Education of Tomorrow | ✉️ #15

Illustration for MKDEV Dispatch #15 featuring a stylized person in glasses and a hoodie looking thoughtful with a paper airplane, titled 'Education of Tomorrow'.

Hey! 👋

LEONID SUSCHEV

When we were recording the last episode of the podcast to be released next week, we tackled the important topic of learning in the age of artificial intelligence. To me, a non-programmer, the idea of memorizing commands and logical sequences doesn't seem complicated. After all, conceptually, it's not much different from memorizing algorithms of work in marketing or design. I learned how to draw circles, and then I drew some circles so that it would turn into.... I don't know... a snowman.

This kind of learning process is built from memorizing some basic concepts, and memorizing how, by combining basic concepts, to create something more complex. And the process of increasing my professionalism is to learn and memorize more and more complex combinations. And so far, this learning process has worked for everyone from the barber to the DevOps engineer.

However, now that artificial intelligence has rapidly burst into our lives, this may be changing. I, who am not a programmer, can task an artificial intelligence to create a video game for me by describing the concept. If I want a video game about football, I don't need to remember what commands and lines of code are used to create an object in the shape of a football field. I don't need to know which command to set the field to green. I don't need to remember how to combine these basic concepts to make a football field. I just task the AI to create a game for me in which the human models run on the football field. And the AI will unmistakably reproduce for me a football field that looks like a football field: rectangular, green, with markings. What do I need to learn then?

How to properly formulate tasks for the AI for example. So that it creates for me a field to play ACTUAL football, not the abomination that Americans call football. Or how to manipulate artificial intelligence so that the result is better than that of a competitor using the same AI. How to distinguish a high-quality result from a low-quality one. Remember when we used to learn how to correctly create search queries on Google so that we would get relevant results? Or how to properly ask questions to mentors or senior developers? We even have an article about that.

It seems to me that getting skills like this is what our education is going to be made up of in the near future. Keep that in mind if you're thinking about learning some new technology, because this “near future” is actually today.

What We've Shared

GPT-4 Creates Your Chatbot: We're going to learn how to create a chat bot that is able to read Reuters.com and use the info as an output in the chat with the new GPT-4 version.
What is umoci and how to install it? Let's check what Umoci is and how this tool helps with container bundles. In this video we're only going to install it to use it for real in next videos.
Is AWS SAM better than Serverless framework? Let's deploy Ruby Telegram Bot with it! We're going to try out AWS SAM, an open source tool to deploy and operate AWS Lambda functions. While deploying a Telegram bot for English to German translation, we will compare some of the features with a Serverless Framework.
'DevOps Accents', episode 7: In this episode of DevOps Accents, Leo, Pablo and Kirill talk about all the ways companies choose technologies for their projects, from thoughtful choices to CEOs making spontaneous decisions on the train to work.

***

What We've Discovered

Event-Driven Architectures vs. Event-Based Compute in Serverless Applications: A good explanation of both event-driven and even-based architectures, as well as how AWS Lambda fits in there.
Decouple Deploys from Releases: It's a good idea that releasing new features and deploying code are two separate processes. Lots to think about.
Platform Engineering Is Not About Building Fancy UIs: A great explanation why developer portal and building a platform are not the same thing. Portals are good, but it should not be the main focus.
Spot Instances at Delivery Hero: An overview of how Delivery Hero uses Spot instances in their EKS clusters to cut costs significantly.
Terraform Inmem State Store: We've never thought about using Terraform without a statefile at all, but Kubernetes deployment use case doesn't sound too crazy!

A random reminder

Our free 'AWS Load Balancer Controller 101' Webinar continues to go out every two weeks. It has moved from Saturdays to Wednesdays, but it's still going! Check out the dates in the registration form to attend.

The 16th mkdev dispatch will arrive on Friday, April 14th. See you next time!