- Nex AI News
- Posts
- Memory and new controls for ChatGPT
Sponsored by
Memory and new controls for ChatGPT
Thursday- March 28 - 2024
Good morning. The memory feature is under testing and has been made available to a select group of ChatGPT free and Plus users. Users have the ability to control ChatGPT's memory, including the option to tell it explicitly to remember or forget certain pieces of information.
In addition to memory, custom instructions have been rolled out, allowing users, including those in the EU & UK, to guide ChatGPT's responses more precisely.
This feature, available on the free plan as well, enables users to provide specific details and guidelines to tailor ChatGPT's interactions according to their preferences.
Gemini 1.5 represents a significant leap forward in AI capabilities, as announced by Google's leadership, including Sundar Pichai and Demis Hassabis.
This next-generation model not only enhances performance across various dimensions but also introduces groundbreaking advancements in long-context understanding and efficiency.
What's New with Gemini 1.5?
Breakthrough in Long-Context Understanding: Gemini 1.5 can process up to 1 million tokens, enabling it to handle vast amounts of data in a single go. This includes the ability to process 1 hour of video, 11 hours of audio, codebases with over 30,000 lines of code, or over 700,000 words.
📈 CCL Carnival Corporation & plc | $17.19 | +0.16 |
🔻 NIO NIO Inc. | $4.6700 | -0.1100 |
📈 TSLA Tesla, Inc. | $179.83 | +2.16 |
📈 MARA Marathon Digital Holdings, Inc. | $22.07 | +1.69 |
🔻 NVDA NVIDIA Corporation | $902.50 | -23.11 |
Today’s newsletter :
Memory and new controls for ChatGPT
Welcome to the Gemini era
Sora is an AI model that can create realistic and imaginative scenes from text instructions
With invideo AI, you can turn any content or idea into video, instantly
Sam Altman's $7 trillion chip dream: Bold vision or delusional fantasy?
Introducing Stable Cascade
NVIDIA ChatRTX - Your personalized AI chatbot
Make-A-Video is a state-of-the-art AI system that generates videos from text.
AI Voice YouTube Monetization: Guide and Tips
Review: Apple Vision Pro
Te memory feature is under testing and has been made available to a select group of ChatGPT free and Plus users.
Users have the ability to control ChatGPT's memory, including the option to tell it explicitly to remember or forget certain pieces of information.
There's also the flexibility to manage and delete specific memories through the settings.
In addition to memory, custom instructions have been rolled out, allowing users, including those in the EU & UK, to guide ChatGPT's responses more precisely.
This feature, available on the free plan as well, enables users to provide specific details and guidelines to tailor ChatGPT's interactions according to their preferences.
One of the benefits of this memory feature is its potential to make the chatbot smarter and more attuned to individual user needs.
For example, custom GPTs like Books GPT will remember which books users have read and their preferred genres, enhancing the personalized experience.
However, with the introduction of this feature, there are privacy considerations, especially concerning sensitive information.
OpenAI assures users that they have control over what ChatGPT remembers and can adjust settings to manage memory.
For more detailed information on the memory and new controls for ChatGPT, you can refer to the official OpenAI blog and the release notes on the OpenAI Help Center.
Welcome to the Gemini era
Summary:
Gemini 1.5 represents a significant leap forward in AI capabilities, as announced by Google's leadership, including Sundar Pichai and Demis Hassabis.
This next-generation model not only enhances performance across various dimensions but also introduces groundbreaking advancements in long-context understanding and efficiency.
What's New with Gemini 1.5?
Breakthrough in Long-Context Understanding: Gemini 1.5 can process up to 1 million tokens, enabling it to handle vast amounts of data in a single go.
This includes the ability to process 1 hour of video, 11 hours of audio, codebases with over 30,000 lines of code, or over 700,000 words.
Efficient Architecture: Built upon a Mixture-of-Experts (MoE) model, Gemini 1.5 is designed for efficiency, selectively activating the most relevant expert pathways for a given input.
This specialization significantly enhances the model's efficiency, allowing it to learn complex tasks more quickly while maintaining high quality.
OpenAI has launched Sora, a revolutionary tool that turns text into realistic videos, marking a significant advancement in generative artificial intelligence.
This innovative technology can generate complex scenes with multiple characters, specific motions, and detailed backgrounds, capturing the essence of the user's prompts and bringing them to life in the physical world.
Technical Breakthroughs Behind Sora
Sora's development leverages the diffusion model technology from DALL-E 3, enhancing it with transformer architecture for video generation.
This unique combination allows Sora to process video data in chunks across both space and time, accommodating a wide variety of videos in terms of resolution, duration, aspect ratio, and orientation.
This method enables Sora to train on a broader range of video types than other models.
