Ein Baum und seine Abenteuer



Generative AI is here to stay!


Some reflections on what 2022 showed us about the future.

Generative AI describes AI that takes a prompt and generates something from it. Prime examples include Copilot which takes code or comments as a prompt to complete the code I want to write, ChatGPT which takes a chat message and replies with a human-like answer and Stable-Diffusion (or Dall-E, or Midjourney) which takes a prompt and returns a fitting image. All of these AIs made massive progress in 2022!

"A path into the AI future" - generated using Midjourney
"A path into the AI future" - generated using Midjourney

Text-to-Image AIs like Stable-Diffusion, Midjourney, or Dall-E went from blurry images in which you could make out shapes to producing high-quality results. While Dall-E is best known in the press, Midjourney produces the best results as of today with the open-source alternative of Stable-Diffusion quickly catching up.

Today though these AIs can not replace artists. They don't yet understand well which composition, lighting, and atmosphere work best to convey a meaning. To turn an image in your head into an artwork you need a collaborative process between artists and AIs. But looking at examples of this we can see of these AIs will expedite exploration phases and help artists to venture into different territories.

Lexica.art is a great place to see what's possible and play around with a free account.


Text generation AIs like ChatGPT or Copilot also made tremendous progress in 2022. They went from testing things out to being super useful in production.

Copilot can write code easily. It helps developers enter new languages and prevents us from doing repetitive tasks. Once again it is a collaborative process between AI and humans where I as a developer decide which suggested version best suits my needs and style and where I need to learn how to talk to the AI.

ChatGPT is a personal tutor, an expert advisor on any topic, and just a friendly robot to talk with. Personally, I already asked it to explain subject matters to me, had it generate outlines and sections for blog posts, and help me write a game. And every time I was amazed!

You can try ChatGPT for free at chat.openai.com.


What all AIs to date have in common is the need for Human-AI-Collaboration. To get the best out of Text-to-Image you need an artist, to get Copilot super productive, you need a developer. And I am not sure whether this need for the "knowledgable user" will go away any time soon.

What also became apparent is that all of us need to learn a new skill! We need to learn how to talk with AI. Same as the introduction of the calculator into Math lessons at school, this will ultimately and fundamentally change how we address problems and find solutions.

I hope for a world of free creativity and personal tutors, where humans can focus on being human. On being social and creative beings with machines taking over more and more of the "work". Within AI I see the potential for such a world within our lifetime!

Robot Emoji
Robot Emoji

Author

Hendrik

I am a JavaScript Enthusiast and developer for the fun of it!
Here I write about webdev, technology, personal thoughts and anything I finds interesting.

More about me

Read next

The Gods of Diffusion

The Pantheon of Diffusion
The Pantheon of Diffusion

I have rencently been spending quite some time with text2image generation. And while looking around for great styles for a project I generated a "Chromatic Goddess of Colors" that I really liked. This sparked the idea within me to generate a whole pantheon of gods and see how far I could push this!

Trying my hands at Deno v1.0

Denos awesome Dinosaur logo.
Denos awesome Dinosaur logo.

Following the release of Deno v1.0 I got excited to try my hands at it. These are my first experiences writing a simple tool in Deno.

A super fast introduction to Deno: Deno is the spiritual successor of Node trying to fix design mistakes that were made early on but recognized only late into the project. Deno supports TypeScript out of the box and relies on web-standards. In Deno you can import ES modules from any URL and use fetch like you would in the browser. To help unify the community on processes and workflows Deno provides a wide array of stdLibs and has build in solutions for bundling, testing and code formatting. You can read more in the Deno v1 release post.

Log to Elasticsearch using curl

Good old tools can do amazing jobs.
Good old tools can do amazing jobs.

Once you get a centralized logging solution like Elasticsearch setup you open up an incredible amount of possibilities. That is if you actually send logs to your central place. On a recent project, we wanted to "just log our deploys from the Jenkins Server". Thus we set out to find the "easiest" way to log to our Elasticsearch. We finally reached curl as our solution and today I will walk you through how to do that.