This weekend I've got to play with Midjourney's...

This weekend I’ve got to play with Midjourney’s last version, v4.

The improvements are amazing 🤯, generating images on the text inputs (or “prompts”) provided to the system.What fascinates me the most is the usability of the system itself.

Talking about using a bleeding edge AI/ML system would usually require some deep technical expertise, dedicated last-generation hardware… but that’s not the case with Midjourney v4.Using it is as simple as interacting with the typical customer-support chatbot that helps you find a missing e-commerce order: you install Discord, join the beta (https://lnkd.in/ezMdB9Gw), send your prompt with a command ("/imagine …"), and you’re in.The bot will use the servers to generate 4 images, that once received can be upscaled, remixed to iterate on some of them - and that’s it!

All the complexity and technical details of the Generative Adversarial Network are hidden, “commoditised” as a “magical” system where your imagination is the limit (quite literally).

You can even see the images improve in quality, as the system finishes it generation.What I find quite fascinating is that it feels like a peek to the future of work.

Playing with MJ has given me a sense of “wow” that reminded me to GitHub’s Copilot system - that focused on writing code, with the prompt being contextual (the content of the file that you’re working on and programming language).I think that the more we evolve the usability of the AI/ML, the more that they might become less “the thing itself”, and they will become part of the toolchain that we use to work - optimizing our work, allowing us to scale, add elements of creativity…

But who knows what the future will bring us!#ml #midjourney #gan #generativeai