has an entire world waiting to be discovered beyond its walls.

The craters also aren’t permanently saved. Once an area reloads, the city forgets all about the destruction and returns to normal. Los Santos has a remarkably efficient recovery system.

The developer has published SA Destruction’s code and technical limitations. The method can be examined, although its performance and stability across different systems still require separate testing.

How Are Older Games Being Brought to New Platforms?

Another project shared on October 6 showed The Legend of Zelda: Twilight Princess running on a PlayStation 5.

Using the Dusklight framework, the project aims to adapt the game’s original behavior to a different system rather than running it through a conventional GameCube emulator.

Early tests reportedly achieved around 60 FPS at 4K resolution. However, displaying an image at 60 FPS doesn’t necessarily mean the physics and game logic are also being calculated at that rate. The developer says the original timing is being preserved.

This is not an official Nintendo release for PS5. The initial demonstrations also aren’t enough to prove that the game can be completed from start to finish without issues or that it will work on every standard console.

Another experiment attracted attention with videos showing GTA V running in the Chrome browser.

The project, shared through PlayGTA5.com, became inaccessible shortly afterward. Some sources claimed the game had been adapted to run directly in a browser environment, while others suggested it might have been using remote video streaming.

That distinction matters.

With a native browser port, the game’s processing happens on the user’s own computer. With cloud streaming, the game runs on another machine, and its video output is sent to the browser.

Both approaches can make it look like GTA V is being played in Chrome. But displaying a game in a browser and actually running it through the browser are two different things. Otherwise, watching a GTA V video on YouTube would also count as a pretty remarkable technical achievement.

Determining how the project actually works would require examining its underlying technology. For now, we have footage, but the method behind it remains unconfirmed.

Can AI Really Develop a Game in Just a Few Days?

Alongside the people bringing games to other platforms, there are also developers using AI to create entirely new games.

One example is Understory, developed by Danny Limanseta with assistance from Claude Sonnet 5.5. Built with Three.js, it’s a short action, puzzle, and RPG experience reportedly featuring five different biomes and five guardian bosses.

A playable version has been released, with the game designed to be completed in roughly 20–30 minutes. Limanseta says he gave the AI considerable freedom during development while continuing to guide the process himself.

However, the claim that all its content works as intended has not been verified through independent gameplay testing.

Around the same time, another claim suggested that Claude Opus 5.5 had been used to add Apex Legends mechanics to Elden Ring in just three days. However, no reliable code repository could be found that would allow the project’s technical details to be examined.

The whole idea of developing a game in a few days also depends on exactly when you decide to stop the clock.

When the first playable prototype works? When every level is finished? Or when the bugs have been fixed, performance issues resolved, and the game runs properly on different computers?

If we only count the time needed to produce the first working version, development times really can become considerably shorter. Of course, that means we also have to stop counting all the remaining testing, optimization, content creation, and maintenance as part of game development.

Unfortunately, the bugs that appear after release haven’t been informed of this new definition.

Research into the impact of AI on software development hasn’t reached a common conclusion either.

In a controlled study conducted by METR in 2025, experienced developers took an average of 19% longer to complete their tasks when using AI. A February 2026 update estimated potential speed improvements for some participants, although the uncertainty made firm conclusions difficult.

Another study involving Google engineers estimated a time saving of approximately 21% for a specific task.

These results come from different conditions, making direct comparisons difficult. The tools being used, the type of work, and the developer’s experience can all affect the outcome.

Yet these details rarely receive much attention in social media stories about games developed in a matter of days. Apparently, talking about a three-day project is easier than explaining exactly where those three days went.

Is Combining Games Actually a New Idea?

Watching today’s videos, you might get the impression that developers have only recently discovered that game engines have boundaries worth pushing. But players have been trying to transform existing games beyond recognition for a very long time.

Back in the 1990s, total conversion mods for Doom could replace maps, enemies, and weapons to create what were practically entirely different games.

Counter-Strike began as a mod for Half-Life. Garry’s Mod turned the Source engine’s physics and object systems into an environment where players could create their own experiments.

Over time, people began recreating not just the content of games, but the engines that powered them.

OpenMW aims to recreate Morrowind’s engine functionality through separate open-source software. Zelda 64: Recompiled uses different techniques to transform older game code into a form that can run on modern platforms.

The developers of Zelda 64: Recompiled have also specifically stated that they do not use generative AI.

So apparently, people were capable of coming up with the idea of running games on other systems even before AI arrived.

Today, the number of available tools and the ways they’re used are changing. Open-source projects, mechanics designed to operate more independently of individual game engines, and AI-assisted coding systems can make certain ideas easier for smaller teams to experiment with.

But being able to apply established techniques using newer tools doesn’t automatically make every resulting project an entirely new invention.

Some projects genuinely explore different approaches. Others apply methods that have existed for years to newer games. Distinguishing between them requires understanding what’s happening behind the scenes rather than judging how impressive the video looks.

But Are These Projects Actually Usable?

It’s easy for a working feature in a viral video to be mistaken for a finished product.

A combat move brought into Skyrim might cause problems when used against different enemies. A system connecting two games might break certain quests. An older game running on a new platform could encounter unexpected bugs in later sections.

It’s hardly surprising that these problems don’t appear in promotional videos. Nobody introducing a new mod is particularly eager to lead with the crash reports. Five seconds of successful gameplay footage tends to be a little easier to market than forty minutes of debugging.

These limitations don’t make the projects worthless. But there’s an important difference between a successful demonstration and software that has been completed and thoroughly tested.

Copyright and terms of use introduce another set of complications.

Rockstar’s modding policies place restrictions on projects that combine content from different Rockstar games or bring its titles to platforms where they haven’t been officially released.

Some developers distribute only the code they’ve created themselves and require players to provide their own original game files. This allows them to avoid directly distributing copyrighted game content.

Even so, this approach doesn’t guarantee legal protection in every situation.

Getting a project to work technically, making it easy for players to install, being able to distribute it, and supporting it over the long term all involve different requirements. The first working version satisfies only one part of that process.

So Are the Boundaries Between Games Really Disappearing?

Recent projects show that a variety of approaches are being used to modify games and make them run on different systems.

Funny-Bones is recreating Elden Ring’s combat behaviors. RedCraft is experimenting with connections between systems running in separate games. SA Destruction is changing the physical environment of GTA San Andreas. Projects like Dusklight, meanwhile, focus on making older games work on other platforms.

Some of these projects rely on techniques that have been developed over decades. Others use modern software tools and AI assistance to make certain stages of experimentation easier.

One perspective is that these developments are allowing smaller teams to bring more complex ideas to life. Another points out that many of the techniques are already well established and that the results shown in viral videos often fail to reflect their technical limitations.

Existing projects provide examples relevant to both interpretations. Some are supported by source code that can be independently examined, while the scope of others can still only be judged from developer demonstrations.

Seeing combat systems ported in five minutes, games built in a few days, and videos of GTA running in a browser is no longer particularly difficult. What remains far less visible is the development work behind those results, the compatibility issues, and the unfinished features.

