How video games grew up: from Pong to the modern gaming industry

Pull up a chair. The video game industry as a thing you can name only goes back about fifty years, which makes it a baby compared to most fields, and that means the entire arc from “weird novelty in a college lab” to “bigger than Hollywood” is small enough to hold in your head. Most of the people who built it are still alive. Some of them still answer email.

The first time anyone described it as an industry would have been the mid-1970s, after Atari’s Pong proved you could put a CRT inside a wooden cabinet, charge a quarter for it, and have people line up around the block. Before that there were a handful of games on university mainframes (Spacewar! at MIT in 1962, OXO and Bertie the Brain a decade earlier, Tennis for Two on an oscilloscope in 1958), and there was Ralph Baer’s Magnavox Odyssey in 1972, which was a real product that real people could buy. But the moment the world clocked that this thing was happening was when Bushnell’s company took a hardware design that didn’t even have a microprocessor in it and turned it into a coin-eating machine that paid for itself in a week.

What follows is the story in four acts. Arcade. Home console. 3D. Modern. Each one redefined what the medium could do, each one ended in a crisis that nearly killed it, and each one left behind technical artifacts and cultural reflexes the industry is still using forty years later. The cybersecurity side gets its own act at the end, because the arc of game security is the arc of computer security in miniature. Every defense the industry has tried, every attack the underground has invented, every economic incentive that turned a hobby into a multi-billion-dollar criminal sector: gaming was usually the first place it played out, because gaming was where the money and the kids both were.

Phase one: the arcade era (1972 to 1983)

#

The arcade was the first place anyone outside a university computer lab encountered a video game. It started with Pong in November 1972, two paddles and a ball on a black-and-white CRT in a wooden cabinet at Andy Capp’s Tavern in Sunnyvale, California. Bushnell’s company put the prototype in for a week as a test; on the second day the bartender called Bushnell because the machine had stopped working, and when Al Alcorn (the engineer who built it) came down to fix it, he found the coin box overflowing, jamming the mechanism. That was the entire market research the industry needed.

What followed was a six-year boom that produced the cultural artifacts everyone over thirty-five still recognizes by sound alone. Space Invaders (Taito, 1978) is the one that lifted the medium from “novelty in a bar” to “national obsession.” Tomohiro Nishikado designed and engineered it more or less single-handedly, including modifying off-the-shelf Intel 8080 hardware because the existing CPUs couldn’t draw bitmap aliens fast enough. The game shipped with what we’d now call a bug-as-a-feature: as more aliens were destroyed, the CPU had fewer sprites to render per frame, the game loop ran faster, and the surviving aliens descended quicker. Nishikado noticed it during testing and decided not to fix it, because the gradual speed-up was the most engaging pacing curve anyone had ever produced in a video game. Players described it as the game getting angry at them. Japan briefly ran out of 100-yen coins.

Pac-Man came in 1980, from Namco, Toru Iwatani’s deliberate design exercise to broaden the audience past the teenage-boy demographic that Space Invaders and Asteroids had captured. He picked a maze chase because it didn’t involve shooting; he picked yellow because food; he reportedly designed Pac-Man’s silhouette from the image of a pizza with a slice taken out, although Iwatani has given different versions of this story over the years and probably the truth involves more pencil-on-napkin and less single-flash-of-inspiration than the legend suggests. The original Japanese name was Puck-Man, which Midway changed for the US release because they were rightly afraid arcade vandals would scratch the P into an F.

The cabinets themselves were the technical Wild West. Some used discrete logic: literally just wired-up TTL chips with no microprocessor at all. Pong was built that way. Others used Z80s, 8080s, or 6502s as the cabinets got more ambitious. The 1980 arrival of the Motorola 68000, the same CPU that would later power the original Macintosh and the Sega Genesis, opened up another tier of complexity. Defender, Galaga, Donkey Kong, Centipede, Tempest, BurgerTime: all between 1980 and 1982, all running on hardware nobody would have called a computer five years earlier.

