Jeff Kaplan: World of Warcraft, Overwatch, Blizzard — 主题精读稿
Jeff Kaplan: World of Warcraft, Overwatch, Blizzard — 主题精读稿
前言：一个从拒信堆里走出来的游戏设计师 (00:00 - 05:00)
Jeff Kaplan is one of the most consequential game designers in history — the lead quest designer and eventual game director of World of Warcraft, then the creative director and game director of Overwatch. This five-plus-hour conversation with Lex Fridman is not really about games. It is about the architecture of a creative life: how obsession becomes craft, how failure becomes wisdom, how a person survives the gap between what they imagined building and what they actually shipped.
The thread connecting Kaplan's entire story is the question of what it means to find your people. He spent his twenties in isolation — writing alone, drinking alone, rejected by 170 literary magazines in a single year, so broken he eventually underwent electroconvulsive therapy. Then he found a guild in EverQuest, then a dorm-like office in Irvine where people in black T-shirts threw frisbees, and for the first time felt fully himself. Everything followed from that.
一、The Foundation: Zork, Arcades, and the Intimacy of First Person (00:00 - 25:00)
Growing Up in the Age of the Arcade
Kaplan grew up in the early era of coin-op arcades — Pac-Man, Galaga, Asteroids — the moment when video games first became a mass cultural force. For a kid who was already drawn to storytelling and world-building, the arcade represented something more than entertainment. It represented the possibility of inhabiting another reality, even briefly.
The family IBM PC brought that possibility home. Text adventures like Zork became formative for Kaplan in a way that graphical games wouldn't until much later. Zork demanded a specific kind of engagement: you typed commands — "go north," "take lamp," "attack grue" — and the game responded in prose. The world existed only in the text and, crucially, in your imagination. There were no graphics because none were needed; the image was always richer in your head. Kaplan describes the experience of playing Zork with his mother and brother — huddled around the IBM, figuring out how to keep the lights on or the Grue would eat you — as one of the core emotional memories of his early gaming life.
When Zork later got a graphical remake in the CD-ROM era of the 1990s, Kaplan found it disappointing in a specific and instructive way: the visuals replaced the imagination, and the version in his head had always been better. The graphics shattered the Zork he knew. This would become a lasting design instinct — the importance of what you leave for the player to fill in.
Richard Garriott's Ultima series deepened the love for complex virtual worlds. Kaplan started with Ultima II and was struck by the richness of a world that had its own rules: you could get on a rocket ship in the middle of a fantasy RPG, you could try to rob merchants, you could attempt to kill Lord British himself (extremely difficult, and very much the goal of every kid playing the game). The bouncers in town taverns, the merchant interactions — small details that made a world feel inhabited, consistent, alive.
The Carmack Revelation: Technology as Art Form
Then came id Software, and the breakthrough that Kaplan credits as the reason he's sitting in the interview at all. John Carmack, John Romero, and the id team transformed gaming with Wolfenstein 3D and Doom — not just graphically but experientially. Kaplan had played the original 2D Wolfenstein, the one where you run around dressed as a German soldier. And then 3D Wolfenstein arrived.
"You look back at screenshots of it now and it seems almost childish — why were you so excited about that?" Kaplan says. But he was transported. The specific thing that transported him was the intimacy of first person: hands in front of you, holding a gun, being dropped into Nazi Germany as the hero fighting back. The camera position made the experience personal in a way no game had achieved before. You weren't watching a character. You were the character.
Doom deepened the effect. Kaplan was a huge Army of Darkness fan — the Sam Raimi horror-comedy with the boomstick and the deadites — and when Doom arrived, he experienced it as Army of Darkness: The Video Game. The graphical advances were real, but more important was the smoothness, the responsiveness, the sharpness of play. The feel of a game, its kinetic quality — how it responds to your inputs — was already, in Kaplan's mind, inseparable from its quality as an experience.
He gives Carmack and Romero unreserved credit: "As somebody who worked on an FPS, that wouldn't have existed without them. Credit where credit's due." And he notes, typically for him, that his range as a gamer was already broad — he loved the first-person shooter intimacy and the world-rich exploratory MMO, two modes that seemed incompatible but would eventually converge in his own work.
Online: The Magic of Knowing There's a Real Person on the Other Side
The transition to online gaming began with Doom and Duke Nukem's primitive LAN and modem modes. Kaplan describes the specific emotional experience of first seeing another entity in a video game and knowing it was a human being somewhere else in the world. "That was magical. Like that moment happened and that person could be in another room or across town from you." The encounter with genuine otherness — another real mind inside the game — changed what he understood games to be.
Quake refined this. The Quake community had websites, mods, clans, a social infrastructure that prefigured what internet gaming communities would become. Kaplan read Blue's News religiously — to this day, he says, it's the only gaming website he follows — and it was through Blue's News that he encountered the .plan file update from a programmer at id Software named Brian Hook. Hook announced he was leaving id to go work on something called EverQuest. Kaplan was baffled: how does anyone leave id, the greatest institution in gaming, for anything else? Either Hook was crazy, or whatever EverQuest was, it was worth finding out.
二、EverQuest and the Discovery of Community (18:00 - 58:00)
What EverQuest Was
Kaplan had missed Ultima Online — he was in grad school when it launched and too busy to play — which he describes as an unfortunate miss, because Ultima Online was the earliest laboratory for grief-play, emergent social dynamics, and the chaos of putting real people in a persistent world together. The famous Ultima Online stories involve players poisoning apples and leaving them on the ground, waiting for someone's horse to eat one, stealing everything when the horse died. "It was the earliest grief-based experiment. Really, like, when you're treating the humans like ants in the ant farm."
EverQuest arrived as a more structured alternative: still massively multiplayer, still a persistent world with real strangers, but with quests, classes, and a clear progression system. For Kaplan, coming off the depression and isolation of his failed writing career, EverQuest arrived at exactly the right moment.
He logged 6,000 hours. He rose from player to officer to acting guild leader to full guild leader of Legacy of Steel, one of the best guilds on the Nameless server. Leading raids in EverQuest was his first experience with real leadership — the kind where you have to motivate people who have no obligation to listen to you, who all have different goals, and who are connected by nothing but their shared presence in a virtual world.
