How much money does it take to start a website?
It turns out, not much. A domain will run you $12/year from Google. Webhosting on Google App Engine or Amazon Web Services is free to start — they only cost money once you start getting more users. And if you know how to code, that’s pretty much it — $12 per year.
But let’s say you have no technical background — not a problem. Setup a Wordpress or Wix website for FREE. No coding required, granted the options are limited. But in less than one hour, you are online. And don’t worry you are in good company. Wordpress powers over 26% of ALL websites.
For a little extra money you can get more themes and functionality from companies such as Weebly, Squarespace, or Bluehost to name just a few. I am talking less than $10/month on average. So for the price of two Venti Iced Caramel Macchiatos per month you can run a respectable website with little to no technical experience. Want this site to sell the products you have? Shopify and Stripe have you covered with monthly rates or small transaction fees. This is the beginning of what we shall call Coding as a Service. We are starting to take this for granted in 2017, but even just 10 years ago — most people would have thought you were crazy.
As a reminder here were the phones we were using in 2007.
History Lesson. The first server Google owned in 1998 was the Sun Ultra II. This cost between $16,500-$60,000 new depending on the specs. That was extremely cost prohibitive, especially for startups. But that was 20 years ago, surely the price has gone down. Well, yes — but for a large company running servers it’s still expensive, which the table below from Sherweb conveniently shows for 2015. Sherweb makes usage and spec assumptions (2 vCPUs, 8GB of RAM, 512GB of disk storage) in their analysis, so this isn’t one-size-fits-all assessment, but you’ll get the idea. One note, the year 1 and 5 costs are much higher due to server upgrade cycles.
In 2006 Amazon was the first major tech company to realize the cost and headache of servers was a lucrative business opportunity, and they began hosting servers with Amazon Elastic Cloud Compute or EC2 for short. This changed the world. Today large organizations and startups like Snap(chat) both are running on Infrastructure as a Service (IaaS) known as the cloud. In the future this means that computers won’t need to have much compute power at all, because calculations will all run in the cloud. Devices will evolve to be smaller and more connected. This is what is paving the way for the Internet of Things revolution, Amazon Echo anyone?
One further step in this evolution is Coding as a Service (CaaS) — just like IaaS before it. The current web builders all rely on human programmers to make the templates for us. This is time consuming and labor intensive, but soon that will no longer be the case. Artificial Intelligence will enable more complex combinations of code to be fit together through an automated process. Think of the code behind websites as puzzle pieces. With current technology website builders are creating, let’s say, 100 piece puzzles. There is some complexity, but still they are pretty basic in their granularity. They are drag and drop or formulaic templates. Now when AI gets integrated into the mix, these website builders will essentially be creating 10,000 piece puzzles. The quantity and quality of customization with AI will be staggering. And for the end user this means it will be at once simpler and cheaper.
But Let’s see this in practice…Imagine speaking to an AI whose sole job in life is to build websites. The AI has natural language processing, so it understands your spoken intents perfectly. The AI also has instantaneous and tireless coding abilities coupled with a flawless understanding of functionality. So the AI will build and change your website right before your eyes, the limiting factor being how quickly you can say you want the background color to be Razzle Dazzle Rose. The end result, a website that looks and acts exactly how you want it to, all by just using your voice.
As you probably have realized by now, CaaS via AI will not be limited to people making ‘interesting’ websites. Ultimately CaaS will drive down the time and money necessary for new capabilities in spaces from enterprise to social. But in the near term CaaS will affect programmers who complete relatively simple or scripted tasks, like querying databases for example. No longer will we need armies of coders who expertly understand how to fit if else statements within loops, whether to use Ruby or Angular. AI will be able to do this much faster, with fewer errors, and a greater understanding of system requirements.
So in short, software will continue to run the world, but fewer and fewer people will build the software.