The technical boundaries between game engines still exist. The methods used to work around them continue to evolve.

Which of the experiments we’re seeing in a few seconds of footage today will become projects supported for years, and which will remain experimental, is still unknown. For now, there’s certainly no shortage of videos showing what games are capable of doing. Figuring out how much of it they can actually do without running into problems might take a little longer.

That is exactly what Weasel Pit, the developer of Iron Pilgrim, said happened to them at the end of September.

In a post published on September 29, the developer shared footage from their own game alongside a second video they said had been sent to them and generated with AI. According to the claim, the sender had found Iron Pilgrim on social media, fed the footage into an AI tool, and created something resembling the game.

The post places Iron Pilgrim footage directly beside the claim that an AI-made clone already exists, which is the visual premise the threat depends on. (Image Source: Weasel Pit on X / InTheIndie)

The demand was straightforward: Send around $200 worth of Steam content, or the supposed “clone” would be released on Steam before the developer’s own game. The message also appeared to give them 24 hours to comply. Weasel Pit said they would not pay.

The email screenshot shows the demand and the 24-hour framing that turned the footage into leverage. It is also why the relatively small $200 figure became central to doubts about whether a release-ready clone existed. (Image Source: Weasel Pit on X / InTheIndie)

But as the story spread, another question emerged.

Had an actual game really been copied?

What They Showed May Not Be a Game at All, Just a Video

In the Iron Pilgrim case, the material shown publicly is not a playable build or demo. What exists is a short video sent to the developer.

Iron Pilgrim’s official Steam art centers the weaponized train that defines the real project’s presentation. A matching-looking video can imitate that surface quickly, but it does not by itself demonstrate the systems of a functioning game. (Image Source: Steam / Weasel Pit)

That distinction matters. AI tools can now generate a few seconds of footage that looks like gameplay. But the existence of a video like that does not prove there is a complete game behind it with functioning mechanics, menus, controls, save systems, and everything else required to make it playable.

Much of the discussion on Reddit focused on exactly this point. Some developers argued that the footage could simply be an AI-generated video rather than evidence of a genuine clone. So far, no independent evidence has surfaced showing that the person behind the message actually developed a working copy of Iron Pilgrim.

Publishing a game on Steam is not as simple as uploading a file either. Valve charges a $100 Steam Direct fee for each product, and developers also have to go through Steamworks registration and review procedures.

AUTOMATON illustrated the incident as an online threat rather than as evidence of a finished commercial game. That distinction fits the gap between generating convincing footage and completing Steam’s actual release process. (Image Source: AUTOMATON)

There is therefore a significant difference between generating a few seconds of similar-looking footage and producing a working game that is ready to be released.

The $200 demand was another reason people on Reddit became suspicious. If someone really could copy a functioning game that quickly and believed they could sell it on Steam, why give all of that up for just $200?

There may be a simple answer: There may not need to be a release-ready clone at all.

For the threat to work, the developer only needs to believe there is one.

Then Other Developers Started Receiving the Same Message

If the Iron Pilgrim case had remained an isolated incident, there would have been much more reason to question whether the story was genuine.

But shortly afterward, Psicodart Games said it had received a similar email from the same sender.

At the time, the studio was developing a game called Circusar. When the shared emails were compared, the wording was almost identical to the message sent to the Iron Pilgrim developer. The main thing that changed was the name of the game. Once again, the developer was asked for $200 and told that an AI-made copy would be released first if they refused to pay.

This is official Circusar key art from Psicodart Games, the second developer named in the report. Its involvement matters because the same $200 template was no longer tied only to Iron Pilgrim. (Image Source: Psicodart Games Press Kit)

On October 2, AUTOMATON reported that multiple developers had received similar threats.

Circusar’s Steam art provides the public-facing material a scammer could discover before release. In the reported emails, changing the game name while keeping the threat structure almost identical is what made the pattern more significant. (Image Source: Steam / Psicodart Games)

This does not prove that the clone shown in the Iron Pilgrim case was a genuinely working game. But it does make it much harder to dismiss the $200 threat as a one-off message aimed at a single developer.

The method appears fairly simple: Find a game still in development, create AI-generated footage resembling it, then send the developer a message saying, “I’ll release it before you do.”

If the real goal is not to release a game but to get money, trying the same tactic on a large number of small developers starts to make more sense.

But There Was Another Problem With the Original Story

Another detail later emerged that made the Iron Pilgrim case more complicated.

When Reddit users looked through the history of Weasel Pit’s account, they noticed that only a few weeks earlier, the developer had posted a remarkably similar threat toward another indie developer.

That developer had shared a game concept built around finding a single four-leaf clover among 35,000 clovers. The Reddit account used by Weasel Pit criticized the game and said that if it was released, they would make a free copy and put it on Steam.

The Reddit screenshot shows the earlier threat to make a free copy and release it on Steam. That visible similarity is why users questioned the later Iron Pilgrim story without proving that the later email itself was fabricated. (Image Source: Reddit / InTheIndie)

In other words, the same account that would later say, “Someone is threatening to copy my game and release it on Steam,” had recently made a similar threat against another developer.

The later Reddit post frames the same account as the target of the kind of Steam-copy threat it had previously made. Seeing the two posts side by side explains why the story’s credibility became a separate issue from the broader scam reports. (Image Source: Reddit / InTheIndie)

Once this surfaced, some users began to suspect that the Iron Pilgrim story might have been a marketing stunt designed to draw attention to the game. InTheIndie also published an article interpreting the incident from that angle.

Weasel Pit later acknowledged that the earlier comment had been wrong and said they had apologized to the other developer. They also maintained that they had not fabricated the extortion message sent to them.

InTheIndie later added the developer’s reply, which is important because the article did not end with the accusation alone. The response sits alongside reports that other developers received similarly formatted messages. (Image Source: Weasel Pit reply / InTheIndie)

Two separate issues need to be distinguished here.

It can be verified that Weasel Pit had previously made a similar threat toward another developer. Naturally, that raises questions about the story they later told.

But concluding from that alone that Weasel Pit fabricated the entire extortion attempt would go beyond the available evidence. In the days that followed, other developers said they had received messages using the same format, and some said the messages came from the same sender.

That leaves us with a strange picture: The developer who first brought attention to the story has a questionable history, but messages resembling the method they described really were sent to other developers as well.

The Fear of Games Being Copied With AI Is Not Entirely Imaginary Either

In these extortion cases, no one has yet demonstrated that fully working games were actually created. Still, it is difficult to say that developers’ fears of being rapidly imitated with AI have no basis at all.

A separate example recently involved the developers of 204Snake.

They said that a few weeks after putting up the Steam page for their unreleased puzzle game, they discovered another game online using the exact same name. According to comparisons shared by the developers, the other game used the same core concept and control scheme, featured similar level layouts, and attempted to imitate the original game’s visual style.

