Midjourney v6: bridging realism and text in AI-generated images

Pivot 5: 5 stories. 5 minutes a day. 5 days a week.

1. Midjourney v6: bridging realism and text in AI-generated images

Midjourney v6 marks a significant advancement in AI-generated imagery, offering improved realism and text creation. Users can now specify more details in prompts, control composition through natural language, and add frames or borders. While the update has slowed down the generation process, the overall quality has dramatically improved.

Midjourney v6's ability to generate more realistic figures like Einstein and Curie shows a substantial leap forward. Additionally, its new text capabilities, despite needing refinement, represent a notable improvement over its predecessor.

Read the full story here 

2. AI diagnoses autism with remarkable accuracy, Korean study shows

A Korean research team has made significant strides in employing artificial intelligence for the accurate diagnosis of Autism Spectrum Disorder (ASD) in children. The AI system, trained to detect retinal patterns indicative of autism, was tested on 958 children and successfully identified every child with ASD without false positives. 

While it provided a numerical spectrum score, its precision varied compared to standard assessments. The study, a substantial leap in AI medical diagnostics, underscores the necessity for further research to refine the tool's effectiveness for younger children and enhance its generalizability.

Read the full story here 

3. Alphabet sets election query limits for bard and AI search

Alphabet's Google announced plans to restrict the types of election-related responses its chatbot Bard and AI-based search can provide ahead of the 2024 U.S. Presidential election and other significant global elections. Set to commence in early 2024, this move aims to mitigate the spread of misinformation and enhance the integrity of election-related information.

This approach aligns with broader industry trends, as Meta also imposes restrictions on AI-generated political content, and X revises its political ad policies. These measures reflect the growing need for stringent AI regulation to combat the challenges of misinformation and maintain the reliability of digital platforms during pivotal election periods.

Read the full story here

4. AI predicts mortality: Danish researchers' groundbreaking model

Researchers from the Technical University of Denmark have developed an AI model called "life2vec," capable of predicting various life outcomes, including the approximate date of death. Published in 'Nature Computational Science,' the study utilized health and labor data from Denmark's six million population to build the transformer model.

Life2vec translates inputs like birth location, health status, and occupation into outputs predicting outcomes from early mortality to personality nuances. While the potential accuracy and ethical implications of predicting life events, especially death, are significant, the model opens new discussions on the use of AI in understanding and possibly foreseeing the trajectory of human lives.

Read the full story here 

5. ChatGPT now allows to archive chats

The feature is out now on the web and iOS, and OpenAI says it’s coming to Android “soon.