And then it crashed. The home-console side imploded first, in 1983, dragging arcades down with it. The standard story blames Atari’s E.T. the Extra-Terrestrial tie-in (rushed, terrible, allegedly buried in a New Mexico landfill, the buried part of the story is actually true and was verified by an excavation in 2014), but the deeper cause was the same one that’s killed every consumer hardware boom since: too many low-quality products chasing too small a market, no real quality control, and consumers who learned to stop trusting that the next $50 cartridge would be worth playing. Arcades survived the crash longer than home consoles did, but the genre’s commercial dominance peaked in 1983 and never came back. The kids who’d been spending quarters at the mall arcade started spending them at home on something a Japanese toy company was about to release.

Phase two: the home console era (1983 to 1995)

#

Nintendo had been making playing cards in Kyoto since 1889. They’d done toys, taxi services, instant rice, love hotels. They’d made some arcade games and some early home consoles. By 1983 they had a research division (R&D1, led by Gunpei Yokoi) that had figured out a console architecture cheap enough to be a Christmas toy and powerful enough to do real arcade-quality games, and that July they launched the Famicom in Japan: a red-and-white plastic box, two controllers permanently wired to the unit, a cartridge slot on top. It sold half a million units in two months. They held it out of the American market until October 1985, partly to perfect it and partly because nobody in America wanted to touch the word “video game” after the 1983 crash. The American version got the name “Nintendo Entertainment System” specifically to avoid the toxic phrase.

The thing nobody talks about that explains Nintendo’s success in America is the strictness of the licensing. The crash had been a quality crisis: too much shovelware, no way for parents to tell the good cartridges from the bad. Nintendo’s response was a “seal of quality” you couldn’t get unless your game went through their internal review, plus a chip in every cartridge (the 10NES lockout) that prevented unlicensed games from running on the hardware. Third-party publishers had to play by Nintendo’s rules (limited number of titles per year, manufacturing through Nintendo, royalties up front) or they didn’t get on the platform. The system was deeply anti-competitive and the FTC later opened an antitrust investigation, but it worked: by 1988 the NES had something like 90% of the American market and the post-crash “video games are dead” narrative was over.

What ran on it was the canon. Super Mario Bros. (Miyamoto, 1985) wasn’t the first platformer (Donkey Kong was) but it was the first one where the level design felt like architecture, where every screen taught you a mechanic you’d need three screens later. The Legend of Zelda (1986) shipped with a battery-backed save chip in the cartridge, which let players spend twenty hours in the same game across multiple sessions and changed what a console game could ask of its audience. Metroid (1986) gave the player a continuous map to explore instead of discrete levels, and named the genre we now call Metroidvania (the vania half wouldn’t show up for another decade, with Symphony of the Night, but the architecture is from Metroid). All three were R&D1 or R&D4 (Miyamoto’s team) projects, all three are still being remade and homaged in 2026.

Sega had been Nintendo’s arcade rival for years, but their first home pushback was the Master System (1986), which lost decisively to the NES in America despite being technically capable in some dimensions Nintendo’s box wasn’t. The real war started with the 16-bit generation. Sega’s Mega Drive (Genesis in America, 1988 in Japan, 1989 in the US) used the same Motorola 68000 the original Macintosh ran on, paired with a Z80 for sound. It was faster than the SNES (1990 Japan, 1991 US), but the SNES had Mode 7 (a hardware-assisted background rotation and scaling trick that produced effects like F-Zero’s perspective road), a richer color palette, and the SPC700 audio coprocessor that lets Final Fantasy VI sound the way it does. Nintendo and Sega spent the early 1990s in an arms race of system features, mascot characters, and marketing campaigns that were openly hostile in ways the games industry hasn’t matched since. Sega’s “Genesis does what Nintendon’t” was an ad campaign. Sonic the Hedgehog (1991), designed by Naoto Ohshima with code by Yuji Naka, was a deliberate spite mascot: faster, edgier, blue, with attitude in the marketing that was specifically constructed to make Mario look uncool.