This 204Snake screenshot shows the grid-based puzzle structure and numbered tiles that define the original game’s look. The developers said a separate AI-made version copied the premise, controls, similar layouts, and visual approach without making a $200 demand. (Image Source: Thinky Games / 204Snake)

The 204Snake team said the copy had been made using AI.

But there is one major difference between that incident and the Iron Pilgrim case.

No one asked the developer for money.

The official Steam header makes the original project’s public visual identity easy to recognize online. That public visibility is exactly what the 204Snake developers said was followed by discovery of a same-name imitation elsewhere. (Image Source: Steam / Robotonic)

So it would be misleading to present 204Snake as part of the same extortion campaign. It is better understood as a separate example of how quickly a developer’s publicly shared ideas and visual approach can now be imitated with the help of AI tools.

What Is Behind the $200 Threat?

Based on the information currently available, it is impossible to say for certain whether the people behind these messages are actually developing games, running an automated scam, or simply trying to frighten developers into paying.

The existence of working clones has not been verified.

The AI-generated videos that were shown are real.

It can also be confirmed that similar messages were sent to multiple developers.

And the fact that the Iron Pilgrim developer had previously directed a similar threat at someone else adds another layer of doubt to the story.

Once genuine examples of games being imitated with AI are added to the picture, the situation cannot simply be reduced to “AI can now copy entire games in a few hours.”

What may be happening is much simpler.

Generate a few seconds of footage resembling a game someone has spent months developing, send it to them, and claim the entire game is already finished.

The person receiving the threat does not need to know whether that is true.

They only need to be afraid that it might be.

What remains unknown is whether there are actually functioning games behind those videos. In the threats sent to Iron Pilgrim and the other developers, no playable clone has yet surfaced that would confirm that claim.

For now, it would be premature to describe these incidents as proof that people are using AI to fully recreate entire games.

But turning a developer’s publicly shared work into a convincing imitation and then using it as leverage against them is no longer purely theoretical.

—

The number of games on Steam disclosing some form of AI use is also rising quickly. Small teams are using AI tools to take on technical work that would previously have been out of reach. At the same time, developers are worried about job security, the value of creative labor, and stores being flooded with low-effort games.

So this is hardly a one-sided picture.

Is all of this actually pushing the gaming industry in a worse direction? Or are we simply seeing a new production model emerge as the money, time, and team size required to make games begin to shrink?

And more importantly: If making games gets easier, does making good games get easier too?

Game Developers Are Already Using AI

More than 2,300 industry professionals took part in the 2026 GDC State of the Game Industry survey. Thirty-six percent said they use generative AI in their work.

But “using AI” covers a very wide range of things.

Among those using AI, 81% said they use it for research and ideation. Forty-seven percent use it for everyday tasks, another 47% for coding assistance, and 35% for prototyping. The figure drops to 19% for asset generation and just 5% for generative features that players directly interact with.

So when we say 36% of developers use AI, that does not mean one out of every three games has its visuals or levels generated by AI. A programmer using AI to help write code counts, and so does a designer using it to brainstorm ideas.

Another figure from the same survey makes the picture more complicated.

Fifty-two percent of respondents said they believe generative AI is having a negative effect on the game industry. Only 7% viewed its impact positively. Negative sentiment was even higher among visual and technical artists, designers, narrative developers, and programmers.

In other words, a significant share of people in the industry are using AI while still feeling uneasy about where the technology may be taking their field.

There is also an important caveat to the GDC numbers. The survey methodology changed in 2026, and more than half of respondents were based in the US. So lining the results up with previous years and claiming that “anti-AI sentiment increased by X amount every year” would be misleading.

Are 35% of Steam Games Really Made With AI?

One of the most striking numbers in the AI debate comes from Steam.

An analysis published in August 2026 examined roughly 17,000 Steam store pages from a pool of more than 62,000 games released after July 2023.

Among 4,766 games released between January and August 2026, 1,670 included some kind of AI disclosure.

That is roughly 35%.

The figure was 12.44% in the 2024 sample and 21.58% in 2025.

But that does not mean “35% of Steam games are made by AI.”

In the dataset, a game that used AI for a few lines of dialogue and a game that used it across a large share of its visuals or music both count as a single AI disclosure. The amount of AI involvement is not measured. Only 4.91% of the 2026 games examined disclosed AI-written code.

Steam’s disclosure system also does not capture every AI tool developers may be using behind the scenes.

Valve clarified its policy further in early 2026. AI-generated content that players actually consume, such as artwork, audio, narrative, or localization, still falls under disclosure rules. Some coding and productivity tools used only to speed up development do not fall under the same requirement.

So the 35% figure matters, but its meaning is limited: Games openly disclosing AI use on Steam are no longer a tiny minority.

What that number cannot tell us is how much of any given game was actually made with AI.

Steam Was Already Flooded With Games Before AI

Another major concern is that lower production costs will fill stores with thousands of low-effort games.

It is difficult to say whether that is already happening because Steam’s release volume began exploding long before generative AI became mainstream.

Steam saw 436 new releases in 2013. By 2016, that number had reached 4,647. There were 9,634 releases in 2020, 18,452 in 2024, and 21,270 in 2025.

AI was not responsible for that growth.

Accessible game engines, asset stores, digital distribution, and systems like Steam Direct had already been lowering the financial and technical barriers to game development for years.

AI is another tool pushing in the same direction.

If writing code, creating placeholder art, handling localization, building prototypes, or completing certain production tasks becomes cheaper and faster, it is hardly surprising that more people will be able to make games.

The problem is that players do not suddenly get more hours in the day.

SteamDB data from late 2025 showed that out of more than 19,000 games listed that year, 9,327 had fewer than 10 user reviews. Another 2,229 had no reviews at all. That does not tell us why those games went unnoticed or whether they were any good, but it does show how difficult discoverability on Steam has become.

AI did not create that overcrowding on its own. But if it lowers production costs even further, it could add even more games to an already packed marketplace.

For indie developers, the biggest obstacle may gradually shift from “Can I actually make this game?” to “How do I get anyone to notice it among all these other games?”

One Developer and Two AI Agents Ported GoldenEye to PC

The 2026 GoldenEye 007 PC Port project shows a very different side of AI.

Its developer used two AI coding agents to work on a large, decompiled C codebase. A local Qwen model and Claude Code were used, with a single human developer managing both agents part-time.

After 36 days, all 21 missions had been tested end to end. The result was a 60 FPS PC version running on Windows, Linux, and Steam Deck.

But saying “AI ported GoldenEye to PC in 36 days” leaves out a lot.

The agents were not starting from nothing. Years of human work had already gone into the GoldenEye decompilation project. Infrastructure from the Perfect Dark PC port was also used.

The human developer was not simply sitting back and waiting for the result either.

Building the project, running the game, playtesting it, checking visuals, verifying bugs, and integrating everything still required human involvement. One of the main bottlenecks was that the AI agents could not actually play the game and judge whether what appeared on screen was correct.

That is what makes the GoldenEye example interesting.

AI did not remove the human from the process. It expanded the amount of technical work one person could handle in the same amount of time.

