Not the End of Software

A while back, there was a Google Docs write-up with the title “The End of Software” that was widely circulated. It went viral as people quoted and reposted it, often sharing their opinions on the matter, either agreeing or disagreeing with the points made. I was one of those who had an opinion on the matter. I wrote a lengthy note as I pondered what was written. That lengthy note, however, did not leave my Obsidian vault, not until today.

I am going to start with an assumption. Since there is no byline on the Google Docs itself, and I can’t figure out from the Google Docs who the author is, I will assume that the original author is Chris Paik. It was his tweet with the link to the Google Docs that people were retweeting, so I assume Chris Paik is the original author.

The write-up made very bold claims about the future of software. Bold claims that, I think, are wrong, which I will elaborate on. Give “The End of Software” a read if you have not already.

The document says that “history tends to rhyme, if you listen.” To understand the future of software, we need to study how technology has changed other industries. The author chooses to compare the software industry to the media industry. I would not make the comparison myself, but let's assume for a moment that software creation is similar to media or content creation.

To sum up the write-up, the author claims media is expensive because content is expensive to create, and therefore it had to make money. The advent of the internet then changed the equation, and now the cost of creating and distributing content is zero. This then creates a Cambrian explosion of content. The traditional media companies that produce their own content, such as newspapers, magazines, books, cable TV, pay-per-view, all struggle to compete with hordes of new content creators. These new content creators could easily produce and distribute content at no cost, probably on social media platforms like YouTube, TikTok, Facebook, Instagram, and others. These traditional media companies then wither and die.

The author claims that the same will happen with software. With the coming of more sophisticated LLMs that can generate better and more efficient code, the cost of software creation will become zero. Just like the Cambrian explosion of content, software will also see its own Cambrian explosion. And that will be the end of software, as the title of the write-up claims.

I see a few flaws in the whole hypothesis proposed by the author. One is the comparison between software and content. While I do see both of them as a form of expression (yes, software can be a form of expression), I see software as being more complex than content. Content by itself cannot produce anything. Software, on the other hand, can. There are software products that produce content, and there are software programs that produce other software. Software can be entertaining, for example, games, but software can also be a utility, something people use. I don’t see this as being the same as content.

Even if we ignore the differences between software and content, the hypothesis still feels flawed to me. The author seems to use the terms ‘media’ and ‘content’ interchangeably, while I think they are different.

The word ‘media’ comes from Latin, which is the plural form of ‘medium,’ something that is ‘in between.’ The modern understanding of the word ‘media’ is that it refers to the multiple ways of communicating content, the plural of medium. And ‘content’ here I will refer to as ideas, stories, or anything that one wants to express or convey. These contents have to be packaged and stored in order to be distributed widely. Newspapers, magazines, books, and so on are the mediums in which content is packaged and stored in a way that it can be distributed widely. This is what I mean by ‘media.’ I wasn’t completely sure that this is what the author meant by ‘media.’

It is important to separate those two, content and media. Content creation is cheap, and it has always been at no cost at all. What is expensive is the storage and distribution of the content. For example, the story you have about an orphaned boy who discovers that he is a wizard and later goes to a school that teaches magic is free. You can tell that story verbally to the person next to you for free. The only cost to factor in is the willingness of that person next to you to spend their time listening to you tell your story.

The cost starts to add up when you want to tell that story to more people. You need a ‘medium’ to store that story of yours, preferably in a form that can be distributed. You’ll need to write the story down in book form. Now, the person you are telling your story to does not need to be sitting next to you. That person can just read it from the book you wrote. The person could be halfway across the world; if your book reaches them, they can still read your story. Paper costs money, printing costs money, shipping the books to bookstores all around the world costs money. This is the part where it gets expensive.

History rhymes, right? Even in ancient times, the oral tradition has been around for communicating content — ideas, stories, or anything that one wants to convey — and it comes at no cost. Socrates, credited as one of the founders of Western philosophy, never wrote a single book. It was his students, Plato, and the like, who wrote his teachings down.

Writing it down, making copies of what has been written, and distributing those copies to multiple locations all come with a cost.

The point I am making, in which I feel the author fails to emphasise, is that content creation is free, but storing and distributing content has always been expensive. This was true before the internet, and it remains true today.

While it may seem to you — as someone who makes content and distributes it only on free platforms like YouTube, TikTok, Instagram, etc. — that there is no cost at all, in reality, someone else is paying for it.

Someone is paying for the initial job of laying down the undersea cables. Someone is paying for the job of maintaining these undersea cables so that someone halfway across the globe could read this long-winded essay from some random guy on the internet, at no cost to the reader, and the author.

Meta, the parent company of Facebook and Instagram, is paying for the cost of storing that digital photo of that prawn olio pasta you had for lunch. They are paying the electricity bills and the people whose job is to keep their servers and data centers up and running. This is so that when someone wants to look at that photo of that prawn olio pasta you just had, they can send the photo to them on their device.

The cost of storing and distributing content is still expensive, and, like the author said, they have to make money somehow. What incentivises companies like Meta (Facebook), Google (YouTube), and the like to keep storing and distributing your content on their platform is that they can distribute content they get paid to distribute next to your content, also known as ads.

When the internet happened, the economics of content storing and distribution changed. The more content these platforms store and distribute, the more ads they can include, and the more money they can make. This is why we see the explosion of diverse content, regardless of its quality.

Traditional media companies mentioned by the author, like newspapers, magazines, books, cable TV, and so on, used to own and control the storing and distributing of content. Either willingly or unwillingly, they relinquished their tight control over the distribution of content to those newer media companies with their digital social networks. Perhaps traditional media thought that the value is in the content, not in the storing and distribution (media) of the said content. This is why they are adamant in keeping the rights to the content they produce. Now we see that the real value is in the storing and distribution of the content. Even low-quality content, if it goes viral, the platform company can monetise the views by placing ads around it.

The point I am making is that the idea that the internet made creating, storing, and distributing content cost nothing is flawed. The cost just shifted to the platform. That being said, the idea that with the advancement of LLMs, the cost of software creation will become zero is also flawed. Again, the cost shifts; someone is paying to keep the machine running to produce better LLMs. Someone is paying to keep the lights on.

If it is free to you, most probably someone else is paying for it, and this ‘someone’ would want to make money out of it.

Lastly, why I feel that the whole premise is flawed, to begin with, is this: the title claimed “the end of software” because the cost of creating software is zero, just like the cost of creating media in the age of the Internet. Even if it is the case that the cost of creating and distributing media is zero (which it is not), it is not the end of media. Content and media are still around. New types of media companies started to appear. Perhaps we’ll see a new type of software company in the future.

With that, I would say, “the reports of the demise of software have been greatly exaggerated.”