By 1995 the 16-bit generation was over and the next one was already being built. Cartridges had run out of room (a Super Nintendo cart maxed out around 32MB; a CD held 650MB), and the storage gap was the constraint that would define the next era.

Phase three: the 3D era (1995 to 2005)

#

The PlayStation was a divorce. Sony and Nintendo had been working together on a CD attachment for the Super Famicom called the Play Station (two words, hyphen sometimes), and the deal collapsed publicly at CES 1991 when Nintendo announced they were going with Philips instead. Sony, humiliated on stage in front of a watching industry, kept the project running on their own and turned it into a standalone console that launched in Japan in December 1994 and in North America in September 1995. The grudge mattered: Sony’s hardware decisions, their pricing strategy, the developer-friendly licensing terms they offered specifically because Nintendo’s were so adversarial, all of it was about beating Nintendo at the console game. They did.

What made the PlayStation different was the dedicated 3D rasterizer (the GTE, Geometry Transformation Engine) paired with the CD-ROM as the storage medium. Cartridges had topped out around 64MB by the end of the SNES era; a CD held 650MB. Suddenly developers could ship games with pre-rendered backgrounds (the Resident Evil approach), full-motion video cutscenes (Final Fantasy VII’s opening, with its CGI of Cloud disembarking from the train in Midgar, was the moment a lot of people realized the medium had grown up), and digitally recorded soundtracks. The trade-off was loading times, which were new and immediately hated, and which the industry has spent every console generation since trying to make go away.

Nintendo’s response was the Nintendo 64, launched in Japan and North America in 1996, which committed to cartridges out of stubbornness and from an honest belief that loading screens would alienate kids. The N64 had genuinely impressive hardware (the SGI-designed Reality Coprocessor, an analog stick that made 3D character control feel natural for the first time, a controller with three different ergonomic configurations because Nintendo couldn’t decide which they liked) but the cartridge decision cost them third-party support. Square went to Sony for Final Fantasy VII, Capcom went to Sony for Resident Evil, Konami went to Sony for Metal Gear Solid. What Nintendo had instead was their own studios producing what are still considered some of the best 3D games ever made: Super Mario 64 (1996, the game that defined how a 3D platformer’s camera and controls should work), The Legend of Zelda: Ocarina of Time (1998, which sits at a 99 on Metacritic alongside Tony Hawk’s Pro Skater 2 and GTA IV), Banjo-Kazooie (1998, Rare, the studio Nintendo would later sell to Microsoft and never quite get over).

Sega’s Saturn was the third console in the war and the one that lost. The hardware was actually capable, but the architecture was so weird (two CPUs, two graphics processors, a bizarre quadrilateral-based renderer instead of triangles) that developers couldn’t extract its full performance, and Sega’s American launch in May 1995 surprised retailers and developers alike by happening months earlier than announced. Sega never recovered the developer goodwill. They tried again with the Dreamcast (1998 in Japan, 1999 in NA), which was technically excellent and shipped with a modem built in (online multiplayer in 1999, three years ahead of Xbox Live), but the company was already running out of cash and exited the hardware business in 2001 after the PlayStation 2’s launch buried them.

The PC side ran its own 3D arms race. Doom (1993) and Quake (1996) didn’t wait for 3D hardware; id Software’s Carmack wrote them around the limits of software rendering, then around the limits of the early 3D accelerator cards (3dfx’s Voodoo, the company nVidia eventually bought) when those arrived. Half-Life (1998) and Unreal Tournament (1999) raised the bar again. By the early 2000s the GPU was a separate, premium-priced component every PC gamer recognized by manufacturer (3dfx, then ATI and nVidia after 3dfx imploded in 2000-2002). The PC-vs-console framing the industry uses today, where each side has its loyalists and its specific advantages, started in this period.