That could matter a lot for old game ports, recompilation projects, mods, and preservation work. Some projects that would normally demand too much time to justify may become much more realistic.

Generating Code and Making a Good Game Are Not the Same Problem

AI can write code, generate images, suggest ideas, and build prototypes.

But does putting all of those things together actually produce a good game?

Keywords Studios tried to explore that question through an experiment called Project AVA. More than 20 experienced developers spent six months trying to build a commercially viable game using generative AI tools.

AI was useful in some areas. It helped produce concept art variations, draft early dialogue, generate story ideas, and assist experienced programmers who were getting used to Unreal Engine 4.

The team still struggled to generate a meaningful share of the planned code and assets at the required quality level. Human specialists also had to step in on the narrative side.

That does not mean AI is useless in game development. The GDC data already shows where it is helping most: brainstorming, coding assistance, and prototyping.

But producing more things faster and deciding which of those things should actually stay in the game are two different jobs.

A combat mechanic can still be boring. A level can still drag. An economy can still break. A visual style can still become inconsistent. A great-looking idea can still add nothing to the finished game.

All of those things still have to be tested and judged.

And when production speeds up, those decisions do not disappear. In some cases, you simply end up with more options to evaluate.

What Helps a Small Team Can Become a Cost-Cutting Tool Somewhere Else

The effect of AI on human labor is the most tense part of this debate.

Twenty-eight percent of respondents in the 2026 GDC survey said they had been laid off within the previous two years. Among US respondents, the figure was 33%. The employment crisis in the gaming industry is very real.

But the same survey also asked workers what reasons their companies had given for layoffs.

The most common answer was restructuring at 43%. Budget cuts and market conditions were both at 38%, project cancellations at 32%, and strategic shifts at 31%.

Automation or AI implementation appeared in 6% of responses. Participants were allowed to select more than one reason.

That 6% figure does not mean AI has nothing to do with job losses. In some companies, AI and automation were explicitly cited as part of the reason.

But the available evidence also does not support blaming the industry’s recent wave of layoffs primarily on AI.

The harder question is what happens next.

If AI allows a team to complete a certain task 20% or 30% faster, what does the company do with that extra capacity?

More iteration?

A bigger game?

A shorter development cycle?

Or a smaller team?

For a solo developer, the same tool might make it possible to finally finish a project that used to be too ambitious. For a large company under pressure to cut costs, the same productivity gain could become a reason to recalculate how many people are needed.

The technology itself does not make that decision.

Do Players Actually Care Whether AI Was Used?

A common argument goes something like this: “If the game is good, who cares how it was made?”

Steam reviews suggest not everyone sees it that way.

A study published in August 2026 analyzed more than 508,000 English-language Steam reviews. Games with AI disclosures had a recommendation rate of 68.4%. The procedural-generation comparison group was at 86.3%.

Researchers also manually reviewed 600 comments. Some players were not judging AI use only by the quality of the resulting art or audio. In some reviews, the use of AI was treated as a signal of how much effort or care developers had put into the game.

That does not mean “use AI and players will hate your game.”

The study was observational. The two groups were not identical in genre, price, target audience, or market positioning. It also could not account for games using AI without disclosure.

Real-world examples show that reactions can vary dramatically from one game to another.

ARC Raiders disclosed that AI-based tools could be used during content creation and still attracted a large audience and broadly positive reviews.

Jurassic World Evolution 3 went in a different direction. After players reacted negatively to the discovery that some scientist portraits had been created with generative AI, Frontier announced that the portraits would be removed.

The Finals uses a mix of human voice actors and AI text-to-speech. Embark has said TTS is useful for contextual callouts that need to be produced quickly, while more dramatic moments still rely on human performances. That decision also sparked a separate debate around voice acting and creative labor.

There is no single unified player attitude toward AI. The type of content being generated, the way the technology is used, and the quality of the finished game can all affect the reaction.

What Is Becoming Cheaper in Game Development Is Changing

The evidence we have right now does not support the claim that “AI is killing the gaming industry.”

It also does not support the idea that “AI is here, so now anyone can make a AAA game alone.”

The clearest change is happening in specific parts of production.

Drafting code, brainstorming, prototyping, generating certain visuals, working on localization, or handling repetitive tasks across a large codebase can all become faster.

That is not a minor shift.

Game development has always been constrained by money, manpower, and technical capacity. As some of those limits come down, more people and more projects can enter the market.

But Project AVA still needed human specialists to reach the required quality. In the GoldenEye port, AI agents could generate code quickly, but human verification became the bottleneck. Steam was already a marketplace receiving tens of thousands of new games every year.

As production capacity grows, choosing, cutting, testing, and getting noticed do not disappear.

They may become even more important.

So going forward, asking “Was AI used in this game?” probably will not tell us enough on its own.

We will also have to ask where it was used and what happened to the productivity gains it created.

One studio might use them to iterate more with the same team. A small developer might finally finish a project that had been out of reach for years. An old game might become economically viable to port to PC. Another company might choose to produce the same amount of work with fewer employees. Thousands of games that otherwise would never have existed may also appear on storefronts.

The data we have in 2026 does not yet tell us which of those outcomes will dominate.

But one change is already becoming visible:

As making games gets cheaper, deciding what is actually worth making becomes more important.

But all of these options lead to the same question: What should players understand about the full game by the time they close the demo?

A demo is not just a small piece cut out of the game. Players are trying to understand how the controls feel, whether the pacing suits them, how deep the systems become, whether they like the atmosphere, how the game runs on their computer, and perhaps most importantly, whether they want to spend dozens of hours with it. For developers, that same demo can also generate wishlists, visibility, sales, festival exposure, interest from content creators, feedback, and technical bug reports.

Seen this way, the question becomes less about how much content the demo provides and more about whether the content it does provide is enough for players to make a decision.

Starting From the Beginning

Using the opening section of the full game as the demo is a natural choice. Players are experiencing the game they would actually be buying. The tutorial, the way the story is introduced, the pace of progression, and the order in which systems are unlocked all remain intact. If save transfer is supported, players who buy the full game also avoid having to replay the same section.

Resident Evil 4 Remake’s Chainsaw Demo is a clean example of this approach. It uses the opening of the game, but adjusts it so that it can work as a self-contained demo.

The main question here is how well the game’s first hour represents everything that comes after it.

The answer is not the same for every game.

Role-playing games, strategy games, management games, and system-heavy action games often cannot put most of their mechanics in front of the player within the first few minutes. Instead of overwhelming a newcomer with dozens of systems, complexity is introduced gradually. Enemy variety is limited, abilities are unlocked one by one, and crafting or economic systems are brought in step by step.

That can be good tutorial design. For a demo, however, it creates a different problem. The easiest part of the game to understand may not yet show the things that make it interesting later on.

In a factory-building game, the section where you set up your first machines and the point twenty hours later where you are managing interconnected production lines may come from the same systems, but they do not have the same level of complexity. A character action game may also begin with only a few basic moves before expanding into completely different combinations as new abilities are unlocked.

