Apple xrOS, XR display progress, ChatGPT launches

Apple xrOS, XR display progress, ChatGPT launches

Our weekly recap: Apple's XR OS is said to have received a new name internally, there is progress with MicroLEDs and OpenAI presents an impressive chatbot with ChatGPT.

realityOS becomes xrOS

Apple journalist Mark Gurman reports that the operating system developed specifically for the Apple headset is now called "xrOS." The previously rumored name was "realityOS". Gurman refers to sources close to the company. The rumor is a good sign insofar as Apple apparently continues to tinker with the device. Gurman expects a market launch next year, as usual.

MicroLEDs for XR headsets

A Taiwanese institute is presenting a new generation of MicroLED displays that are supposed to score with more colors, higher brightness and less power consumption. These screens could be the basis for a new and slimmer generation of XR glasses.

Ein taiwanesisches Institut stellt eine neue Generation von MicroLED-Displays vor, die mit mehr Farben, höherer Helligkeit und weniger Stromverbrauch punkten sollen.

A Taiwanese institute is introducing a new generation of MicroLED displays that are said to score with more colors, higher brightness and lower power consumption. | Image: ITRI

Neuralink plans brain chip tests on humans

Elon Musk's brain-computer interface startup shares the results of its research in a nearly three-hour presentation. Neuralink plans to venture into human trials for the first time within the next six months, including helping blind people see again.

Levitation is a great mod for Half-Life: Alyx

The free Half-Life: Alyx mod “Levitation” continues the story of the best VR game. It's definitely worth your time if you like VR games and Half-Life in particular. The large, graphically impressive environments are its biggest strength.


Levitation's strength lies in the varied and beautifully designed environments. | Bild: Steam

ChatGPT is OpenAI’s first chatbot and it’s not sentient

ChatGPT is an AI model from OpenAI optimized for dialogs. Like the latest language model for GPT-3, ChatGPT has been trained with human feedback. The chatbot is temporarily available online in a trial phase and is versatile, for example, it can generate extensive texts and even computer code. Is OpenAI possibly collecting more feedback data for the training of GPT-4?

More AI news at THE DECODER: An artist has trained GPT-3 with her diaries and can now have a conversation with her inner child. Plus: Are hidden abilities the norm in large language models?