Microsoft entered the console business in 2001 with the original Xbox, a deliberately PC-architected box (Intel CPU, NVIDIA GPU, hard drive built in) that Microsoft’s internal team had originally pitched as the “DirectXbox” before marketing trimmed it. The Xbox shipped with Xbox Live in 2002, which was the first online console service that worked the way you’d want one to work, and which more or less invented the modern friends-list-plus-matchmaking-plus-voice-chat experience that every console online service has had to clone since.

Phase four: the modern era (2005 to now)

#

The modern era is the one we’re still in, which makes it harder to write about because the boundaries are blurry and the story isn’t finished. Roughly: broadband became universal in developed markets in the early 2000s, the iPhone shipped in 2007 and rewrote what a gaming platform looked like, World of Warcraft (Blizzard, 2004) showed everyone that massively multiplayer was the future, and the Xbox 360 (2005) and PlayStation 3 (2006) generation built online services into the console operating system as a first-class feature. By the early 2010s the industry was bigger than Hollywood, by the late 2010s it was bigger than Hollywood and music combined, and by 2026 the global gaming market is sitting somewhere around $250 billion a year depending on whose numbers you trust.

What defines the era technically is the convergence of always-on connectivity, photorealistic graphics, and the smartphone as a parallel console. World of Warcraft peaked at about 12 million subscribers in 2010 and proved that a virtual world could hold real people’s attention indefinitely if it was rich enough. Minecraft (Markus Persson / Mojang, 2011) became the best-selling video game of all time by units (around 350 million as of 2026), partly because it works on everything from a phone to a high-end PC and partly because the construction-sandbox model is genuinely timeless. Fortnite (Epic, 2017) and PlayerUnknown’s Battlegrounds (PUBG Corp / Bluehole, 2017) invented and then mainstreamed the battle royale, a genre that previously existed only as a mod for ARMA. Among Us (Innersloth, 2018, but global only in 2020 during the pandemic) showed that a small four-person studio could build a phenomenon entirely on word-of-mouth.

Mobile turned out to be its own industry. The iPhone in 2007 created a platform that didn’t exist before, and the App Store in 2008 gave developers a distribution channel that bypassed publishers entirely. Angry Birds (Rovio, 2009) was the first canonical mobile-only hit. Candy Crush Saga (King, 2012) productized the free-to-play model, free download, currency-purchase loop, ad-supported plus IAP, that now generates more revenue than the entire console industry. The gacha mechanics built into games like Genshin Impact (miHoYo, 2020) and Honkai: Star Rail (HoYoverse, 2023) are why the industry is having a serious regulatory conversation about whether the loot box is a slot machine. (Belgium, the Netherlands, and increasingly the UK and Australia have decided it is. The US hasn’t, yet.)

VR and AR landed in the mid-2010s with the Oculus Rift (2016), the HTC Vive (2016), and the PlayStation VR (2016). VR has been a real but smaller market than the early hype suggested; Meta has spent over $50 billion on the Reality Labs division and the consumer-VR side still hasn’t broken into the mainstream the way console gaming did. AR had its consumer moment with Pokémon Go (Niantic, 2016) and has been waiting for the killer headset since. Apple’s Vision Pro (2024) is the most expensive single attempt at making AR mainstream and is selling at a pace that suggests it’ll take another generation.

Esports became a real industry in the 2010s, with League of Legends (Riot, 2009), Dota 2 (Valve, 2013), and Counter-Strike: Global Offensive (Valve, 2012) at the center. Twitch (founded 2011, acquired by Amazon 2014) gave it a broadcast medium; The International (Dota 2’s annual championship) has paid out single tournament prize pools above $40 million. The 2024 League of Legends World Championship final in London sold out the O2 Arena, which is the same venue that does Adele residencies.

The dark side of all of this, gambling-shaped monetization aimed at children, the rise of in-game harassment as a category of social problem, the labor practices in AAA studios that have produced an unprecedented wave of unionization, the environmental footprint of always-online infrastructure, the way the entire industry has consolidated under three or four publishers (Microsoft acquired Activision Blizzard for $69 billion in 2023, the largest deal in tech history at the time), is the part the next decade is going to spend untangling. Gaming is no longer a fringe culture. It’s the dominant interactive medium and it carries all the responsibilities and contradictions that comes with.