For that reason, the opening section can be faithful to the full game while still failing to show its full character.

A Section From Later in the Game

The alternative is to place the player directly at a point where more of the game’s systems have already opened up.

More weapons, abilities, enemy types, and system interactions can be shown immediately. Players do not have to wait several hours to see where the mechanics eventually lead.

But what makes a later section meaningful is not simply the number of features available in it. By that point, the player has also learned how those systems work.

Giving a first-time player eight abilities, a crafting system, several resource meters, and multiple characters may show how extensive the game is. But for someone who does not yet know when or why any of those things should be used, the same screen may simply look overwhelming.

Progression systems perform two roles here. They give players new options, but they also give them time to learn how to use those options. If you remove that process and only present the end result, a system that feels understandable in the full game may feel unnecessarily complicated in the demo.

The problem becomes even more obvious in story-heavy games. A later mission may work perfectly from a mechanical perspective, but the player may have no idea who the characters are, why they are there, or why the events taking place matter. Explaining that context can take too long, while leaving it unexplained can weaken the impact of the scene.

Choosing the most impressive section of the game is therefore not automatically the solution. Some sections work precisely because the player understands how they got there.

A Separate Section Made for the Demo

Instead of taking a specific section from the full game, creating a separate experience specifically for the demo gives developers far more control.

Story spoilers can be removed, core mechanics can be shown within a shorter period, and the pacing can be designed around the length of the demo. Tools that normally appear much later in the game can be introduced early. It is even possible to create an entirely new section that does not exist in the full game at all.

That freedom also comes with risks.

It is possible to take the best enemies, most interesting mechanics, and strongest moments from a twenty-hour game and compress them into half an hour. Naturally, such a demo may feel more intense, more varied, and better paced. The full game, however, does not necessarily maintain that same level of intensity for twenty hours.

Still, the fact that a custom demo differs from the full game does not automatically make it less representative.

The Stanley Parable is an interesting example. One of the game’s most important qualities is its ability to surprise the player, so showing a direct section from the main game could have used up some of those surprises before the player even bought it. Instead, the developers created a separate demo that followed the logic of the full game without using its actual content. Players did not see the same rooms or events, but they still got a fairly accurate idea of how the game communicates with them.

Resident Evil Village’s Maiden is a more limited example. You do not control the main character, there is no combat system, and the story does not take place at the same point as the full game. Even so, the environment, atmosphere, visual quality, and horror tone can all be experienced directly.

A demo may not need to prove every system in the game. The problem begins when the things it leaves out create the wrong expectations about the full experience.

Why Should One Section Have to Do Everything?

If the opening is better at showing how the game teaches the player, while a later section is better at showing where the mechanics eventually lead, there is no reason both jobs have to be forced into the same piece of content.

Final Fantasy XVI’s demo solves this with two separate sections. First, players experience the real opening of the game, and their save can be carried over into the full version. After that, a separate combat section unlocks with a broader set of abilities that are not yet available at the beginning.

Players first see how the game starts, then get to experience how its combat system expands later on.

Final Fantasy VII Rebirth follows a similar idea in a slightly different way. Alongside a section that establishes story context, it includes another area that demonstrates broader exploration and combat options. Establishing the narrative and showcasing the systems are not left to a single section.

This kind of structure is not necessary for every game. Preparing separate sections requires more design, testing, and maintenance work. But if there is a major difference between the game’s opening hours and the level of mechanical depth it reaches later, it removes the need to force everything into a single section.

How Long Should It Be?

One of the easiest questions to ask about a demo is how long it should last. But there is no single number that works for every game.

In games where movement, basic combat, or the core gameplay loop can be understood within a few minutes, 15 to 30 minutes may be enough. In a role-playing game, however, it can take much longer for the characters, world, and progression systems to settle into place.

Octopath Traveler II offers a demo that can be played for up to three hours, with progress carrying over into the full game. Dragon Quest XI S pushes that to roughly ten hours. The fact that the same concept can work across such different lengths shows that duration alone does not determine whether a demo is good or bad.

As the demo gets longer, another issue becomes more important: repetition.

Replaying a ten-minute section after buying the full game may not be a major problem. Starting over after several hours of progress is different. That is why save transfer becomes much more valuable for longer demos.

Where the demo ends is just as important as how long it lasts.

If a short section allows the player to complete the entire basic gameplay loop and provides a natural stopping point, even half an hour may feel satisfying. A longer role-playing demo, on the other hand, may work better if it ends just as new systems begin to open up, leaving the player with more still to discover.

The ideal stopping point is not a fixed number of minutes. It may be the moment when the player has just started to understand the system but has not yet exhausted what can be done with it.

The intended length and the amount of time people actually spend playing are also not the same thing. Even a short demo will not be completed by everyone. That means saving the game’s main appeal for the final minutes may result in many players never seeing it at all.

The Role of the Demo Has Changed

In the past, the purpose of a demo was easier to describe: play a little before deciding whether to buy the game.

During the shareware era, that idea extended all the way to the distribution model itself. The first episodes of games such as Doom were distributed widely, and players who wanted the rest could buy the full product. Demo discs and later digital trial versions served the same basic need in different forms.

Today, players have more options. Refund systems on stores such as Steam allow people to try a game after purchasing it, provided certain conditions are met. Yet demos have not disappeared, because they no longer serve only as a way to test a game before buying it.

Being discovered during a festival, generating wishlists, giving content creators a playable build, collecting early feedback, or identifying technical problems can all be part of the same purpose.

That makes it difficult to judge the success of a demo only by how many people later buy the game.

In the past, data was published suggesting that demos could significantly reduce sales. But that data did not establish a direct causal relationship between releasing a demo and lower sales. Major brands that skipped demos, marketing budgets, and the circumstances surrounding individual games could all influence the results. More recent festival data suggests that demos can create strong momentum in discovery and wishlists, but even there, it is difficult to completely separate the impact of festival visibility, publisher attention, and the demo itself.

More than anything, a demo gives the player more information.

Sometimes that information strengthens their desire to buy the game. Sometimes it helps them realize, before spending any money, that the game simply is not for them. In the second case, the sale may not happen, but from the player’s perspective, the demo may still have done its job.

The Right Solution Depends on the Game

The reason no single demo format works for every game is that players are not evaluating the same things in every game.

In a platformer, movement can be understood within a few minutes. In a horror game, atmosphere and the feeling of vulnerability may matter from the very beginning. In a puzzle game, a separate section that teaches the game’s problem-solving language without using up the solutions from the main game may be more useful.

In strategy and management games, a controlled scenario may explain the rules more clearly than dropping the player into a massive late-game map.

Factorio’s demo is a good example. Instead of simply shrinking down the free-play experience, it uses missions that teach core systems such as mining, conveyor belts, furnaces, automation, and production chains. Before showing players how impressive a gigantic factory can become, it gives them a chance to find out whether they actually enjoy building a small production line.