The Texture of an EverQuest Raid
Running a 50- or 60-person raid in EverQuest required a specific combination of organizational skill, emotional intelligence, and genuine in-game competence. "Organizing people in an online game like EverQuest is like herding cats," Kaplan says. Some players were loot-motivated. Some cared about guild rank. Some were lonely and came for the company. Some were deeply depressed — Kaplan was among them — and the game provided escape and structure when nothing else did.
The diversity of motivations and personalities made leadership genuinely difficult in ways that had no parallel in his previous experience. "When do you force people versus encourage them?" He had the safety net that the original guild leader would sometimes log in — but over time that safety net became more theoretical than real, and Kaplan found himself building out his ability to lead de facto.
He also met his wife in EverQuest. He emphasizes this without editorializing: she was in the guild, they connected through the game, and the life he has built — career and family — traces directly back to those 6,000 hours on the Nameless server. EverQuest was not an escape from real life. It was where real life happened.
The Blizzard Lunches
Among the guild members Kaplan eventually met was someone known in the game as "Ariel" — a female wood elf who turned out to be Rob Pardo, then lead designer on Warcraft 3. Ariel/Pardo noticed that Kaplan had been making levels for Half-Life (another game that shipped its editor on the CD, inviting players to create content) and asked him to mail a CD with the levels to an address in Irvine.
The lunch that followed was transformative in a very specific way. Kaplan had always been ashamed of his love of games — had never told friends or family how deeply he was invested, had hidden this part of his identity. Gaming carried a stigma, and he carried the stigma as internal shame. At that lunch in Irvine, sitting with several Blizzard employees who also turned out to be guildmates, talking about dragons and raid tactics and shit-talking other guild members — he felt, for the first time in his adult life, fully like himself.
"I'm out to lunch with these guys and we're talking about dragons and swords and raid tactics. And I literally had this moment where I felt like myself. I just felt so comfortable." The coming-out metaphor is one he uses deliberately: he was hiding who he was, and this was the moment he stopped.
The lunch with "Barfa the Troll Warrior" added the final piece. The Hole dungeon story — Kaplan's rogue guild saving Barfa by handing over an expensive teleport potion when the entire raid had wiped out — had been a small act of generosity to a guildmate he didn't know. It turned out Barfa was Alan Adham, co-founder of Blizzard. At lunch, Adham said: "You saved me in the Hole that time." Kaplan hadn't known who he was. He'd just done the right thing in the dungeon.
The Blizzard job posting appeared the next time Kaplan checked the site. Associate quest designer, World of Warcraft. Requirements: strong writing skills and a creative writing degree. "You guys set this up for me — like they were just interviewing me for six months."
The Hiring Gauntlet
Kaplan applied with "all his heart" and went through a six-month interviewing process — unusual because Blizzard typically promoted designers from within, from QA or tech support, never from outside the company. Hiring someone off the street was a significant departure.
His most memorable interview was with Chris Metzen and Kevin Jordan. Jordan was quiet; Metzen owned every room he walked into. They went to an Italian place across from Blizzard. Metzen stopped to buy cigarettes on the way. Every other word out of Metzen's mouth was profanity. Kaplan had come from his father's professional recruiting world, where you never cursed in interviews. "Being around people who didn't care about what the corporate norms were was so inspiring. And then my last interview was with Alan and Rob and Bob Fitch. They took me to an ARCO station that had a Jack in the Box. And I remember thinking to myself: these guys just brought me to a Jack in the Box that's in an ARCO station. I need to work here. Like, these are my people."
三、Depression, ECT, and the Work That Pulled Him Out (1:09:00 - 1:16:00)
Lex asks directly about the low points — the depression and alcohol that marked the writing years. Kaplan's answer is unusually honest.
The drinking was bad. He describes a specific memory: drunk watching the Oscars alone, the contrast between his own sense of failure and the polished success on screen making everything worse. The cycle of trying not to drink but drinking to feel better became its own trap. Therapy helped — but crucially, he emphasizes that you have to find the right therapist, not just any therapist. Checking the therapy box without finding someone who genuinely helps you get out of the rut in a way that works for you doesn't accomplish much.
Antidepressants didn't help either — he hated the feeling of being chemically altered, and never responded to them well. Then came the revelation he'd never shared publicly before. He underwent ECT — electroconvulsive therapy, electric shock therapy. He is careful not to endorse it as a cure-all: "That was, I was at such a low point that people were very worried about me and my well-being and what was going to happen. And that was sort of an extreme pull the rip cord, like there's nothing else to lose moment." He believes it was a difference maker. That, and starting at Blizzard.
The deeper point Kaplan makes is about the specific loneliness of the introvert. He had been in a writers' group in New York and that community had kept him functional. He left for a girl, the relationship ended within two months, and he was alone in California. The social infrastructure that writers have — giving each other's stories to people who will genuinely criticize them, not just your parents — had been severed. Introverts, he says, sometimes need other people more than extroverts do, not less, and are worse at finding and maintaining those connections.
Finding his people at Blizzard — the gaming community in general, the company specifically — was as therapeutically significant as the ECT. "I think introverts almost need people more and we don't always know how to engage in the right healthy ways." The work itself, too — the specific creative hard work on a game he loved — was what finally pulled him out. Not just the people, but the thing they were all doing together.
四、Early Blizzard: The Culture That Made Great Games Possible (1:16:00 - 1:45:00)
What the Office Actually Looked Like
On his first day in May 2002, Kaplan arrived at Blizzard's offices on the UC Irvine research and development campus. The surrounding buildings housed Cisco and other tech companies — professional, corporate, formal. Outside the Blizzard offices, people in black T-shirts and shorts were throwing frisbees and riding scooters. Inside, there were futons (for the inevitable late-night crunch sessions) and posters on the walls. It looked like a dorm.
Blizzard had fewer than 200 people at Kaplan's arrival, including the Blizzard North studio in San Mateo. The WoW team — Team 2 — was the scrappy underdog within the studio. Team 1 was the revered RTS group responsible for Warcraft and StarCraft. Team 2 had tried and failed to spin up multiple times before WoW. A previous game called Nomad had been scrapped. Alan Adham had steered the remnants of Team 2 toward World of Warcraft.
