Welcome to this edition of
The AI Reverie 

The best newsletter to get a short, concise overview of the recent AI- and tech-news.

The Gentleman’s agreement:
We don’t need donations or gifts of any kind, all we ask from you, dear reader, is that you open each email and click at least one link in it. This is all we ask for. In return you will get the best content we have to offer.

Thank you - Now let’s get to it.

FRIDAY 26/01/2024
LATEST AI NEWS

OpenAI Upgrades AI Models, Cuts Costs

OpenAI has announced upgraded text embedding models and reduced pricing for its AI models. The new text-embedding-3-small and text-embedding-3-large models offer improved performance at a lower cost, with the smaller model offering a 5x cost reduction while maintaining accuracy.

OpenAI also plans to release an upgraded version of GPT-3.5 Turbo, featuring higher accuracy and multilingual support, along with a 50% reduction in input token prices and a 25% reduction in output token prices. These advancements demonstrate OpenAI's commitment to enhancing the performance and affordability of its AI models.

Scientists are researching a device that can induce lucid dreams on demand

A technology startup, Prophetic, is developing a wearable headband called Halo, which aims to induce and stabilize lucid dreams using transcranial ultrasound (TUS) technology. The device, currently in the prototype stage, is designed to read EEG data of users and interact with neural activity to enable the experience of awakening in a lucid dream.

The company envisions potential benefits for mental health, creativity, and consciousness exploration. While the possibilities and limits of the technology remain unclear, the research and development of this device represent a significant step in the field of neuroscience and its potential applications

Google Unveils Lumiere: A Breakthrough in AI Video Generation

Google has introduced Lumiere, a cutting-edge AI video generation model that surpasses existing technologies. This state-of-the-art tool can produce highly realistic or surrealistic video clips up to five seconds long, setting a new standard in generative AI video.

Lumiere's "space-time U-net architecture" enables it to generate the entire temporal duration of the video in a single pass, resulting in visually impressive output. While still a research project, Lumiere's capabilities have attracted widespread attention and are poised to revolutionize video generation and editing