For a long role-playing game, a multi-hour opening with save transfer may make more sense. A system-heavy action game may benefit from combining the real opening with an advanced combat section. If the way the story is introduced is a fundamental part of the experience, the opening itself may already be the most appropriate demo.

So before asking, “Should we use the opening, choose a mission from the middle, or create a separate section?” another question needs to be answered first: What does the player need to experience in order to understand this game?

The controls?

How much deeper the combat becomes later on?

The atmosphere?

The production loop?

The tone of the story?

Technical performance?

When the answer changes, the right demo format changes with it. For the same reason, an opening section, a later section, a custom scenario, or a hybrid structure is not inherently superior to the others.

Using the opening of the full game can sometimes be the best choice. But if the beginning makes the game look simpler than it really is, being chronologically faithful is not enough. A later section may reveal the potential of the mechanics, but if the player does not know how to use them, it may communicate very little. A custom-made demo can solve these problems in a more controlled way, but then it risks presenting an experience that is more intense than the full game itself.

The job of a demo is not to show everything the game contains.

It is to give players a sufficiently accurate understanding of what they will encounter in the game they are buying.

Sometimes the opening is enough. Sometimes it takes several hours. And sometimes a completely separate section that does not exist in the full game may explain it better.

A demo does not have to be a miniature version of the full game.

It just needs to avoid making the full game look like something it is not.

WHAT SHOULD A GAME DEMO BE LIKE? was originally published in Babel on Medium, where people are continuing the conversation by highlighting and responding to this story.

Not every example here is literally more famous than the game it came from. You can’t really reduce something like Final Fantasy VII, Skyrim, or Modern Warfare 2 to a single scene. But every mission and moment below became recognizable enough to be talked about separately from the rest of its game.

Call of Duty: Advanced Warfare | Press F to Pay Respects

At Will Irons’ funeral, you walk up to the coffin and the PC version asks you to press F to pay respects.

The scene lasts only a few seconds. Its life on the internet lasted much longer than that. “F” gradually became a quick response to loss across Twitch chats, comment sections, and social media, sometimes sincerely and sometimes as a joke.

It’s completely possible to know what “F” means without ever playing Advanced Warfare.

Call of Duty 4: Modern Warfare | All Ghillied Up

Years before the main story, Captain Price and Captain MacMillan move through the abandoned outskirts of Pripyat in full ghillie suits, trying to stay hidden from Russian patrols.

The mission slows Call of Duty down in a way the series rarely did at the time. At one point, you’re lying completely still in the grass while soldiers and armored vehicles pass just a few feet away, with MacMillan telling you when to move and when to stay down.

Call of Duty 4 has plenty of memorable missions, but All Ghillied Up is still one of the first names that comes up when its campaign is discussed. The Pripyat setting, MacMillan, and that crawl through the grass became some of the most recognizable images from the original Modern Warfare.

Zero Wing | “All Your Base Are Belong to Us”

“All your base are belong to us.”

The line comes from the famously broken English translation in the 1991 European Mega Drive version of Zero Wing. In the early 2000s, forums, images, and Flash videos turned it into one of the internet’s earliest major memes.

The fact that plenty of people know the quote without knowing Zero Wing says enough about how far those few opening lines travelled.

Star Fox 64 | “Do a Barrel Roll”

In Star Fox 64, Peppy’s “Do a barrel roll!” is really just a simple gameplay instruction.

The internet never let it go. The line spread through forums, reaction videos, and memes, and Google even turned it into an Easter egg on its search page.

Star Fox 64 is still a well-known game, but those four words travelled far beyond the series itself.

The Elder Scrolls V: Skyrim | The Opening Cart Ride

The screen fades in, your hands are tied, and you’re sitting in the back of a cart. Then Ralof turns to you and says:

“Hey, you. You’re finally awake.”

Those few seconds from Skyrim’s opening became the punchline to countless bait-and-switch videos. A clip can start as something completely unrelated, suddenly cut to the cart ride, and the joke works without needing any explanation.

Skyrim is an RPG you can spend hundreds of hours in. Yet one of its most recognizable moments happens before you’ve even created your character.

GTA: San Andreas | Wrong Side of the Tracks

Big Smoke is riding on the back of the bike while CJ follows the train. The goal is to take out the Vagos members standing on top of it.

Fail, and Big Smoke delivers the line:

“All we had to do was follow the damn train, CJ!”

The mission has long been one of San Andreas’ most talked-about sections. The failure quote went even further and kept circulating as a meme long after the mission itself.

GTA: Vice City | Demolition Man

You control a tiny RC helicopter and have to carry four bombs to different floors of a construction site before the timer runs out.

The mission became notorious largely because of the helicopter controls, especially in the original PC version. You’re trying to steer through a cramped building, place each bomb in the right spot, and keep an eye on the clock at the same time.

Vice City has plenty of missions people remember fondly. Demolition Man is usually remembered for very different reasons, and its reputation has stuck around for decades.

Driver | The Parking Garage Test

Driver doesn’t immediately let you loose in its open world. First, you have to complete a list of driving moves in a closed parking garage, including burn-outs, reverse 180s, and slaloms.

You have 60 seconds.

And all of this happens before your first proper mission. That’s where Driver’s reputation for putting a wall in front of the player before the game has even really started comes from.

There’s no reliable figure showing how many players failed to get past it, but the parking garage remains one of the most frequently discussed parts of Driver.

Battletoads | Turbo Tunnel

In its third stage, Battletoads puts you on a speeder bike and starts throwing obstacles at you faster and faster.

Turbo Tunnel carries a huge part of Battletoads’ reputation for brutal difficulty on its own. Even players who never saw the later stages often know this one.

There’s a reason Battletoads conversations so often get to Turbo Tunnel before anything else.

The Legend of Zelda: Ocarina of Time | Water Temple

Ocarina of Time’s Water Temple has been one of the game’s most controversial dungeons since release.

Its reputation followed it for years. When the Nintendo 3DS version was developed, complaints about the Water Temple were specifically addressed, with changes made to make parts of it easier to read and navigate.

Ocarina of Time has dozens of famous moments, yet “Water Temple” is recognizable enough that many players immediately know which game you mean.

Ghostbusters | The Zuul Building Stairway

Ghostbusters on the NES isn’t just a game about climbing stairs. You drive around New York, buy equipment, and catch ghosts.

Then, near the end, you reach the stairs of the Zuul Building. You have to keep pressing A to climb while avoiding ghosts on the way up.

The section is often described online as being 23 floors, but the game’s own manual says the target is the 22nd floor.

The rest of Ghostbusters’ systems get talked about far less today. The stairs, on the other hand, still come up constantly whenever the NES game is mentioned.

Teenage Mutant Ninja Turtles | Underwater Dam

In the 1989 NES game, Area 2 sends you underwater. You have to defuse eight bombs before time runs out while navigating electric barriers, and the swimming controls don’t exactly make things easier.

Mention the NES TMNT game and the underwater section is still one of the first examples people bring up.