Kaplan arrived as one of the very few designers ever hired directly from outside the company. The ID badge on his first day read "Associate Game Designer." His offer letter was for $35,000 a year.
The Five Disciplines and Why Their Tension Makes Games
Kaplan walks through the full architecture of a game development team: engineers (writing code, from server networking architecture to gameplay systems), artists (3D modeling, environment art, animation, tech art, lighting), game designers (creating the player experience — quest systems, loot tables, combat tuning, progression), producers (project management), and audio (sound designers and composers, the most underappreciated discipline on most teams).
The creative tension between disciplines — artists who might be literally drawing on any available piece of paper while in a meeting, and programmers who think in crystalline logical structures — is, in Kaplan's view, the actual source of what makes great games. The best outcomes come from pairing people whose minds work completely differently around a shared goal. A great graphics programmer and a great lighting artist, working together on a specific visual problem, produce something neither could alone.
Audio and lighting are the two disciplines that no one notices until they're missing. Play a game without a finished audio mix or with placeholder lighting, and something feels profoundly wrong. Add great audio and great lighting, and you've tapped into senses that the player wasn't even tracking consciously. Kaplan learned this deeply on Overwatch, where the audio team had a dedicated audio programmer doing nothing but audio hooks — a level of resource investment in sound that most studios wouldn't consider justifiable.
Small Teams: Why Everyone Having a Loud Voice Matters
The power of small teams is Kaplan's most consistent design philosophy. On a team of ten people at a game's incubation stage, all ten are in the room for every decision. The junior engineer is in the art style meeting. The associate game designer is in the server networking discussion. Nobody is siloed. Nobody is compartmentalized. As teams grow, this changes — not because anyone decides to change it, but because it becomes physically and logistically impossible to keep everyone in every room.
The psychological consequence of compartmentalization is subtle but serious. When you stop interacting with people from other disciplines, you start to view their work with skepticism. You begin making statements like "the artists just don't get it." And when you examine that statement carefully, Kaplan says, it's obviously an asshole statement. "The artists signed up for this. This game is going to be as much theirs as it is mine. So who am I to say a statement like that?" On his Overwatch team, the practice he tried to instill was: assume we've recruited the best prop artist in the industry, because we have. When that person suggests something, listen, because they might be right.
The Lesson Pardo Taught Him About Leadership
In his early days as a lead on WoW, Kaplan fell into the insecure new manager trap: the compulsion to be right. When people in meetings threw out ideas that seemed wrong to him, he would — politely, but systematically — explain why the idea was wrong and what they should do instead. Rob Pardo pulled him aside after one such meeting. "You're a very smart designer," Pardo said, "but you shouldn't do what you just did to those people. You should always listen to what people have to say and try to make their ideas work."
That correction — which Kaplan initially received as a rule to follow rather than a belief — gradually transformed into one of his core operating principles. Try to make other people's ideas work. When an idea that your instinct says is wrong turns out, on development, to be brilliant, the most gratifying outcome is not that you were proven wrong but that the person whose idea it was gets all the credit. An insecure early-2000s Kaplan would have killed the idea in the meeting. Instead, it shipped and became something players loved, and someone else got the recognition for it.
五、Building World of Warcraft: Quest Design, Zone Flow, and the Path of Least Resistance (1:45:00 - 2:30:00)
What World of Warcraft Is (Really)
Chris Metzen, Blizzard's creative director, had a phrase that Kaplan absorbed as foundational: "The lead character of World of Warcraft is the world." Not the player character. Not any NPC. The world itself — Azeroth — is what the game is about. The goal of design is to create a place that is exciting and dangerous but also comfortable and uncomfortable, massive but navigable, the kind of place people want to spend time living in, not just visiting.
Kaplan offers a visual description of the early WoW design sessions: whiteboards everywhere, Chris Metzen drawing as he talked — gorgeous maps emerging in real time from a mind that seemed to contain the entire world already formed. Here's the Dwarven lands; here's Ironforge; here's Elwynn Forest where the humans start; here's Westfall with the Defias Brotherhood and the Deadmines. The place names and their associations were inseparable from the stories Metzen was already imagining in them.
The world was split into two warring factions — Horde and Alliance — a decision that was fiercely controversial internally. Kaplan and Pardo both came from EverQuest, where all races could group together, and they argued against the split. Rob Pardo's view was influenced by Dark Age of Camelot, an MMO with three factions, where being on a team from minute one removed the anxiety of being a loner in a world. Alan Adham championed the split. Kaplan and Pardo argued and argued. Then Alan retired (he left before WoW shipped, to go run a hedge fund and get into poker and finance — "kind of nuts"), and when Pardo took over as lead, his first act was to reverse his own position. "Alan's a smart guy. The fact that he was fighting so hard for Horde/Alliance — we got to do it." The Horde/Alliance split became one of the deepest cultural markers in gaming history. People have Horde symbols tattooed on their bodies. It is who they are.
The Quest Experiment That Changed MMO Design
The original design philosophy for WoW's leveling was based on the EverQuest model: do a cluster of quests in an area, then kill creatures in the area while grinding for experience points, then move to the next area. Pat Nagle designed the Elwynn Forest starting area with roughly 20 quests and assumed players would spend several hours grinding after finishing them.
After a team playtest — a couple of hours in Elwynn Forest with players who came from shooter backgrounds, not MMO backgrounds — the feedback was unanimous and alarming: "My God, that was horrible. I ran out of quests right away. I expected quests to just keep going the whole time."
The team had a collective "oh shit" moment. They had massively underestimated the number of quests needed. And the realization led to a deeper design principle: design along the path of least resistance. In EverQuest, the fastest way to level was to find the easiest creatures and kill them in the same spot for hours — what some players find meditative and others find mind-numbing. In WoW, the team decided to make quests the path of least resistance by overloading experience into quest completion. The fastest way to level became: do quests.
