1. OpenAI Unveils Sora: A New AI for Video Generation

Sora, OpenAI’s latest generative AI model, is revolutionizing video creation with its ability to generate videos up to a minute long based on textual cues. Moreover, it can extend these videos while maintaining stunning visual quality.

Sora’s capabilities go beyond understanding text and accurately rendering cues. It can create multiple shots within a single video, preserving the characters and visual style consistently. The model also supports generating images, converting still images into videos, expanding existing videos, bridging two videos, and filling in missing frames.

What sets Sora apart are its impressive 3D simulation capabilities. These capabilities show promising prospects in various fields, including the creation of short videos, animations, cinematics, and even rendering video games. With such a diverse range of applications, Sora is truly setting new standards in AI-powered video generation.

Address: https://openai.com/sora

2. Amazon Launches 1 Billion Parameter Text-to-Speech Model on 100K hours of data

According to the Arxiv website, the Amazon team launched a text-to-speech (TTS) model called BASE TTS on February 12, representing a new surge of capabilities for large-scale adaptive streaming TTS. As the largest TTS model to date, BASE TTS was trained on 100,000 hours of public speech data to improve speech naturalness.

Paper address: https://arxiv.org/pdf/2402.08093.pdf

Audio generation sample address: https://amazon-ltts-paper.com

3. NVIDIA Releases Chat With RTX

NVIDIA has launched Chat with RTX, enabling users to create a personalized chatbot experience. The app works on a local Windows PC or workstation, ensuring data privacy and control. It allows users to train a language model with their own data, including documents, notes, transcripts, etc., to build a chatbot tailored to their specific needs.

The app requires users to download Mistral or Llama 2 models separately for training.

To find out more, please read: https://www.nvidia.com/en-us/ai-on-rtx/chat-with-rtx-generative-ai/

4. ChatGPT’s Memory Management Goes Live

OpenAI announced on its website that they are testing a memory management feature for ChatGPT. The feature remembers what users have discussed in all their chats, and also allows users to not have to retype information and be more helpful in future conversations. This week OpenAI will roll out the memory management feature to a small group of both ChatGPT Free and Plus users.

5. Several AI Startups Receive Funding as Investor Enthusiasm Rises

According to recent reports by The Information, AI startups are gaining significant attention from investors. Scale AI, which provides high-quality training data for AI, is currently in talks with investors to raise more funds. Perplexity AI, an AI search engine startup, has received an unsolicited offer that could value it at more than $1 billion. Meanwhile, Qualcomm and Intuit have invested $15 billion in AI startup Anthropic. Additionally, medical AI startup Abridge has successfully closed a new funding round, with a post-investment valuation of $850 million.

6. General Motors to Expand Super Cruise Network to Rural Roads

General Motors (GM) unveiled plans on Thursday to extend its hands-free Super Cruise technology network to include minor highways and rural roads. This expansion will increase the network’s coverage to approximately 750,000 miles across the U.S. and Canada.

Launched in 2017, Super Cruise recently expanded its network to 400,000 miles. With this new addition, GM establishes Super Cruise as the largest truly hands-free operating domain in North America. The coverage is nearly six times larger than other hands-free driver assistance technologies currently available in the market.