The 2022 Cowabunga Collection softens some of the original pressure with features like rewind and save-anywhere. The original level’s reputation was already firmly established by then.

Superman 64 | Flying Through Rings

Superman 64 has escort missions, bosses, mazes, and objects to carry around. But when most people picture the game, none of those are usually the first thing that comes to mind.

Instead, it’s Superman flying through a foggy city and trying to pass through rings.

And it isn’t even just one “ring level”; the game keeps bringing the same idea back. Eventually, those rings became enough to represent Superman 64’s entire reputation for bad design.

Gabriel Knight 3 | Cat Hair Mustache Puzzle

In Gabriel Knight 3, you need a fake mustache to impersonate someone else.

The solution? Collect hair from a cat, stick it to your face with maple syrup, and then draw a mustache onto the passport photo of the man you’re impersonating because he doesn’t actually have one.

Yes, that is genuinely the solution.

The Cat Hair Mustache puzzle has spent years as one of the classic examples of adventure game logic that leaves players wondering, “How was I supposed to think of that?” Even plenty of adventure game fans who never played Gabriel Knight 3 know the story behind it.

Call of Duty: Modern Warfare 2 | No Russian

Modern Warfare 2’s campaign has no shortage of memorable missions. Even then, No Russian stands apart from the rest.

You take part in the attack at Zakhaev International Airport alongside Makarov and his group. The player does not have to shoot civilians, and the mission can be skipped entirely.

That didn’t stop it from becoming a major controversy at launch, one that spread well beyond gaming media. Even later Call of Duty games with similarly shocking scenes still tend to get compared to No Russian.

Final Fantasy VII | Aerith’s Death

Final Fantasy VII is already one of the most famous RPGs in gaming history, so saying Aerith’s death became bigger than the game itself would be a stretch.

But the scene reached a different level of recognition.

Someone can know what happens to Aerith without ever playing Final Fantasy VII. Most gaming spoilers lose some of their weight over time, but this one can still sit right at the center of conversations about FFVII.

It’s one of those rare gaming scenes that can be recognized separately from a game that is already massive in its own right.

Mass Effect 3 | The Ending

Mass Effect 3 brings years of major storylines to a close, including the genophage, the geth and quarian conflict, and the relationships Shepard built across the series.

After launch, though, a huge amount of the conversation narrowed down to the final minutes.

The backlash led BioWare to release the free Extended Cut, adding new scenes and explanations to the ending. Mass Effect 3 as a whole has been reassessed plenty of times since then, but the ending debate never fully disappeared.

BioShock | “Would You Kindly?”

Throughout BioShock, Atlas tells you what to do. One phrase keeps appearing in the way he speaks to you:

“Would you kindly?”

At a certain point, you find out that it isn’t just a polite way of asking.

The twist is remembered for more than the surprise itself. BioShock takes the fact that players usually follow mission instructions without questioning them and makes that part of the story.

That’s why “Would you kindly?” is often one of the first things people remember when BioShock comes up.

Metal Gear Solid | Psycho Mantis

The Psycho Mantis encounter doesn’t play out like a normal boss fight.

He can react to certain games on your memory card, mess with you through controller vibration, and in the fight’s most famous trick, you’re expected to move your controller to the PlayStation’s second port.

In 1998, these were the kind of tricks that made players wonder, “How does the game know what I’m doing on my console?”

Metal Gear Solid already has a strong lineup of bosses. Psycho Mantis still stands apart because, for a few minutes, the game makes the console itself part of the fight.

Spec Ops: The Line | White Phosphorus

In Chapter 8, Walker orders the use of white phosphorus. You mark targets from a distance and carry out the strike.

Then you go down into the area.

Spec Ops doesn’t treat the aftermath like something you can simply skip past. The game makes you walk directly through the area you just attacked, and that sequence became the most talked-about part of its wider discussion about war, orders, and the player’s role in what happens.

If people pick one scene to represent Spec Ops: The Line, the white phosphorus sequence is very often the one that comes up.

Press F is a few seconds of button input. Driver’s parking garage appears before the first proper mission. Mass Effect 3’s ending comes at the end of a story that spans three games.

What puts them on the same list isn’t whether they’re good or bad. It’s that at some point, they stopped being just another mission or scene inside a game and became moments people mention separately whenever those games come up.

Total overhaul and total conversion mods belong to the second group. Some keep the original game’s world but redesign the systems you know. Others use the game’s engine to create what is almost a separate game, with new maps, characters, and stories.

That is why not every project on this list is the same kind of mod. We are putting mods like Enderal, which creates an entirely new RPG, under the same umbrella as projects like XCOM: Long War, which rebuild the systems of the existing game. What they share is how far they go: even someone who knows the vanilla game well has to learn it again.

Morrowind | Tamriel Rebuilt + OpenMW

Tamriel Rebuilt is not a typical expansion that simply adds a few new areas to Morrowind. For years, it has been building mainland Morrowind beyond Vvardenfell piece by piece, complete with cities, settlements, quests, and factions.

OpenMW replaces the 2002 engine with a modern foundation. Combined with Tamriel Rebuilt and other major mods, it turns Morrowind into something both much larger and far more flexible on a technical level.

The goal here is not to turn Morrowind into another universe. Quite the opposite: it is about creating more Morrowind on a scale Bethesda never completed.

Skyrim | Enderal

Enderal does not even use Skyrim’s world.

It runs on the same technological foundation, but it has a new continent, different lore, new characters, and a completely separate main story. Its character development and progression systems also work differently from vanilla Skyrim.

Calling Enderal a large Skyrim quest mod does not really do it justice. It is much closer to an independent RPG built on Skyrim’s engine.

Gothic II | The Chronicles of Myrtana: Archolos

Instead of adding new quests to Gothic II, Archolos uses its systems to build a new Gothic game.

Alongside a new protagonist, a separate open world, and an entirely new story, the economy, crafting, equipment, and character progression have all been reworked. Professional voice acting and original music also play a major role in making the project feel more like a full-scale RPG than a fan-made expansion.

Gothic II’s combat and controls remain familiar, but the world and campaign you play through are completely different.

Fallout 4 | Fallout: London

Fallout: London does not expand Fallout 4’s Commonwealth map. It takes the game out of America and moves it to London.

It introduces new areas, factions, characters, weapons, and a separate story. Fallout 4’s Creation Engine foundation is still recognizable, but the familiar American Fallout iconography is replaced by a different post-apocalyptic world built around British culture.

The result feels less like a Fallout 4 mod and more like another regional Fallout game Bethesda never made.

S.T.A.L.K.E.R. | Anomaly + GAMMA

GAMMA’s transformation does not come from a new story or an entirely new map.

Hundreds of changes built on top of Anomaly bring survival, crafting, equipment repair, the economy, combat, and progression together under one hardcore framework.

You can revisit locations you already know from vanilla S.T.A.L.K.E.R., but the way you survive in them changes completely. Finding a weapon is not enough; the condition of its parts, your resources, repair options, and the economy all have to be considered constantly.

