🕹️ PS42013Naughty DogVERY HARDThe Last of Us rebuild deck
VERY HARD
Perception, investigation and cover are all promptable inside a week; the product is a fifteen-hour performance-captured script, and there is no prompt for a performance.
You will get a good prototype. The product is what comes after.
The clearest NOT REALLY on the list, and it is worth having one. The individual systems — perception, investigation, cover — are all promptable and make a great prototype. But this game is 700 team-years of performance capture, writing and art direction, and no amount of prompting produces a performance. Build the encounter to learn stealth AI. That is the honest version.
- Difficulty
- VERY HARD · 4 of 5
- First playable
- A project
- The original
- 700 team-years
- Released
- 2013 · PS4
- Genre
- Action Adventure
Build a third-person stealth-survival encounter prototype. three.js, ONE room, ONE encounter. Do not attempt the game. Requirements: - Scarcity-driven combat: 4 bullets, 1 crafting resource, enemies that take 2-3 hits. Every encounter must be solvable by stealth, by combat, or by running, and combat must be the worst option. - Enemy perception as an explicit, debuggable system: a vision cone with distance falloff and line-of-sight raycasts, a hearing radius that scales with player movement speed, and a three-state alertness (unaware, suspicious, alerted) with decay timers. Render the cones and radii in a debug view — an unreadable stealth system is an unfair one. - Investigation behaviour: a suspicious enemy walks to the last known position, searches nearby, and calls to others. Alerted enemies flank rather than charging. - Listen mode: hold to reveal enemy positions through walls as outlines, at the cost of being unable to move quickly. - Crafting from a shared resource pool, so a health kit and a molotov compete for the same materials, performed in real time without pausing. - Cover: contextual, no button — the character lowers automatically near waist-high geometry, and aiming from cover uses an over-shoulder camera with a tight FOV. - A companion AI that follows, stays out of your line of fire, does not break stealth, and is never seen by enemies. Cheat openly: make companions invisible to enemy perception. Every studio does this. Definition of done: I get through the room without firing a shot, then replay it loud and it feels like a mistake.
- Everything. This is a 700-team-year production and the product is performance, writing and art
- Motion capture, facial animation, and voice work across a 15-hour script
- Companion AI that never breaks the fiction, which is one of the hardest problems in games
- Environmental art at a density no procedural system reproduces
Enemies that act on last known position
The moment a guard commits to where you were rather than where you are, it stops being a sensor and starts reading as a mind. Separating belief from truth is the entire trick, and it is also what makes the system debuggable, because a belief is a thing you can draw on screen.
ScopeAbout a week— build just this, not the game
Build a stealth-perception toy as a single self-contained HTML file, Canvas 2D, top-down, no build step. One room with walls, one player, three guards. Each guard holds a belief about the player: a last known position, a confidence value, and one of three states — unaware, suspicious, alerted. Perception writes only to the belief, never straight to behaviour. A vision cone with distance falloff plus a line-of-sight raycast raises confidence quickly; a hearing radius scaled by the player's movement speed raises it slowly. Confidence decays every tick. Crossing thresholds changes state. Behaviour reads only the belief. A suspicious guard walks to the last known position and searches around it. An alerted guard goes there and shouts, which writes that position into the other guards' beliefs — so they can all be wrong together. Render the cones, the hearing radii, each guard's believed player position as a translucent ghost, and the state above their head. Definition of done: I break line of sight and watch a guard commit to a spot I am no longer standing in.
Not a clone — The Last of Us re-themed and cut down to something you can finish. The angle is the important reel: it is what turns a project into a weekend. 24,360 ways to land.
Gumshoe Idle
The Last of Us, but an idle version that plays while you are away, set in a noir city. Rendered in Game Boy four-shade green.