The consequences radiated outward in ways they couldn't fully anticipate. Quest-driven leveling moved players through the world, which meant players would see all the content designers had built. It enabled storytelling. It created directed gameplay that felt optional but structurally wasn't. "WoW allowed you to play as a single player. And what makes an MMO massive is having the other people there — they're so important, or else the world feels kind of wrong and dead. But the concept that we have to force you to interact with them to do anything is very off-putting to a lot of people." By making quests the default path, WoW opened the genre to a vastly larger audience.
Kaplan argues this shift — which seems obvious in retrospect — is why WoW scaled to tens of millions of players while EverQuest peaked in the hundreds of thousands.
Zone Flow: Designing the Experience of Discovery
The zone flow document that Kaplan shared with Lex before the interview represents one of the fundamental design artifacts of WoW's original structure: a map of how players would move through the world and what they would experience in what order. The principle was simple: the scary places — Burning Steppes, with its lava and dragons — belong at the end, as destinations that have been planted in the player's imagination long before they're ready to visit.
But the implementation involved a quiet masterstroke: Blizzard gave dwarves and humans a free flight path between their starting cities, Ironforge and Stormwind. Level-five newbie dwarves could fly to meet their human friends — or vice versa — and the flight path took them over the Burning Steppes and Searing Gorge. "You look down and you're like, holy shit, that looks scary and dangerous. Plant that seed of things to come." The world communicated its own scope and stakes through the experience of traversing it.
Green Hills of Stranglethorn: A Designer's Honest Mistake
Kaplan's most infamous quest design is the Green Hills of Stranglethorn — universally hated by WoW's original playerbase, and one he examines with characteristic honesty. He wrote a short story (paying homage to Hemingway by naming the quest-giver Hemet Nesingwary — Hemingway's name rearranged — and including another character whose name rearranges Kerouac), broke it into pages, and scattered the pages across creature drops throughout the entire zone. Players had to collect all the pages to complete the quest.
The intention: players would socialize in zone chat, trading pages they had duplicates of for pages they were missing. The fantasy was of a spontaneous community forming around a shared scavenger hunt. The reality: WoW's inventory system didn't allow pages to stack, pages clogged limited bag space with duplicates nobody could trade efficiently, and the "community" that formed was mostly people being angry in zone chat.
Kaplan uses the phrase "ant farm designer" to describe his mindset when designing it: treating players as experimental subjects in a behavioral study, optimizing for the designer's entertainment rather than the player's. He invokes Sid Meier's three types of fun — fun for the player, fun for the designer, fun for the computer — and notes that this was very clearly an example of fun for the designer.
What he's prouder of, in retrospect, is how he handled it: he was the first person to publicly acknowledge that the quest was terrible and explain exactly why. "It's always easier if you're the first one to go out and say, hey guys, I think I made one of the shittiest quests in the game and here's why." That willingness to be publicly critical of his own work opened the door for the whole team to be more honest about what wasn't working. When the designer admits failure first, it removes the defensive energy that would otherwise shut down that conversation.
六、The Elements of Fun: A Game Designer's Psychology (2:06:00 - 2:45:00)
What Fun Actually Is
Kaplan's analysis of fun is the closest he comes to a unified theory of game design. He rejects the idea that fun is monolithic or that any designer can fully map it from the inside — one of the hardest parts of the job is that designers, like all people, lack self-awareness about what actually motivates them. You might think you're a mastery-focused player but actually be loot-motivated. You might claim you love narrative when you actually spend all your time in cow-level grinding. "We're all different types of gamers, but if you ask me to describe the type of gamer I am, I might be giving more of a picture of the type of gamer I wish I was, or the type of gamer I want you to think I am, versus the type of gamer I actually am."
The elements of fun he identifies:
- Progression: the satisfaction of investment being recognized. Leveling, gold accumulation, gear improvement — these tap into the sense that what I put into this game is coming back to me in the form of a character that's more powerful today than yesterday.
- Mastery: pure skill. The test of whether you can click heads, execute a raid mechanic perfectly, beat the boss with a suboptimal build. This is often what makes competitive games compelling even when the loot is trivial.
- Creativity / Customization: the ability to make something yours. A character that looks like you want them to look. A build that no one else thought of. The whirlwind barbarian that's your whirlwind barbarian.
- Exploration / Story: the desire to see what's over the next hill, to find out what happens next. Kaplan identifies himself primarily as this type — "I'm more motivated by seeing the content the world has to offer" — while acknowledging that exploration often leads him back into loot motivation temporarily, because to see the next content you need gear you don't yet have.
The insight he offers about motivation drift is practically important for designers: "I don't think about what makes this fun. I'm thinking about what makes this not fun. But I'm also watching everyone around me. My wife plays games. My kids play games. Understanding, like, what do they do? What's frustrating? What did they miss?" Design for the player who isn't you, not the player you imagine yourself to be.
The Psychology of Loot Boxes
The Overwatch loot box system was, Kaplan insists, designed by a game designer — not by a commercial person or business executive — and the visceral experience of opening a box was as carefully considered as any combat mechanic. The sound it makes. The way items spill out and animate. The dopamine of the reveal. There is a "lizard brain" component (I see chest, it will feel good, open chest) and a spreadsheet component (is this an upgrade?), and the best designers know how to tap both simultaneously. The loot box connects extrinsic motivation (what did I get?) with intrinsic satisfaction (I earned this by playing).
He makes a similar point about shooter fun. The definition is almost reductive: "Clicking heads." But that simplicity conceals a complex hierarchy. First-person perspective creates an intimacy no other camera angle matches — your hands are in front of you, you are embodied in the world. PvP adds the irreducible tension of knowing the opponent is a real mind trying to defeat you, responding to your moves, adapting, unpredictable. The pure visceral test of skill — can you react faster, aim more accurately, predict the other person's movement — is engaging in a way that PvE (player versus environment) can never fully replicate, because NPCs are fundamentally predictable.
The Multiplayer Director Problem
Kaplan's description of what it feels like to design multiplayer games is one of the most evocative formulations in the conversation: "Imagine if you were going to be a movie director. You've got actors, set designers, props, writers, scripts. Your goal is to get a certain movie made. But you're going to leave the room. You can set it all up ahead of time, and then you're not allowed to be there or talk to anybody involved in it. Now you need the actors to have an experience."
