Introduction - the paradigm shift
A few months ago I went on a surfing trip to Greece. There I met a surfing instructor, Jessie, who taught me a very important metaphor for life.
“It’s like surfing” she said “You wait for a wave, until you see one coming behind you. Then you have a choice: you either decide to try to catch it or you let it go and wait for the next one. But, if you decide to catch it, you have to start paddling as fast as you can and jump up at the right moment. Sure, you can fail and then you have to start again. But either way if you decide to do it you have to work as hard as you can to try to stay on the wave.”
I believe that we are in this moment in history and AI is that wave that is coming up behind us. We can try to catch it or we can wait for the next one. But if you decide to catch it you have to embrace it fully.
10x developers
Have you heard the term 10X developer? ChatGPT describes it like this:
”The term 10X developer refers to the concept of a software developer being ten times more productive than an average developer. This idea is contentious within the tech community, with opinions divided on whether such individuals truly exist or if productivity can be quantified in this manner.”
I personally believe that there are developers who are much more productive than average. People like Linus Torvalds - Creator of Linux and Git, Guido van Rossum - Creator of Python, John Carmack - Co-founder of id Software and the lead programmer on Doom and Quake and many others. (Chances are if you are a BDFL, you probably are a 10x dev)
Unfortunately not many of us can code as quickly and efficiently as those guys. Most of us are, by definition, average. However when you harness tools like ChatGPT, Copilot or Claude for code generation, you can start feeling like a 10X dev.
Granted, you have to know what you are doing. These tools are not perfect, they make many mistakes and sometimes mislead you. In order to benefit from them and minimize risk, you must have an ability to distinguish between good and bad code. But if you DO know what you are doing, then current LLMs are a force multiplier like you have never seen before. You can create a rapid prototype in a fraction of the time it took you before. If you have an idea and you want to see it working - just provide a detailed description to ChatGPT, copy paste it into your IDE and you can see if it holds water. The speed is mind-blowing. I used to do this kind of stuff earlier in my career, but it took much longer and debugging was a frustrating process that slowed me down.
When you are using AI, if you have errors, copy paste those errors into the prompt, you don’t even have to describe your issue, let the AI help you identify and fix the problem since it already has the context of your previous work. It’s pretty amazing.
"Zero coding"
A few years ago, the “zero code” concept came into prominence thanks to companies like Zapier, Flutterflow, Wix and others. On the other hand, PAAS infrastructure allows you to run code in the cloud with minimal effort by abstracting dev-ops and coding tasks to easy to use interfaces where you don’t need to know exactly how it works as long as you have a clear vision of what you need.
Let’s ask ChatGPT what it means:
Zero code, or "no-code," refers to a development approach that allows individuals to create software applications without writing manual code. Utilizing visual programming environments and tools, users can design applications through graphical user interfaces and configuration, democratizing software development by making it accessible to non-programmers. This approach aims to simplify the application development process, reducing the time and cost involved and enabling a wider range of people to innovate and automate processes efficiently.
Well, with the new GenAI tools this concept becomes even more relevant. Now you can develop stuff that you have expertise by yourself, using AI as a force multiplier. On the other hand, for areas where you have knowledge gaps, you can fill them with AI. For example, let’s say you are a full stack developer and you can write frontend and backend comfortably, but you are lacking knowledge in design and dev-ops. You can ask numerous AI tools to fill the gap. Ask DALL-E to generate some assets and ask Copilot to write you a pleasant looking CSS. Then, when you have to run this in the cloud, ask for instructions on how to deploy your newfangled MVP container to Cloud Run. If you encounter errors, copy-paste them into ChatGPT and ask for help.
I recently wrote a solution for one of our clients that involved a FastAPI application with some basic javascript at the front. The solution was supposed to be deployed to the customer’s Wordpress site. Wordpress, as you may know, runs on PHP. The last time I touched any PHP coding was about 10 years ago, so building and deploying a Wordpress plugin should have been fairly complex. However by using AI I got a very specific set of instructions on what needed to be done, right down to what tools I had to download and use, where to upload my plugin files and how to test the solution. In the end, there was a working production code running on the client’s website in half a day. This productivity could not have been achieved just a few months ago.
1 person unicorn
We all know of famous Unicorn companies. Those are the companies that reached a $1 billion valuation. We also witness that time to a $1 billion valuation shrinks as time goes on:
Spotify, which was founded in 2006, became a unicorn by 2011, after about 5 years.
Airbnb achieved unicorn status by 2011, taking around 3 years.
Slack - Launched in 2013, and reached unicorn status in about 1.5 years by 2014.
Wiz, a cloud security startup, became a unicorn incredibly quickly, reaching a $10 billion valuation just two years after its founding in March 2020
Now a question - can there be a “1 person Unicorn?” What does our favorite AI think?
Regarding the possibility of a "1 person unicorn," it's theoretically possible for a startup to reach this valuation with only one founder or employee at its inception. However, as the company scales, it would likely require a larger team to manage its growing operations and development needs, making the concept of a "1 person unicorn" more illustrative of exceptional potential rather than a common occurrence.
As you see this statement is a bit skeptical, especially because AI doesn’t like to give a definitive answer to avoid some recent controversies like the GeminiGate. I would be more bold. I think there is going to be a one or two person unicorn company in the near future. Maybe even this year. Maybe it already happened, we just don’t know it yet.
Guys like Multi-On are making a lot of waves with very small teams. The future is unknown but it is bright.
What is the future?
We already witness productivity levels that we’ve never seen before. Tasks that used to take days or weeks can now be achieved in hours. You just have to know what you are doing. I know that I won’t be hiring a programmer that doesn’t use AI as part of their process. It’s just not worth it.
As you may have noticed this post was written with substantial help from ChatGPT but it was not written by AI. Original ideas don’t come from AI. I think I shall leave this topic for another time.
Comments