Unique Capabilities of Sora
One of Sora's standout features is its ability to animate existing images and extend videos by filling in missing frames.
It showcases remarkable flexibility in video generation, supporting videos in various sizes and aspect ratios.
This adaptability makes Sora suitable for different platforms and devices, offering rapid prototyping at lower resolutions before scaling up to full resolution.
AI-Powered Scripts: Generate engaging scripts tailored to any video topic, saving time and enhancing content quality.
Voiceover Generator: Offers human-sounding AI speech in various languages, adding life to your videos.
Diverse Content Options: Access to over 16 million stock photos and videos, ensuring a wide range of choices for your projects.
Intuitive Interface: Designed for ease of use, facilitating quick navigation and video editing without the need for extensive technical skills.
Customization and Collaboration: Customize every video aspect and collaborate in real-time with other users, making it suitable for team projects.
User Experience and Pricing:
InVideo has garnered high praise for its user-friendly platform and the quality of videos users can create, with an impressive overall rating based on hundreds of reviews.
The platform offers flexible pricing plans, starting as low as $4 per month, making professional-quality video creation accessible and affordable.
Comparisons and Alternatives:
While InVideo stands out for its comprehensive features, other platforms like Synthesia, Canva, and Pictory offer alternative solutions for AI video creation.
Synthesia specializes in non-camera videos using avatars, Canva combines graphic design with video editing, and Pictory provides diverse effects and templates for various video editing needs.
How much should be invested to "secure our collective future"? $7 trillion in chip factories, if you ask Sam Altman.
According to a report by the Wall Street Journal, the CEO of OpenAI is in talks with investors to raise between $5-7 trillion to reshape the global chip industry and thereby ensure the supply of the computing power required for artificial intelligence developments.
"We believe the world needs more ai infrastructure--fab capacity, energy, datacenters, etc--than people are currently planning to build," Altman wrote on X earlier this month.
"Building massive-scale ai infrastructure, and a resilient supply chain, is crucial to economic competitiveness."
While his words may sound reasonable, the reported pricing of the project is unusual by any standard.
News
What else is nex?
Introducing Stable Cascade
Today we are releasing Stable Cascade in research preview, a new text to image model building upon the Würstchen architecture. This model is being released under a non-commercial license that permits non-commercial use only.
Stable Cascade is exceptionally easy to train and finetune on consumer hardware thanks to its three-stage approach.
NVIDIA ChatRTX - Your personalized AI chatbot
ChatRTX is a demo app that allows you to personalize a GPT for large languages (LLM) connected to your own content, such as documents, notes or other data.
Thanks to retrieval-augmented generation (RAG), TensorRT-LLMen RTX acceleration allows you to query a custom chatbot to quickly get contextually relevant answers.
And because everything is running locally on your Windows RTX PC or workstation, you get fast and safe results.
Make-A-Video is a state-of-the-art AI system that generates videos from text
IN 2023, OPENAI told the UK parliament that it was “impossible” to train leading AI models without using copyrighted materials.
It’s a popular stance in the AI world, where OpenAI and other leading players have used materials slurped up online to train the models powering chatbots and image generators, triggering a wave of lawsuits alleging copyright infringement.
AI Voice YouTube Monetization: Guide and Tips
Unleashing the Power of AI Voices on YouTube: Your Guide to Monetization Success!
🚀 Calling all aspiring YouTubers! 🚀 Ready to supercharge your channel and break free from the camera?
Dive into the world of AI voices for YouTube and unlock the secrets to monetization success.
Join us as we explore the pros, cons, and everything you need to be aware of in this AI-driven content creation revolution!
Review: Apple Vision Pro
BY NOW, YOU have likely read a great deal about the experience of using Apple's Vision Pro mixed-reality headset, which the company is positioning as the future of computing.
There have been countless unboxings, teardowns, reviews, and even reports of damaged front screens.
I encourage you to read my colleague Lauren Goode's excellent write-up about watching movies—and crying—with the headset on.
TOOLS
100+ AI Business Ideas
What you can gain from subscribing to our exclusive : Over 100 AI Ideas.
http://embeds.beehiiv.com/f94e1508-2161-4798-8a3e-27d1009b683f
Feel free to share the link with friends and family!
THAT’S ALL PEEPS
Did you find our newsletter/information enjoyable? Share the love with your friends or subscribe for more delightful updates!
See you next week !
Join these Groups on FB to get more into AI communities!
The AI Rush - The Latest AI Tools, News, Tips & Guides
Artificial Intelligence (ChatGPT & more) and Machine Learning
The Rundown is the world’s fastest-growing AI newsletter, with over 500,000+ readers staying up-to-date with the latest AI news and learning how to apply it.
Our research team spends all day learning what’s new in AI, then distills the most important developments into one free email every morning.
|
Reply