The designer of a multiplayer game sets the stage, writes the rules, crafts the affordances — and then steps back and watches what the players do with it. The most magical outcomes are player stories, emergent narratives that no designer could have scripted. The story of giving Barfa the teleport potion and meeting him in real life. Lex accidentally blowing a townsperson's head off in Red Dead because he confused the greeting button with the gun button and then being chased across the county. "Those are the stories that I think are more interesting from games" than any designed narrative.
This doesn't mean Kaplan dismisses good writing in games — he cites Naughty Dog, Valve's Half-Life 2 and Portal, and Red Dead Redemption 2 as examples of exceptional game narrative. But he calls himself "the anti-shitty story guy, not the anti-story guy." He holds game writing to a high standard precisely because he loves what it can be when done well.
七、Blizzard Polish: QA, Hotfixes, and Studio Culture as Design Tool (2:30:00 - 2:55:00)
What Polish Actually Requires
"Polish" is one of those words that gets used about Blizzard games constantly without much examination of what it actually means mechanically. Kaplan breaks it down.
The foundation is cultural: a studio-wide refusal to accept bugs as an inevitable feature of software. "No one can be satisfied with a bug." The speed at which you fix bugs, and the urgency with which you treat them, reflects how seriously you take the player's time and experience. Blizzard games had bugs — all games do — but the institutional response to a bug was immediate, because the people finding and fixing the bugs were themselves passionate players who were motivated to fix the bugs on behalf of the community they were part of.
The QA department was, in Kaplan's view, the best in the industry — not because of their processes (though the processes were excellent and systematic, far from the cliché of people just playing games all day) but because of their motivation. Blizzard QA was composed of people who deeply loved the games, many of whom wanted to be developers themselves, and who brought the scrutiny of genuine passion rather than professional obligation.
The tactical implementation of that passion: sitting QA team members as close to the development team as possible, encouraging direct communication across the historical chain-of-command barrier. "Some of our QA members knew the game so inside out, you would just say to them: message me anytime, here's my home number, if there's a bug and you think we're going to get raked over the coals on this, you've got to speak up."
A specialist QA member who could snipe at 100 meters in a shooter and feel a single frame of input delay was more valuable than any automated test for catching the subtle feel issues that would make competitive players churn. That person, paired with a graphics engineer, could identify and fix responsiveness problems that most teams would never have caught.
The Engineering of Hotfixability
The other technical pillar of the Blizzard approach: architecting the game to be hotfixable from the start. A hotfix is a server-side patch that requires no client download. The game keeps running while the fix is deployed. For a live game with millions of players, the difference between a hotfix and a required client patch is the difference between a 30-minute response window and a multi-day incident.
The specific scenarios Kaplan describes: a hero in Overwatch has a game-breaking exploit discovered. You disable the hero. The player who mains that hero only plays Overwatch for that hero — they'll leave if the hero is disabled for three days waiting for a patch cycle. A hotfix gets that hero back live in 30 minutes. You've kept a player. You can feel when developers love their product and aren't just putting it on a shelf. Part of that felt love is the speed and attentiveness of the response when something breaks.
The balance dimension: when a build or character becomes obviously dominant — every player is a whirlwind barbarian, no other class is worth playing — that's a failure of tuning that players experience as neglect. The willingness to hotfix balance issues, not just crashes, signals to the community that the developers are still in the game with them, still paying attention, still caring.
八、WoW as Game Director: Server Fires, BlizzCon, and What Success Actually Feels Like (2:45:00 - 3:10:00)
The First Experience of Running a Successful Game
Rob Pardo left Team 2 for StarCraft II shortly after WoW shipped. He put Kaplan and Tom Chilton in charge of WoW. Kaplan's title was still Senior Game Designer — not even a lead title. He was effectively running the world's most popular online game at 28 years old.
"I thought it was totally normal and I thought what we were experiencing with WoW was just normal for making a video game because it was the first video game that I had worked on." The servers were barely running under the unexpected load. WoW had exceeded every subscriber projection — which sounds like a good problem to have, except that the infrastructure had been built for the projections, not the reality. They were hiring database programmers desperately, trying to figure out how to handle a scale of player activity that simply hadn't been anticipated.
The specific texture of that period: working all day, going home for dinner, logging into WoW with his wife for four more hours, going back the next morning to work on WoW again. The entire life was the game. The game master team calling his home phone at 3 a.m. when something went wrong — a faction exploit in Stranglethorn Vale requiring the spawning of Guardians of Blizzard (giant infernals used during beta to keep players out of restricted areas) across the zone, Kaplan whispering his responses so as not to wake his wife sleeping next to him. "And I loved it. I loved the thrill."
Multiple Team Exoduses and Staying Upright
The post-launch period was also defined by people leaving. Two significant groups walked out within the first year after WoW shipped. Carbine Studios formed — taking most of the WoW animators and strong programmers, eventually making WildStar, which took ten years to develop. Red 5 Studios formed — taking the team art director and team lead Mark Kern. Kaplan watched these people leave from his desk, which faced Mike Morhaime's office: "I watched them all go in and quit."
Left behind were the people who believed in the game enough to stay, making patches without animators, keeping the ship afloat with a skeleton crew while simultaneously being demoralized by the departures and overwhelmed by the server issues. "The morale was just in the shitter. Everybody felt very down on Team 2. That we had somehow failed."
Then BlizzCon happened. The first BlizzCon, 2005. The initial announcement got almost no response — the team was reading forums full of server complaints, assuming their perception was accurate. Then Mike Morhaime insisted on putting the tickets on the launcher rather than just the website. Every active WoW player saw the announcement. The tickets sold out instantly.
Kaplan arrived at the convention and encountered something he wasn't prepared for: nothing but love. The same people who had been furious on the forums about server downtime, in person, were overwhelmed with affection for the game. They loved World of Warcraft. They had gathered in Anaheim to be around other people who loved World of Warcraft. They wanted to hear what was coming next and they were grateful for what they already had. The forums had given him a fundamentally distorted picture of the community's relationship with the game.
The Lesson About Online Toxicity
This experience feeds into one of Kaplan's recurring concerns: the structural incentive for online negativity. Social media is designed in such a way that maximum hyperbole gets maximum attention. "If I say, that's a pretty nice mug, I've seen nicer, but I like this one — no one's interested in that. I have to either love this thing, or better, this thing's a crime against humanity in some way."
