From Giant Brains to Decompiled Games: Preservation Is the Story of Computing

In 1992, WGBH Boston and the BBC told the story of the computer in a five-part series, The Machine That Changed the World, first shown in the UK as The Dream Machine (Baio, 2008). Its first episode, “Giant Brains”, follows the machine from the human “computers” of the nineteenth century to Alan Turing. In October 2026, a SomeOrdinaryGamers video showed communities using AI-assisted tools to keep games running long after their hardware, and sometimes their studios, have gone (SomeOrdinaryGamers, 2026).

Read together, the two make one argument. The computer’s defining idea is the separation of a machine from the task it performs, which is what lets software outlive the hardware it was written for. But that separation only preserves anything if someone does the work, and the documentary itself nearly vanished before an enthusiast digitised it from VHS. This post follows each video as a chapter, with the video embedded at the top of each; the timestamps jump to the matching moment in that chapter’s video.

Chapter 1: Giant Brains (1992)

The embedded episode is an unofficial upload of the broadcast. The series was produced by WGBH Boston and the BBC and, as of Andy Baio’s 2008 account, had never been commercially reissued.

When computers were people

The episode begins with a reminder of what the word used to mean. In the nineteenth century, a computer was a person who calculated, and the work was slow, tiring and error-prone (watch from 8:00). The schoolteacher William Shanks spent years calculating pi by hand to 707 places, and made a mistake at the 528th. Mathematical tables, compiled by human computers, were riddled with errors that propagated into navigation and engineering (watch from 10:04). Dictionaries, the narrator notes, only switched the definition of “computer” from a person to a machine between about 1935 and 1945 (watch from 19:15).

Babbage separates the machine from its purpose

Charles Babbage, frustrated by errors in tables, wished they could be calculated by steam (watch from 11:04). The difference engine could do one job, so Babbage moved on to a better idea: a machine that could do many (watch from 13:07). The analytical engine separated the “mill”, where operations happened, from the “store”, which held information, and took its instructions from punched cards (watch from 14:07), an idea borrowed from Jacquard’s loom (watch from 15:08).

One of the episode’s historians puts the consequence plainly: for the first time, someone designed a machine without deciding in advance what it was for. What we now call software, the documentary argues, was born in that separation. Ada Lovelace’s published notes on the engine, with their detailed programs, earned Lovelace the title of the first computer programmer (watch from 16:10).

Zuse’s switches

A century later in Berlin, the engineering student Konrad Zuse wanted to avoid tedious calculation and concluded that a flexible machine needed a simple, repeatable element (watch from 20:30). That element was the telephone relay, an on–off switch, and the arithmetic that suited it was binary (watch from 22:31; 23:34). By the end of 1941, with programs punched into discarded film stock, Zuse had a programmable general-purpose computer (watch from 25:38). A proposal to build an electronic version with vacuum tubes was rejected because the German command would not fund long-term projects (watch from 27:38).

ENIAC and the cost of rewiring

In the United States, the bottleneck was artillery firing tables. A single trajectory took a human computer 30 to 40 hours with a desk calculator, and one table needed about 1,800 of them (watch from 29:38). John Mauchly and J. Presper Eckert proposed a machine with some 18,000 vacuum tubes, which experts predicted would fail every few seconds; careful worst-case design made it work (watch from 31:39). ENIAC could perform 5,000 additions a second (watch from 32:40).

Its weakness was programming. Changing ENIAC’s task meant setting thousands of switches and replugging cables, which could take days while an expensive machine sat idle (watch from 36:43). The fix was to store instructions in memory alongside data (watch from 37:43), the design that became the blueprint for later computers.

Turing’s universal machine

The episode closes with Alan Turing, whose 1936 paper showed in principle that a single machine could carry out any procedure that can be written down as steps (watch from 44:47; Turing, 1936). At Bletchley Park, the Colossus machines applied electronic speed to code-breaking, evidence that computers were not merely arithmetic engines, though wartime secrecy hid them for decades (watch from 45:52). Turing went further, arguing that a computer could learn from experience and proposing a test of machine intelligence in 1950 (watch from 49:56; Turing, 1950). Turing died in 1954, after being prosecuted for homosexuality, and did not see the industry those ideas created (watch from 51:56).