The cybersecurity side, as a parallel history

#

The thing nobody planned is that the gaming industry would become the largest single category of organized cybercrime by transaction volume. Account theft, currency markets, DDoS-for-hire (originally called “booters” by the kids who used them to knock rivals off Xbox Live), data breaches, ransomware against studios. Every major attack class has its gaming-side history, often years before the same technique started showing up in enterprise compromise reports.

Phishing for game accounts has been around since EverQuest in 1999. The pattern is the same as corporate phishing, just with different bait: “your account has been flagged for review,” “click here to claim your exclusive item,” “your friend sent you a trade request, log in to accept.” Steam accounts in particular get phished at industrial scale, because Counter-Strike 2 cosmetic items can carry four-figure real-world prices on third-party marketplaces. Riot’s anti-phishing team and Steam’s account-recovery flow are the same kind of operation as a bank’s fraud department, and they handle comparable transaction volume.

Malware bundled with cheats and mods is the single most reliable infection vector for consumer Windows in 2026. A kid downloads what they think is a Fortnite aimbot, runs the installer, and ten minutes later their Discord token, browser cookies, and saved passwords are being exfiltrated to a Telegram channel run from Russia. The info-stealer market (RedLine, Vidar, Raccoon, Lumma) lives off this. The “cheat” sometimes even works, which makes the infection cycle longer.

DDoS attacks were a gaming problem before they were an internet problem. Xbox Live and PSN took regular hits through the 2010s; the Lizard Squad attacks against both networks on Christmas Day 2014 are the canonical example. The Mirai botnet, which became the largest DDoS source the internet had ever seen and brought down Dyn (and most of the US east coast’s connectivity) in October 2016, was built specifically to take down Minecraft servers, the authors had a competing Minecraft DDoS-protection business and were trying to drive customers to it. They got prosecuted and the case set the modern precedent for botnet sentencing.

The cheating arms race is its own subdiscipline. Aimbots, wallhacks, ESP overlays, and recoil-elimination scripts have been in every major competitive shooter from Counter-Strike 1.6 through CS2, Call of Duty, Valorant, and Fortnite. The anti-cheat side has progressively escalated from user-mode hooks to kernel drivers (Riot’s Vanguard, Activision’s Ricochet, Easy Anti-Cheat, BattlEye) that are functionally indistinguishable from rootkits, raising privacy and security concerns the industry hasn’t fully reckoned with. The cheat side has responded by moving to external hardware: a $400 device that sits between the game console and the monitor, processes the HDMI output to identify enemies on screen, and feeds movement commands back through a USB-connected controller emulator. Software-based anti-cheat can’t see hardware-based cheats. The arms race continues.

Data breaches against studios have produced some of the largest data leaks in history. Sony’s 2011 PlayStation Network breach exposed 77 million accounts and took the service down for 23 days. The 2020 leaks of Cyberpunk 2077’s source code (CD Projekt Red, REvil ransomware), the 2022 leak of GTA VI gameplay footage and source code (Rockstar, attributed to the Lapsus$ group), and the 2023 leak of Insomniac Games’ entire Spider-Man 2 development infrastructure (Rhysida ransomware) all happened during ransomware extortion against the studios. The economic incentive is exactly the same as enterprise ransomware: pay or your IP gets dumped.

Ransomware against studios is the newest category. Capcom (2020, Ragnar Locker), CD Projekt Red (2021, REvil), Bandai Namco (2022, ALPHV/BlackCat), Insomniac (2023, Rhysida), and Squarespace’s Squarespace acquisition target Square Enix (multiple) have all eaten major ransomware events. The studios are uniquely vulnerable because they’re under release-date pressure that ransomware operators understand perfectly: every day of delay before launch costs them millions in projected sales, which means the calculus on whether to pay the ransom is often more brutal than for a normal enterprise. Some studios pay. Some publish the source code rather than pay. Neither outcome is good.