He acknowledges he helped create this culture, inadvertently, through his early years writing in gaming communities — the hot takes and strong opinions that survived 30 years of internet memory. "I had this reputation for being edgier than I really was. There were a couple notable posts that people like to look back on, but they don't look back on the ones where I'm just being chill."
The consequence he worries most about: creators retreating from the field. He describes Jay Wilson, game director of Diablo 3 — "one of the great design minds" — who took so much online abuse that he effectively retired from making games for over a decade. "We lost the great creative mind of somebody who made a game that millions of people loved." The question is haunting: would Van Gogh have existed if Reddit had been around to comment on his paintings in real time? Would Beethoven have kept composing if thousands of strangers could send him direct messages about every note? The level of access that online communities have to creators is without historical precedent, and Kaplan believes it is actively silencing people who would have created beautiful things.
His plea, especially to young people: be vulnerable. Say you love a thing if you love a thing. The default mode of online discourse — ironic detachment, pre-emptive mockery of enthusiasm — is learned as protection against looking like a nerd. "If you love a video game, if you love Overwatch, say you love it."
九、Titan: Seven Years, $83 Million, and the Anatomy of Failure (3:07:00 - 3:45:00)
The Vision
As WoW was succeeding beyond all expectations, Blizzard leadership began planning for the post-WoW era. The assumption was that WoW would be successful for perhaps five years before aging out, and the studio needed another massively multiplayer game in development. Around 2006, serious planning began for what would eventually be called Titan.
The design vision was genuinely ambitious and, in Kaplan's description, genuinely interesting. The game would take place in near-future Earth. Players would have day jobs — running businesses, building houses, living in a neighborhood (drawing heavily from Animal Crossing, Harvest Moon, The Sims). By night, they would become secret agents with over-the-top shooter abilities. The daytime and nighttime modes were meant to create a dual existence with genuine investment on both sides.
Technically, the game was going to attempt something that WoW had never managed: a single server for the entire global playerbase. No shards. No realms. One world. The game was building out San Francisco (called Bay City), Hollywood, and the geography of California between them, with future plans for Cairo and London. It had full GTA-style driving. A new engine, built from scratch.
The Failure
Titan was canceled in 2013. Kaplan is direct about the causes: "The failure of Titan lies with leadership, team leadership, myself included. There's just no getting around that." He identifies three simultaneous failures: art, engineering, and design — all three at once, which is rare even in failed games. Usually you can point to one discipline as the breaking point.
The art failure wasn't about quality. Some of the best art ever made at Blizzard was made for Titan. The failure was cohesion: "the art looked like it could have come from ten different games." Without a unified art direction that everyone was working toward, brilliant artists in isolation produced individually impressive work that didn't cohere into a world.
The engineering failure was severe: the engine crashed so frequently that in a 40-hour week, a technical artist could only actually work for 20 hours. "Imagine having fighter pilots but not letting them fly." Creative frustration at that scale demoralized some of the best people in the industry.
The design failure was the foundational one: the team never had what Kaplan distinguishes as a vision, as opposed to an idea. He defines the distinction carefully: an idea is "let's make a secret agent who runs a flower shop by day." A vision is an idea plus a technological plan, an art style, a design plan, a production reality, and the ability to inspire a team to believe in all four simultaneously. "Titan kind of was like — that was the hubris of Blizzard in that era. We were now in the era of like, we made World of Warcraft, we can do no wrong."
Anticipatory Hiring: The Root Cause
The single most concrete mistake was what Kaplan calls anticipatory hiring. His own philosophy is to hire nobody until the team is physically working overtime, until not having someone on board is visibly threatening the ship date. Titan did the opposite: over the course of a single year, they hired roughly 70 environmental artists from around the world — before the art style was defined, before the world's rules were established, before anyone knew what those artists should be building.
These were the best people in the industry, attracted by the mystique of working on Blizzard's secret next MMO. They arrived on day one and were told nothing they could hold onto. A guy who flew in from Belgium on his first day asked: "Are we making World of StarCraft?" They weren't. No one could tell him what they were making.
The correct way to incubate a game: start with the smallest possible team, prove out the core idea using whatever art and technology you can borrow or improvise, expand only once you know what you're building. "You're making up busywork. You're just burning cash faster than anybody on the planet."
By 2009, Kaplan knew the game couldn't ship in its current form. By 2010, he went to Mike Morhaime and said: "Shut us down. We're just burning money." Morhaime wouldn't. The studio believed it could be salvaged. It was eventually canceled in 2013, after seven years and $83 million.
十、Six Weeks to Overwatch: Three Pitches and the Right Question to Ask (3:24:00 - 3:55:00)
The Setup
When Titan was finally canceled, most of the 140-person team dispersed to other Blizzard projects — Heroes of the Storm, the Diablo 3 expansion, Hearthstone, WoW. Kaplan fought to keep the core creative group together. Mike Morhaime gave them six weeks to pitch a new game. Two constraints: ship within two years, and have World-of-Warcraft-level revenue potential. Kaplan immediately set aside the revenue requirement as noise: "It's all fucking Monopoly money to me. That's someone else's problem." He focused on the schedule constraint.
The group was demoralized, possibly at the lowest point of their careers. "I didn't know if I was going to be fired. I didn't know if that was the end of my career at that point." They decided to spend two weeks on each of three pitches, with the rule that everyone had to be fully committed during each two-week window — no hedging, no "but the other idea might be better."
Pitch One: StarCraft Frontiers
The first pitch was a StarCraft MMO. Not the standard take — not playing as a Space Marine in the giant RTS armies — but something more intimate. The concept grew from a Chris Metzen sketch of a space prospector: a lone figure on an alien planet, cigar in mouth, foot on a Hydralisk skull, a med-vac silhouetted in the background. Not a soldier in an army. Indiana Jones in space.
"I don't want to be a minion in a giant army. I want to be on the ground — what's it like there? What does a lone prospector find when the giant battles have moved on?"
