Photo from Anatol Rurac at Unsplash
I've been watching AI educational videos here and there for the past months. I tested OpenAI ChatGPT when it was launched to get the first insights of what AI was able to do but did not feel very interested back then. Over the months, I have also watched several interesting AI documentaries such as AlphaGo and The Thinking Game, as well as reputed interviews to some of the top minds in the space. I specially recommnend this one featuring Godfather of AI Geoffrey Hinton
During all this time, I've been more an spectator, keeping an eye on how the industry is evolving from the fence. I am by no means a coder, although I can understand software architecture and feel familiar on technical conversations. Yesterday I decided to jump into one of the platforms, Google AI Studio, to play around and create a simple App to mint NFTs on an Ethereum based architecture. I wanted to see how deep I could get to create a final product.
I started yesterday afternoon, and I have been immersed in the process until now, more than 18 hours with only a 3 hour break to sleep.
The first step was just use the prompt to request a basic minting app with a backend to manage the administration. A basic minting was created but not published in a real blockchain. That did not surprise me too much, although it was a pretty nice starting result from scratch. The game started getting more and more interesting as I was starting to finetune the whole project requesting to add, change or delete features and finally managed to get all the transactions on Ethereum testnet blockchain funding the transactions from one account to make things easy for users. I published the app and got a url to share with friends to test and give me their feedback.
All this, with a free google account. It is true I had to wait sometime for my credits to increase as I burned them requesting too many things ona short amount of time, but I am impressed on how everything gets created, documented, and errors can be fixed when requesting to do so while interacting with the AI platform.
True you need to know what you are doing, need to understand basic coding and blockchain concepts and architecture, know how to deal with an external explorer to check if transactions are deployed on the chain and request the features on a way they can be well understood, but from a senior coder perspective is as having a team doing lots of work which would have taken hours to do before. I've also been keeping updated copies of the code on a Github account to have the complete history of the development.
Here's a screenshot of the app, while I am still working on it.
After this short initial experience, I've been able to better understand the warnings on how many jobs will change and/or disappear and how important it is to interact with the tools to get to understand them and avoid being left out. New requests keep jumping into your mind as you move forward and once the app can be launched you can also create an analytics dashboard to take a look at how the app use is evolving and also ask the AI to give me its thoughts after taking a look at the created databoard to improve it. That can also be done with several different platforms to see the results you get not just from one but from different ones.
It is clear that AI is here to stay and digital industry is the one which will get disrupted quicker. But every industry will be affected, last week I went shopping with my wife and bought some pants on a shop, Selected them, tried them on and paid for them with no assistance at all, the most amazing part was the automatic cashier, just left the pants on top of the counter, they were identified and the price appeared on the screen requesting for confirmation and payment with the credit card. The whole store is now probably managed by a two people instead of 10.