The lesson the rest of the security industry is finally learning from gaming: where there’s a real-money market, there will be theft; where there’s competition, there will be cheating; where there’s social trust, there will be social engineering. Gaming had to figure all of this out twenty years ago because it had real-money markets, competition, and social trust before the rest of the internet did. The mitigations the gaming industry built (account two-factor with hardware tokens, kernel-mode anti-cheat, behavioral fraud detection on transactions, automated bot-detection on chat) are the ones the enterprise security industry has been quietly copying ever since.

Technical tidbits

#

A handful of the architectural and programming-trick details that made these games possible, in roughly chronological order:

1. Discrete-logic arcades. Pong and its immediate successors didn’t have CPUs. The whole game ran on a few dozen TTL chips wired together so that the physics and rendering were emergent behavior of the circuit. Space Invaders used an Intel 8080 (2 MHz), a real microprocessor, but earlier cabinets really were just hardware.
2. The Space Invaders speedup. Not designed; the Intel 8080 was too slow to render 55 aliens at full frame rate, so the game slowed down with more aliens on screen and sped up as they were destroyed. Tomohiro Nishikado noticed during testing and decided not to fix it. Players experienced it as the game getting tense as they got closer to winning.
3. NES color palette. A 54-entry master palette in the PPU, of which 25 could appear on screen simultaneously (one universal background plus 12 background palette entries plus 12 sprite palette entries, with transparency overlap). Hard constraints that shaped how the games looked.
4. NES sprite limits. 64 sprites total per frame, 8 per scanline. Games used “sprite multiplexing” (cycling which sprites were enabled on which scanlines, fast enough to fool persistence of vision) to fake more.
5. PSG vs FM synthesis. Early consoles used Programmable Sound Generators producing square, triangle, and noise waveforms (the NES’s APU is the classic example, the sound of the entire 8-bit era). The Sega Genesis added Yamaha’s YM2612 FM synthesis chip, which is why Genesis music sounds like Streets of Rage 2 and NES music sounds like Mega Man 2, same era, different chip family.
6. Cartridge ROM versus CD storage. Cartridges mapped game code directly into the console’s address space, so loading was instant. CDs held 100x more data but had access latency in the hundred-millisecond range, which is why the PlayStation generation introduced the loading screen as a cultural artifact.
7. Fixed-point arithmetic. Pre-PS2 3D consoles had no FPU; vertex coordinates were stored as fixed-point integers, with arithmetic done in integer math and the binary point treated as implicit. This caused “polygon wobble” (vertices snapping to the nearest integer position) and is why the PS1’s geometry looks distinctively jittery to anyone who played it.
8. Z-buffering. Solving the visibility problem (which polygon is in front of which) by storing a per-pixel depth value alongside each color. The N64 had hardware Z-buffering; the PS1 didn’t, which is why its games used painter’s-algorithm rendering and you can see polygons popping in front of each other in the wrong order.
9. Perspective-correct texture mapping. Early hardware applied textures with linear interpolation across the polygon (cheap, fast), which warped textures on surfaces seen at an angle. The PS1 used the cheap version; the N64 had perspective correction; this is why N64 floor textures look right and PS1 floor textures swim.
10. Mode 7. A SNES PPU feature that let a background layer be rotated and scaled in real time, producing pseudo-3D effects without actual 3D math. F-Zero’s perspective road and Super Mario Kart’s view are Mode 7 backgrounds being rotated and scaled per-scanline.
11. Parallax scrolling. Multiple background layers moving at different speeds to fake depth in a 2D game. Universally used through the 16-bit era; the Sega Genesis had hardware support for two parallax layers, which is why Sonic the Hedgehog feels like it has depth.