Arnold Tsang and Peter Lee executed concept art for this pitch that Kaplan describes as pre-order-worthy on the strength of a single image. (That space prospector drawing eventually became McCree in Overwatch — they kept the character, changed the universe.) The design work was serious: class systems, progression mechanics, expansion roadmapping. But the honest assessment was inescapable. Even with a focused team, even if nothing went wrong, this game would take five years to make. With 40 people and a two-year mandate, it was a vision without a path to reality.
Pitch Two: CrossWorlds
The second pitch was a Chris Metzen-driven concept — a meeting point at the edge of the universe, a Mos Eisley spaceport where every civilization in a multiverse had agreed to interact. Seedy, criminal, full of smugglers and diplomats and travelers from wildly different worlds. The neutral planet where the game takes place lets designers hint at a vast multiverse without having to build it: the implied home worlds are more interesting precisely because they're left to imagination.
This pitch was making real progress — class meetings, skill system design (class-based vs. skill-based progression), concept art — when Kaplan had the idea that became Overwatch.
The Origin of Overwatch
During a CrossWorlds design meeting about class systems, designer Jeff Goodman said something offhand that Kaplan couldn't stop thinking about after the meeting ended: "I wish instead of making six classes, we could make fifty classes — and instead of having a hundred abilities on each class, the fifty classes just had one or two things that were really interesting about them."
Kaplan went back to his desk, opened the folder of Titan character designs, and started pulling up Arnold Tsang's art. He began distilling Titan's character concepts: the Jumper class (the one that could blink, recall, use dual pistols — designed after Kaplan's own favorite Modern Warfare 2 loadout, dual G18s) became the clearest version of itself when he stripped it down to essentials. He went to Arnold Tsang and said: "What if this isn't a class? Who is this as a person? What if she's British and her name is Tracer?"
The Gun Jack became Reaper. The Ranger was split and became Soldier 76, and also somehow became Bastion. Each Titan character, distilled to their most essential and interesting capability, became an Overwatch hero. Instead of six classes with a hundred abilities each, fifty heroes with one or two abilities that are really interesting. The idea Jeff Goodman tossed out in a meeting became the design philosophy of Overwatch.
The seven-page deck Kaplan assembled was terrible by presentation standards — all gray with default PowerPoint shapes, "a Jeff deck." Ray Gresko looked over his shoulder while Kaplan was working on it, asked what it was, and immediately said: "Go show Metzen this. This is what we should make." Metzen agreed. The next morning, producer Matt Holly stopped Kaplan in the hallway before he could present it to the team and said he could not walk in there with a deck titled "Monetize Shooter." Kaplan named it Overwatch on the spot — lifting the name from the results of a Titan naming exercise where Overwatch had gotten the most votes before someone else had steered the result to a different name.
Crawl, Walk, Run
The pitch deck's production plan was as honest as possible about the constraints: crawl, walk, run. Ship the PvP hero shooter as the crawl — establish that there's a universe worth caring about. Add PvE co-op content (the walk) once the world has players invested in it. Build toward the full MMO vision (the run) if the audience is there. Kaplan acknowledges that the run stage was aspirational in a way that was partly about saving themselves: "Don't cancel us. This team can make something great." The deck included a mobile strategy — a Photoshop mockup of Arnold Tsang's art on a tablet with the words "and also it'll be on mobile" — because that was required, and because nobody actually believed it.
十一、Overwatch's World and Design Philosophy (3:45:00 - 4:30:00)
A Bright, Hopeful Future
The Overwatch world was a deliberate aesthetic rejection. Most contemporary shooters were gritty, post-apocalyptic, desaturated. Call of Duty, Battlefield, the "realistic military shooter" genre had claimed one end of the market. Kaplan wanted something different: a near-future Earth that was aspirational, colorful, "a future worth fighting for."
The location selection process was one of the most enjoyable parts of development: a group of designers looking at photographs of real places and asking "where do you want to go?" Santorini. A stylized version of Iraq where the game imagined what that region could look like at its most hopeful (the map Oasis). King's Row in London. Hanamura in Japan. "Resist the urge to do the cargo container mazes that you see in every game. We wanted Overwatch to be a world tour of great places that you'd want to go to." The practical design benefit: if players spend hundreds of hours in an environment, that environment should be pleasing, not oppressive.
The heroes were designed with the same aspirational principle. Tracer — the cornerstone, the one on the box — was not just the Jumper class with dual pistols and blink/recall. She was Lena Oxton, a British woman, with a personality and a life implied by the world she inhabits. The backstory of Overwatch lore seeps in indirectly, through environmental details and animated shorts rather than in-game dialogue dumps. "The backstory is implied, told not directly." Players who want to find it can. Players who just want to shoot things can do that.
Reinhardt, the tank, grew directly from Left 4 Dead 2. Kaplan was playing a lot of L4D2 versus mode — playing as the special infected enemies against human players — and the Charger enemy's charge ability stuck with him. Once you press that button, you are a runaway train. The irreversibility is the point. Watching a Reinhardt charge to his death, because the player committed and couldn't stop, is both hilarious and a genuine skill differentiator between good and bad Reinhardt players.
The Matchmaker Problem
The Overwatch matchmaker was designed to maintain players at approximately a 50% win rate. When Lex asks about this, Kaplan's answer is characteristically honest about the gap between what players say they want and what they actually want. Players say they want fair matches. What they actually want is matches where they're slightly better than the opponent — where they win, but it felt competitive. You can't architect that. It's a zero-sum system.
The specific failure of perception: players notice loss streaks and don't notice win streaks. A player who had an eight-game winning streak followed by a six-game losing streak will post furiously on Reddit about the six-game streak. There will be no post about the eight-game winning streak. Kaplan admits to looking up accounts when these Reddit posts appeared: "Yeah, he had the six game losing streak. He had an eight game winning streak before that. There was no post about how awesome it is."
One design decision Kaplan would change in retrospect: Overwatch over-indexed on team outcomes and downplayed individual contribution too aggressively. No scoreboard, a medal system that rewarded both winning and losing teams, maximum emphasis on team wins. Philosophically coherent, psychologically miscalibrated. "If I was making a hero shooter from scratch today, I would actually downplay the team factor and try to put more focus on individual contribution. People are selfish. Not in a bad way. It's just human nature."
