- Sharp Startup
- Posts
- Open AI Introduces Text to Video!
Open AI Introduces Text to Video!
PLUS: Hugging Face launches GPT competitor
Happy Friday!
OpenAI shocked the world this week as they unveiled Sora, an AI model capable of generating realistic video from text input. Also, YCombinator encourages startup applicants in 20 “important” categories and Hugging Face launches Chat Assistants to compete with OpenAI’s GPT Store.
OpenAI unveils Sora, text to video
YCombinator requesting startup applicants
Hugging Face launches GPT store competitor
New Memory Controls for ChatGPT
Overview: OpenAI introduced Sora, an AI model capable of generating realistic and imaginative video scenes from text instructions, designed to understand and simulate the physical world in motion.
Key Takeaways:
Sora is becoming available for red teamers and several creatives to evaluate risks and enhance the model, as OpenAI seeks external feedback to improve AI capabilities.
The model excels in creating detailed scenes with dynamic characters and consistent visuals, interpreting prompts with a deep linguistic understanding to produce emotionally expressive content.
Sora can produce videos up to a minute long while maintaining high visual quality and adherence to text prompts.
Currently, there are limitations such as inaccuracies in physics simulation and difficulty with complex interactions.
Overview: Y Combinator announced a new Request for Startups (RFS), encouraging innovation in 20 specific categories including explainable AI, LLMs for back-office, and AI enterprise software.
Key Takeaways:
The RFS is aimed at identifying and filling gaps in the startup ecosystem, specifically in categories believed to be “important” in the upcoming decades.
The 20 categories specified cover a broad range of sectors including robotics, space, healthcare, and artificial intelligence.
Y Combinator is accepting applications for the Summer 2024 funding cycle with an early apply deadline of February 21st.
Overview: Hugging Face launched Chat Assistants, directly competing with OpenAI's GPT Store, offering users the ability to create personalized chat assistants easily and for free, contrasting with OpenAI's subscription model.
Key Takeaways:
Hugging Face's platform quickly gained traction with over 4,000 Assistants created shortly after launch.
The Chat Assistants can be fully customized including a choice among various open-source language models.
Unlike OpenAI's GPT Store which is only available to users with a “Plus” subscription, Hugging Face is free, promoting open-source accessibility and community-driven development.
Overview: OpenAI introduced a memory feature in ChatGPT, allowing it to retain information from conversations to enhance future interactions, with user control over memory management and privacy.
Key Takeaways:
ChatGPT's new memory feature allows it to recall past interactions for more coherent conversations, with user options to manage and delete remembered information.
How it works: The more you interact with ChatGPT, the more it learns and improves its memory, enhancing future conversations by either remembering what you specifically ask or naturally picking up on details. You can turn off memory at any time (Settings > Personalization > Memory).
The feature is rolling out to a small portion of ChatGPT free and Plus users this week with plans for a broader roll out soon.