That makes GAMMA less a single total conversion and more a massive overhaul pack that unifies many mods under one design philosophy.

Doom II | Modern Total Conversion Ecosystem

The Doom modding community stopped treating Doom as just a game to be modified a long time ago. Its engine is used more like a platform for building other games.

Projects such as Ashes 2063, Blade of Agony, and DemonSteele introduce new worlds, weapons, enemies, and movement systems. Some move far away even from Doom’s classic FPS structure.

At that point, it is not just the content that changes. Even the basic rules of how Doom is supposed to play are placed in the mod developer’s hands.

XCOM | Long War

Long War does not create a new universe. The alien invasion, the soldiers, and the XCOM organization all remain in place.

What changes is how the campaign works.

Strategic decisions, soldier development, classes, resource management, and the pace of the campaign all become much more demanding than in vanilla XCOM. Many habits learned by someone who has already finished the game no longer produce the same results.

It is one of the clearest examples of how far a total overhaul can go without changing the setting.

Mount & Blade II: Bannerlord | The Old Realms

The Old Realms leaves Bannerlord’s Calradia behind and moves the game into the Warhammer Fantasy universe.

Its factions, armies, equipment, and visual identity are all built around a completely different world. Bannerlord’s large-scale battles and campaign map remain at the core, but Warhammer’s fantasy systems are layered on top of them.

Bannerlord’s sandbox combat framework effectively becomes a strategy and battle engine that can be used for another IP.

Mount & Blade: Warband | Prophesy of Pendor

Prophesy of Pendor is not a mod that simply adds a few new factions to Warband’s classic Calradia campaign.

It builds its own world, factions, troop structures, and progression system. Warband’s core combat and campaign systems are still recognizable, but the political order and progression path the player encounters are clearly different from the Native campaign.

Despite its age, Warband remains one of the best-known examples of total conversion culture.

Crusader Kings III | Elder Kings II and AGOT

Crusader Kings III’s moddability goes far beyond alternate-history scenarios.

Elder Kings II brings CK3’s dynasty and character simulation to Tamriel. Historical Europe is replaced by the peoples, cultures, and religions of The Elder Scrolls.

A Game of Thrones projects take a similar approach, using CK3’s political and dynasty-based systems for Westeros.

Here, CK3 stops being just a history game and becomes a framework for running political simulations in entirely different universes.

Hearts of Iron IV | Old World Blues and Kaiserreich

Hearts of Iron IV is another example of the same transformation on the grand strategy side.

Old World Blues brings HOI4’s warfare, production, and diplomacy systems into the Fallout universe. Kaiserreich, meanwhile, rebuilds world history around a completely different alternate-history scenario.

HOI4’s core strategy systems remain intact, but the countries, political order, events, and campaign flow are no longer the same as the vanilla World War II scenario.

Medieval II: Total War | Third Age and Divide and Conquer

Third Age: Total War replaced Medieval II’s historical world with Middle-earth, effectively turning the game into a Lord of the Rings strategy title.

Its factions, map, units, and campaign structure are all rebuilt around Tolkien’s universe.

Projects such as Divide and Conquer, which grew out of Third Age, pushed that transformation even further. Medieval II’s age has not stopped it from remaining one of the first games people think of when total conversions come up.

Civilization IV | Fall from Heaven II

Fall from Heaven II turns Civilization IV’s historical race between civilizations into a dark fantasy 4X game.

Historical civilizations are replaced by fantasy factions, while magic, heroes, and religion systems become part of the game’s progression.

Civilization’s idea of advancing through the ages and eventually reaching the modern world is replaced by a completely different progression model.

Civ IV is still running underneath it all, but the world you are playing in has very little in common with Civilization’s classic historical simulation.

Civilization V | Vox Populi

Vox Populi takes a different approach. It does not add a new world or fantasy factions.

Instead, it reworks Civilization V’s AI behavior, diplomacy, balance, and core strategic systems.

At first glance, it still looks like the same Civilization V. A few turns later, however, you start to realize that the decisions and habits that worked in the vanilla game no longer work the same way.

It is not a total conversion, but a gameplay overhaul that directly redesigns the game itself.

Fallout: New Vegas | Fallout: New California

Instead of adding a few quests to the Mojave, Fallout: New California creates its own map and campaign.

Rather than starting as the Courier, the player begins in a different character context and enters a separate Fallout adventure centered around Vault 18.

New Vegas’s RPG systems and the broader Fallout identity are preserved. That means it does not create an identity as independent as Enderal, but it still offers an expansion-sized campaign that is completely separate from the vanilla main game.

BattleTech | RogueTech

RogueTech turns BattleTech’s story-driven campaign structure into a much more open sandbox.

Along with new equipment, contract types, factions, and planets, the MechLab becomes far more complex. Deciding which parts go into which mech and how to build your force requires much more planning than in the vanilla game.

Finishing BattleTech does not mean you are ready for RogueTech. The whole point of the mod is to rebuild the same systems into something denser and more punishing.

Freelancer | Discovery Freelancer

Discovery changes what Freelancer is played for more than it simply expands the single-player space adventure.

Alongside new systems, ships, and equipment, multiplayer roleplay and a persistent server structure take center stage. Instead of the vanilla campaign, relationships between players, factions, and the long-term multiplayer order become more important.

That is why it would be misleading to think of Discovery as a traditional single-player total conversion. It is a major world and gameplay overhaul that pushes Freelancer toward an MMO-like social space sandbox.

Oblivion | Oscuro’s Oblivion Overhaul

Oscuro’s Oblivion Overhaul does not create a new Cyrodiil.

Instead, it changes Oblivion’s controversial level-scaling system, creating a more dangerous world where enemies, loot, and encounters are not tied so tightly to the player’s level.

Most of the map and quests remain familiar, but it becomes harder to predict what you might run into while exploring.

It is one of the classic examples of an overhaul that changes the basic logic of the vanilla experience without creating an entirely different game.

Star Wars: Empire at War | Thrawn’s Revenge and Major Conversion Mods

Empire at War’s modding community has done more than keep an old strategy game alive for years; it has also created separate campaign experiences for different eras of Star Wars.

Projects such as Thrawn’s Revenge move away from the game’s original era with new factions, units, and strategic campaign structures.

Empire at War’s core structure of space and ground battles remains intact, but the era you fight in, the powers you control, and the way the galactic map works can all change dramatically depending on the mod.

A Total Overhaul Does Not Always Mean Making a New Game

The differences between these 19 projects also show why “total overhaul” does not describe just one kind of mod.

Projects such as Enderal, Archolos, and Fallout: London become almost separate games through their new worlds and stories. Long War, RogueTech, Vox Populi, and Oscuro’s Oblivion Overhaul keep the universe you already know but redesign its systems.

In projects such as GAMMA, the transformation comes from hundreds of mods being brought together under a single progression model. In Doom, the engine itself has become a platform for building other games.

They do not all arrive at the same result. But when they work, going back to the vanilla game no longer feels like seeing only less content; it feels like returning to a different game governed by different rules.