12. The Konami Code. Up, Up, Down, Down, Left, Right, Left, Right, B, A. Created by Kazuhisa Hashimoto in 1986 for the NES port of Gradius as a debug cheat to give the player full power-ups, because Hashimoto couldn’t beat the game without it and needed to test late levels. He forgot to remove it before shipping. The code has appeared as an Easter egg in dozens of subsequent Konami games and as a meme in everywhere else.
13. Antialiasing. Smoothing jagged edges (jaggies) by sampling pixel colors at multiple sub-pixel positions and blending. Multisample antialiasing (MSAA) was the default through the 2000s; modern games use temporal antialiasing (TAA), which reuses information from previous frames.
14. Real-time ray tracing. Simulating light by tracing rays from the camera through the scene. Existed in offline rendering for decades; NVIDIA’s Turing architecture (RTX 20-series, 2018) was the first consumer GPU with hardware ray-tracing acceleration. By 2026 ray tracing is in most AAA console titles, with quality scaling depending on the GPU.
15. Game engines. Unreal Engine (Epic, since 1998) and Unity (since 2005) are the dominant general-purpose engines; Frostbite (DICE/EA), Source 2 (Valve), and Decima (Guerrilla) are the major proprietary ones. Engine choice is a major capital decision; rebuilding a studio’s tech stack is years of work.
16. Netcode and client-side prediction. Modern multiplayer games hide network latency by simulating the player’s actions locally as soon as input is detected, then reconciling against the server’s authoritative state when the round-trip completes. Done well, players don’t notice; done poorly, you get the rubber-banding that ruined Halo 3’s online ranked.
17. Hitboxes versus hurtboxes. The invisible geometry that registers an attack as having landed (hitbox) is separate from the geometry that registers receiving damage (hurtbox). They rarely match the visual character model exactly, and fighting game tournament players spend hours studying frame-data charts that show exactly which frames have active hitboxes.
18. Scanline rendering tricks. CRT TVs drew the screen one horizontal line at a time. Older consoles let the game change the palette mid-frame using “raster interrupts”, a hardware interrupt that fired between scanlines, letting you change the palette before the next horizontal line was drawn. This is how the SNES displayed more than 25 colors on screen.
19. Dithering. Two alternating-pixel colors that the eye blends into a third color, expanding the effective palette. The Sega Genesis used this heavily because its native palette was small; the PS1 used it on transparency. Modern HDMI scalers blur the dithering pattern, which is why some retro purists insist on CRT displays.
20. Blast Processing. A Sega marketing slogan that actually had technical content: the Genesis’s DMA controller could transfer data to the video processor faster than the SNES’s could, which let games push more sprite data per frame. The slogan was hyperbole; the underlying capability was real.
21. Normal mapping. Texturing a flat polygon with a normal map that lies about the polygon’s surface direction, producing the illusion of bumps and grooves without actually adding geometry. The technique made the early-2000s leap to “this looks photorealistic” possible without melting the GPU.
22. Occlusion culling. Not drawing what the camera can’t see. Sounds obvious; doing it correctly without occasionally drawing nothing because the algorithm got confused has cost the AAA industry billions of dollars in QA over the decades.
23. Procedural generation. Generating content algorithmically instead of authoring it by hand. Elite (1984) had eight galaxies of procedurally generated planets in 22KB; Minecraft uses Perlin noise to generate effectively infinite terrain; No Man’s Sky (2016) generates a galaxy of 18 quintillion planets and is the canonical case study for both the technique’s appeal and its limits.
24. Middleware. Third-party engines for specific subsystems. Havok for physics. SpeedTree for vegetation. Wwise and FMOD for audio. Most AAA games license multiple middleware stacks; the “engine” most games are credited to is really the integration shell around six or seven licensed components.
25. The Doom engine (id Tech 1, 1993). Wasn’t actually 3D. Used 2D sectors with variable floor and ceiling heights, ray-cast for visibility, and sprites scaled by distance. You couldn’t have a room above a room. This is why Doom runs at 60fps on a Pringles can and Quake (which was true 3D) needed a Pentium.

