Epic Games’ 2026 AI MegaJam has produced a particularly tidy irony: the game currently leading its public ratings is LSD:FIH, and its creator says that no artificial intelligence was used to make it.
The contest is framed around generative-AI-assisted development and includes the slogan “vibe your way to $50k.” Entrants were permitted to use as many AI agents as they wanted. But permission is not necessarily a requirement, based on the available contest description, and LSD:FIH appears to be making the most of that distinction. Its page opens with an unambiguous statement: “No AI was used to make this game.”
That has put a small, plainly strange indie project at the centre of a much larger argument about what AI tools are for, what audiences reward, and whether a jam designed to showcase a production method can also celebrate an entry that opts out of it entirely.
A game jam result with an unexpected leader
A game jam is a time-limited game-development event. Creators work to a set theme, set of tools, or other constraints, then submit a playable project by a deadline. The compressed timeline is the point: jams often reward a focused idea, a memorable presentation, or a game that finds a hook quickly rather than a broad, polished production.
For the AI MegaJam, the relevant window ran from November 12 through November 15: four days. Submissions have now closed, while voting remains underway for the next several months. The announced results date is January 29, 2027.
At present, LSD:FIH is at the front of the ratings feed. The creator characterizes it as a computer game about “a fih aura-farming,” a description that is more surreal than explanatory. Yet that oddness may be part of its appeal. Jam entries routinely compete for attention in a crowded field, and a deliberately peculiar premise plus a conspicuous “no AI” label is a very effective way to give people something concrete to discuss.
One voter described the entry as the sort of imaginative, enjoyable game people in the 2000s might have expected an AI-made game to be, contrasting it with “generated slop.” That response is an opinion, not a technical assessment of the game or of AI-created work as a whole. Still, it illustrates why the project’s lead matters: the conversation is not only about whether tools were used, but about originality, intent and the feeling players get from a short-form game.
What “vibe coding” and AI agents mean here
Generative AI refers to systems that generate material in response to prompts or other inputs. In a game-development context, that material can include written text, code, images, sound concepts or design suggestions. The term says something about the method used to produce material; it does not, by itself, establish that a resulting game is good, finished, creative or commercially viable.
Related coverage includes An AI-Free Game Is Leading Epic's 2026 AI MegaJam Vote.
Vibe coding is an informal label often used for a prompt-led development process: a creator tells an AI what they want, evaluates the output, asks for revisions and pieces the results into a working project. It emphasizes direction and iteration rather than writing every line of code manually. That does not mean the human participant has no role. They still must decide what to ask for, determine whether an output works, handle errors and shape the project around its limitations.
An AI agent, in this setting, is a tool that can take on multi-step tasks after receiving an objective. Instead of producing a single answer, an agent may attempt a sequence of actions such as proposing code, revising it or working through a development task. The MegaJam’s allowance for entrants to use as many agents as they like indicates an open approach to such tools.
But an allowance is different from a mandate. The available details say participants may use agents, not that every submitted game must include generative AI in its development. That distinction is central to the current situation. If the published rules do not require AI use, an AI-free game is not automatically at odds with the contest’s eligibility conditions merely because the event is focused on AI-powered creation.
Why the entry’s status is getting attention
Epic’s support for AI makes the leaderboard especially notable. Earlier this year, the company described a major shift toward maximum AI acceleration, while Epic chief executive Tim Sweeney advocated an “AI first” approach. Sweeney has also objected to Steam’s AI disclosure requirement, arguing that it is unfair to developers who use AI to make good games.
Against that backdrop, an AI-free game leading an AI-branded jam reads as a neat public contradiction. It is not necessarily a rejection of AI by every voter, and it does not prove that AI-assisted entries are less popular or less capable. Public voting is influenced by plenty of factors: a game’s concept, presentation, visibility, community sharing, the novelty of its premise, and the social debate surrounding it.
It does show that players can respond positively to a game that gives them a clear point of view. In this case, LSD:FIH has made the production method itself part of its identity. The “no AI” declaration is not hidden in technical notes; it is presented as a defining fact about the project. In a contest where many entrants are likely foregrounding which new tools they used, choosing not to use those tools becomes its own form of differentiation.
There is also a practical lesson for jam participants beyond the AI debate. A narrow development window rewards decisions that make a project legible immediately. A concise premise, an unusual name and a clear statement of authorship can help a small game stand out before players even reach its mechanics. That does not mean every developer should build their pitch around controversy. It means that clarity is valuable when people have many entries to evaluate.
What the leaderboard does—and does not—settle
The current rating lead does not settle any broad argument about generative AI in games. It does not establish that human-only work is inherently better than AI-assisted work, just as the event’s AI focus does not establish that tool use alone produces compelling games. A jam vote measures the response to individual entries under specific conditions, during a limited period, with an audience that may be motivated by novelty as much as play.
Nor is there enough information to treat the top position as a final result. Voting is still active, and the published results are not due until January 29, 2027. The standings can change before then. It is also not known whether the organizers will comment on the prominence of an entry whose creator says it used no AI.
Speculation that the game could be disqualified should therefore remain speculation. The key question is the written eligibility criteria, not the event’s marketing language alone. The supplied contest details specifically point to a permissive rule—participants may use as many agents as they like—rather than a stated obligation to use them. Until organizers say otherwise, the sensible reading is that the contest has room for differing workflows.
The useful takeaway for players and creators
For players, the MegaJam’s emerging story is a reminder to look past a development label. “AI-made,” “AI-assisted” and “made without AI” are meaningful descriptions of process, but they are not substitutes for engaging with the game itself. A tool disclosure can provide context; it cannot fully explain whether an idea lands.
For creators, the four-day development period behind LSD:FIH underscores how much can hinge on a project’s central identity. AI tools can potentially speed up parts of a workflow, but speed is only one component of making something people want to play or share. A creator still needs a recognizable idea, decisions about what to include and omit, and enough coherence for the finished game to communicate its purpose.
For contest organizers, meanwhile, the situation is a useful test of language. Events centered on emerging tools should be precise about whether those tools are optional, encouraged or compulsory. Those are three very different policies. Clear rules protect entrants from uncertainty and let voters understand what an event is actually intended to reward.
Whatever happens when the votes close, the present standings have already given the AI MegaJam an unexpectedly human headline: a four-day project publicly defining itself by the technology it did not use is currently beating projects in a contest built to spotlight that technology. For more on the industry pressures and shifts surrounding game development, see our coverage of Firaxis’ job cuts and reported new XCOM milestone trouble.







