Midjourney V6, Images Directly from the Brain, and Life-Saving AI: Top AI News of the Week

#1 — AI Generates Typical Phone Photos

— Let's start with news from the end of last year: Midjourney released the sixth version of its model. It's still in the alpha version, but it’s making a lot of noise. First, the realism of the generated images has significantly improved. Now, you can generate not just stunning studio portraits but also typical phone photos. Second, we are eagerly awaiting the web version — goodbye, Discord! I think this will increase the user base exponentially. Third, the model creates perfect images in the style of... anyone! Scenes from "Rick and Morty," "Dune," "Cyberpunk 2077," etc., are near-perfect replicas. Actually, from a copyright perspective, this should be prohibited. Let's see if the company faces any legal issues.

#2 — Scientists Convert Images Directly from the Brain

— I believe that almost everyone has tried generating images with Midjourney, DALL-E, or Stable Diffusion. Many might have given up because it can be challenging to transform a mental image into a textual prompt. You can't just explain to the model exactly what you want from it. There are several ways to solve this problem. You can extensively practice writing these prompts. You can also take an AI course from aw.club and learn prompt engineering there. The most dramatic option, however, is to simply wait. Scientists are very close to extracting images directly from your brain! Researchers at Osaka University in Japan recently generated 1200 images based on the brainwaves of the research subjects. Yes, the images were distorted, with an accuracy of about 50%, and an fMRI machine is required. But we are clearly stepping into a world where thoughts can be read, and dreams recorded.

#3 — AI Saves Lives

— If the previous news left you a bit unsettled, let's balance it with something positive. It's the beginning of 2024, and we are coming off a groundbreaking year for AI and other technology developments. More than 80 different AI models were released in the field of healthcare last year, for example. Medication selection, interpretation of lab tests, discovery of new vaccines and antibiotics, diagnosis prediction with limited data — all of these potentially lifesaving developments are already in production. And in 2024, we can expect to see even more.

So, AI can not only create fakes, it can also literally save lives. In 2023, it became clear that very cool and powerful tools are now available to almost everyone. In 2024, we need to learn how to use them.