Trivia

#

The kind of things you tell people at parties to see if they’re real gamers, with the actual history rather than the version everyone repeats:

1. Sonic’s “Mr. Needlemouse” internal name is a literal English compound of the Japanese “Mr. Harinezumi” (hari = needle, nezumi = mouse, which together just means hedgehog). The “Needlemouse” version is a translation artifact that became gaming folklore; Yuji Naka has said in interviews he doesn’t recognize it as an official codename. The pre-Sonic prototype really was a rabbit, though, also designed by Naoto Ohshima.
2. The first playable female protagonist is not Ms. Pac-Man (1982). It’s Lady Bug (Universal, October 1981), and Billie Sue in Wabbit (Apollo, Atari 2600, 1982) has a competing claim. Ms. Pac-Man came later than both and the “first female protagonist” attribution is just a popular misremembering.
3. The best-selling video game of all time by units is Minecraft, around 350 million as of 2026. “Highest-grossing” is more contested; Grand Theft Auto V has higher lifetime revenue per most estimates because Shark Cards and microtransactions add up faster than $30 game sales.
4. Charles Martinet (voice of Mario, 1990 through 2023) had never heard of Mario or Nintendo when he walked into the audition in 1990; he saw the audition listed on a flyer, asked “what’s a Mario?”, and improvised the voice that defined the character for three decades. The folklore that he’s never played a Mario game is incorrect; he’s named Super Mario Maker as a favorite.
5. The first video game with recognizable human characters is generally credited to Western Gun / Gun Fight (Taito/Midway, 1975) with its pixel cowboy duelers, or Heavyweight Champ (Sega, 1976), which was the first fighting game and also the first to depict a Black character. Spacewar! (1962) had spaceships, not people.
6. The Xbox was internally called “DirectXbox,” one of three codenames Seamus Blackley’s team used (alongside “WEP” and “Midway”). Marketing tested the shortened “Xbox” against a set of more “appalling” alternatives and it won.
7. The highest Metacritic score is 99. The Legend of Zelda: Ocarina of Time (1998) holds it, but so do Tony Hawk’s Pro Skater 2 (2000) and Grand Theft Auto IV (2008). The “only game ever with a 99” claim is wrong by a factor of three.
8. Lara Croft’s design evolution. Toby Gard’s original concept was a male protagonist with a whip and hat (Indiana Jones-derivative, which is exactly what his producer told him). He cycled through five designs before settling on a female character, briefly named “Laura Cruz” (a braided South American adventurer), then renamed Lara Croft by Vicky Arnold using the Derby phone book.
9. The original PlayStation was indeed “Play Station” with a space, when it was a Sony-Nintendo joint project for a SNES-CD attachment. After Nintendo backed out and Sony went solo in 1993, the space closed up to “PlayStation.”
10. The first CD-based console was the PC Engine CD-ROM² (NEC, Japan, December 4, 1988). The “TurboGrafx-CD” was the US name for the same hardware, but it didn’t ship until August 1990, so the 1988-first-CD-console honor goes to the Japanese version under its original name.

Closing

#

Fifty years in, the medium is still the youngest popular art form, and it’s the only one where the audience interacts with the work in real time. The four-phase arc above (arcade, home console, 3D, modern) is the technological story; the cultural story is that a thing started by engineers in university labs trying to see what a computer could do for fun became the dominant interactive entertainment for two generations of humans and the largest single industry of organized cybercrime by transaction volume. Both of those things are true and they grew up together.

The next phase is already being argued about. Cloud streaming hasn’t killed the local console the way some people predicted in 2020; VR hasn’t taken over the way Meta wagered $50 billion that it would; AI-driven content generation is moving fast and nobody knows where it lands yet. Whatever happens, the kid playing Roblox in 2026 is using technology and following social conventions that descend directly from a quarter dropped into a Pong machine in 1972. The cabinet’s gone. The play is the same.