Build a browser game called "Gumshoe Idle". The pitch in one line: a action adventure in the spirit of The Last of Us, but an idle version that plays while you are away, set in a noir city. WHAT THIS IS Take the SHAPE of The Last of Us — its core loop, and whatever makes that loop good — and rebuild it as your own game in a different world at a smaller scope. Use your own names, art and audio, which you want anyway: a reskin of somebody else's game is less interesting than a new one that works the same way. THE ANGLE — this is the important constraint, honour it above everything else An idle version that plays while you are away. Offline progress on load, big-number maths from line one, and a prestige reset. The fantasy survives; the input does not. This is what makes the project finishable. If a decision would push the scope back towards the original's, take the smaller option every time. THE THEME A noir city. Rain, venetian blinds, and information as the only currency. Everyone is lying by omission and the player must notice which parts. The theme is not a coat of paint. Let it change what the mechanics mean — a reload, a health pack and a locked door should all be things that make sense in this world and nowhere else. LOOK Game Boy four-shade green. A strict four-colour palette, 160×144 internal resolution scaled up with nearest-neighbour, and no anti-aliasing anywhere. REQUIREMENTS - Runs in a browser with no build step. Canvas 2D unless the angle genuinely requires 3D, in which case three.js. - Fixed-timestep update loop with interpolated rendering, so it behaves the same at 60Hz and 144Hz. - Every tuning constant in one CONFIG object at the top, exposed as live sliders in a debug panel. You will find the feel by dragging those, not by prompting. - All content — levels, entities, balance numbers — in JSON or plain data files, never inline in the logic. - Audio generated with the Web Audio API rather than asset files. - Respect prefers-reduced-motion: keep the fades, drop the shake and the parallax. BUILD ORDER 1. The core loop with placeholder rectangles, at the reduced scope the angle demands. No theme, no art, no audio. Make it fun as rectangles first. 2. The angle's consequences. Cutting to bots, or to turn-based, or to one level changes the design rather than just shrinking it — find out how and rewrite step 1. 3. The theme, as content and data. 4. The look, applied last as a rendering layer over a game that already works. Start by writing a one-page plan: the core loop, what the angle removes and what that frees you to do properly, and what you are deliberately leaving out. Wait for me to approve it before writing code.
Half the time one of these already does what you wanted, and the other half it saves you a week of solving a problem someone documented in 1997.
One count per game. No account — it is tied to a salted hash of your IP, which is not stored in a form anyone can reverse. Every build adds 700 team-years to the counter on the homepage.
Then try these
Questions
How hard is The Last of Us to rebuild with an AI agent?
Yes — every game is buildable at some scope. The question is how hard, and The Last of Us is VERY HARD (4 of 5). You will get a good prototype. The product is what comes after. Perception, investigation and cover are all promptable inside a week; the product is a fifteen-hour performance-captured script, and there is no prompt for a performance. Reckon on a project to something you can actually play — not to something that matches the 700 team-years Naughty Dog spent.
What makes The Last of Us hard to rebuild?
Perception, investigation and cover are all promptable inside a week; the product is a fifteen-hour performance-captured script, and there is no prompt for a performance. That is why it sits at VERY HARD rather than a tier either side of it. If you want the idea without the project, the transferable part is enemies that act on last known position — about a week of work.
How long does it take to build a The Last of Us clone?
Wrong question, and it is worth saying why. Nobody rebuilds The Last of Us — the original took Naughty Dog on the order of 700 team-years, released in 2013 on PS4. What you can do is reach something playable and genuinely yours in about a project, and then keep going for as long as it stays interesting. Any figure that claims otherwise is measuring a prototype and calling it a game.
What do I lose by building my own The Last of Us?
Everything. This is a 700-team-year production and the product is performance, writing and art. Motion capture, facial animation, and voice work across a 15-hour script. Companion AI that never breaks the fiction, which is one of the hardest problems in games. Those are the parts a prompt will not hand you.
Which AI coding agent should I use to build it?
Any of Claude Code, Codex or Cursor will handle this prompt. Paste it as your first message, let the agent scaffold the project, then iterate on feel — the second, third and tenth prompts are where a game actually gets good. Work in an empty folder or a fresh git branch so you can throw it away cheaply.
One email when new games go on the list. No other reason to email you, ever.