十二、Kintsugi, AI, and the Future of Games (4:58:00 - 5:16:00)
What Jeff Kaplan Is Building Now
After leaving Blizzard in 2021, Kaplan has been building a new game called Legend of California with a small studio of 34 people alongside Tim Ford. The studio is called Kintsugi, after the Japanese practice of repairing broken pottery with gold lacquer — making the cracks visible rather than hiding them, finding beauty in the repaired imperfection rather than the original wholeness.
"A lot of me and Tim is in that name. We're so appreciative for our time at Blizzard, but we didn't come away unscarred." The name also carries a design philosophy: games are never going to be perfect, the pursuit of perfection is a mistake, and there is beauty in the imperfection that actually ships. The gold in the cracks.
Legend of California is openly in early access, deliberately rough. The world resets monthly, inspired by Rust's world-reset mechanic — the feature that Kaplan considers Rust's greatest innovation. In a world that resets, a player with 5,000 hours and a new player both wake up on the beach with nothing at the start of each cycle. "If you want to play World of Warcraft with me and I'm level 80 and you're level 1, there's no meaningful experience we can have together. But in Rust, we just wait for a reset and we're both naked on the beach from minute one." The reset creates genuine equality and genuine shared adventure.
The game is not PvP-centered (no losing all your gear to raiders, which would eliminate most of the potential audience), but will have PvP. The design challenge Kaplan is working on: make the reset something players are excited about rather than dreading. "I can't wait for the next reset because the adventure starts all over again."
AI in Game Development
On AI as a tool for game development, Kaplan is measured but skeptical in the current moment. "The current state of AI trying to integrate it into development is mostly a hot mess." The analogy he offers for game-making is someone else's formulation he finds apt: making a game is like making a movie if you had to invent the camera every time. AI could theoretically help with that invention process. In practice, at this moment, it's "overconfident in what it tries to deliver" — right about 1 in 10 times on technical questions, which is not a hit rate you can build a development pipeline around.
His ethical position is unambiguous: no one's creative work should be used by AI without their explicit permission. Voice actors, illustrators, writers — using their work without consent is theft. "That's just immoral. It's no different than just stealing." The line between acceptable automation and unacceptable appropriation is consent.
The use case he does endorse: pure tedium elimination. He needed 2,000 images resized after doing them all wrong in Photoshop. ChatGPT did it in a minute. He wasn't going to hire someone to do that. He wasn't going to stay two hours late to do it himself. "It made my life easier. It didn't take a job." As long as that ethical line holds, AI as a tedium-eliminator is unambiguously good.
The thing AI will never replace: "It's never going to draw a picture like Arnold Tsang. It's never going to tell a story like Chris Metzen. The human spirit is irreplaceable." Lex adds what he thinks is missing from AI creative work — something he calls "the edge that's human," the imperfections, the thing that makes a piece of work feel like it came from a particular mind at a particular moment. Kaplan's response: "AI to me right now, currently, it's like an interesting fever dream."
Small Studios Are the Future
His structural prediction for the industry: small studios are the future of gaming. The large studios will acquire them for new IP, but the genuinely innovative ideas will come from teams small enough to move fast and focused enough to prove the core concept before expanding. His call to game developers is a direct attack on the corporate consolidation he has watched consume the industry: "Own the craft. Own our art form. Stop giving it to these fucking corporate jackholes. You are the golden goose. Keep your eggs."
The Greatest Games
Kaplan's personal hierarchy, outside the games he made: Legend of Zelda: Breath of the Wild is the greatest game ever made. Every aspect is thoughtful and well-designed — art matching design matching technology, combat, narrative, exploration, the world as a toy where everything works the way you hoped it would. "Other games are lucky if they get one of those things right and become best in their genre just for getting that one thing right. Breath of the Wild does them all right and the best."
His reverence for Nintendo is almost religious: "Nintendo is the mecca. That's all there is to it." He doesn't understand how they work. He just worships the results. His son, who also makes games, played Breath of the Wild and came away sad: "I know I'll never make anything like this."
Red Dead Redemption 2 occupies a different slot: pure craft. The Tarantino-level dialogue. Arthur Morgan's arc. The Daniel Lanois song that plays as you ride back, with lyrics for the first time in a game that had been instrumental throughout — and the quotes from Dutch and Arthur Morgan surfacing over that song. "I'm just like: God damn, this is art. I know it's supposed to be entertainment. I know it's a business. But the top of the pyramid is art." Great game stories aren't just technically impressive — they hit you emotionally, the way the greatest work in any medium does.
EverQuest and Rust are his personal extremes: two games he would never recommend to a random person, because they require you to come to them. "Rust will come calling to you if you are up to play it." For the right person, Rust is the most defining gaming experience imaginable. For everyone else, it is incomprehensible.
后记：碎裂的陶器与金色的缝隙
This conversation is ultimately about what it costs to build things that last, and what it takes to survive that cost. Kaplan's path runs through depression, alcohol, hundreds of rejection letters, a dumpster full of manuscripts, shock therapy, seven lost years on a failed MMO, and the chronic paradox of creative work: you're always shipping a lesser version of the game you imagined, and somehow that lesser version is still one of the greatest games in history.
The advice he has internalized and still returns to: focus on what you want to do, not what you want to be. Not "I want to be a game designer" but "I want to make worlds that people can live in." Not "I want to be a writer" but "I want to write things that make people feel something." The doing leads you to the identity, never the other way around.
And the hardest thing, which he got wrong before he got it right: try to make other people's ideas work. The ego that says my idea is correct and your idea is wrong is the ego that kills games and breaks teams and loses the insight that was sitting right in front of you. The designers who succeed are the ones who can hold their own vision while remaining genuinely open to being wrong — who can push when the team isn't being ambitious enough and pull when the team is being too creative to ship.
The scarring is the work. The gold in the cracks is what you learn from it. And the thing you build with that knowledge, imperfect as it is, is the most beautiful version of what was possible.
Lex Fridman Podcast #493 | Published: 2026-03-11 | Runtime: ~5:16:00 Original transcript: ~55,000 words | This essay: ~7,200 words | Compression: ~87%