A documentary that nearly disappeared

The series itself is a preservation story. In 2008, Andy Baio wrote that it was out of print and had never been released online, surviving mainly on VHS tapes in school libraries (Baio, 2008). With help from Simon Willison, Jesse Legg and, unofficially, the Portland State University library, Baio tracked down and digitised all five parts, shared them for streaming and download, and pledged to take them down if the series returned to print or the rights holders objected. A later note on the post records that the BBC had blocked one of the YouTube copies.

A history of how software was separated from hardware survived only because someone copied it from one obsolete medium to another.

Chapter 2: Everything is about to change (2026)

The universal machine, applied to old software

Emulation is Turing’s insight put to work on history: one machine imitating another closely enough to run its software. The video shows community forks of an experimental PlayStation 5 emulator running Astro Bot and Demon’s Souls on a PC, with broken graphics and slow frame rates, but months after those builds could barely show a title screen (watch from 2:01; 4:02). Mutahar explains that modern emulators increasingly translate a console’s system calls and graphics work rather than simulating every circuit, which is why a mid-range PC can attempt it (watch from 9:04).

Where preservation happens now

The most meaningful examples are games their owners have stopped supporting. A fork of the shadPS4 emulator runs DriveClub, whose studio has closed, at up to 120 frames per second, so the game remains playable on ordinary computers (watch from 13:07). REDriver2, a human-made decompilation of Driver 2 translated from MIPS assembly to C, has become a base that newer, AI-assisted forks extend (watch from 12:06). The video credits AI assistance for the sheer number of decompilation projects now under way, many of which would never have found a volunteer with the necessary skills (watch from 26:13).

The need is real. A 2023 study by the Video Game History Foundation and the Software Preservation Network found that only about 13% of classic games released in the United States before 2010 were commercially available; the remaining 87% were, in the study’s words, critically endangered (Video Game History Foundation, 2023).

Hardware prices and vanishing discs

Mutahar connects the pace of PlayStation 5 emulation to frustration with the platform: rising console prices, with pre-owned units listed above the price of new ones (watch from 8:02), and a move away from physical discs that leaves players with less control over what they own (watch from 6:02). Whatever one thinks of that argument, it echoes 1992: when a machine becomes expensive or disappears, the software written for it is at risk unless something else can run it.

The same caveats

The video does not treat AI as a free pass. Code that its author cannot read should not ship, and experienced developers who audit what a model writes are using it well (watch from 1:00). Unreviewed AI-written builds can carry serious security flaws, so Mutahar runs them on an isolated machine (watch from 14:07), and a legal response to large-scale decompilation is expected soon (watch from 27:13).

The thread between them

The documentary and the video describe the same promise from opposite ends. Babbage’s machine without a fixed purpose and Turing’s universal machine are why software can be separated from the hardware it was written for. Emulation, decompilation and digitisation are what that separation makes possible when the original hardware, publisher or medium is gone.

They also show that the promise is not automatic. Preservation has always depended on labour, from the human computers who checked tables twice, to an enthusiast copying VHS tapes, to volunteers rebuilding a racing game line by line. AI is changing the cost of some of that labour; it does not remove the need for people who understand the code and take responsibility for it.

And they show the role of rights. Baio pledged to withdraw the series if its owners objected; the Video Game History Foundation notes that copyright limits how even preserved games can be distributed. Preservation is most durable when rights holders, archives and communities have legitimate routes to keep work available. Where those routes do not exist, the history of computing suggests that people will improvise.

References

Baio, A. (2008, June 3). The machine that changed the world: Giant brains. Waxy.org. https://waxy.org/2008/06/the_machine_that_changed_the_world/

SomeOrdinaryGamers. (2026, October 6). Everything is about to change… [Video]. YouTube. https://www.youtube.com/watch?v=yGdA7hhwk4o

Turing, A. M. (1936). On computable numbers, with an application to the Entscheidungsproblem. Proceedings of the London Mathematical Society, s2-42(1), 230–265. https://doi.org/10.1112/plms/s2-42.1.230

Turing, A. M. (1950). Computing machinery and intelligence. Mind, 59(236), 433–460. https://doi.org/10.1093/mind/LIX.236.433

Video Game History Foundation. (2023, July 10). 87% missing: The disappearance of classic video games. https://gamehistory.org/87percent/

WGBH Boston & BBC (Producers). (1992). Giant brains (Episode 1) [TV series episode]. In The machine that changed the world. Unofficial upload retrieved from YouTube. https://www.youtube.com/watch?v=hayi9AsDXDo