AI News: The Best Open Source Model EVER
TLDRThis week's AI news features Meta's release of Llama 3, an open-source large language model with real-time knowledge integration and creative features. The industry anticipates the 400 billion parameter model for advanced capabilities. Nvidia highlights its role in Llama 3's training, and Grock announces its upcoming integration. Hugging Face and Meta's new website offer ways to use Llama 3, including an AI image generator and animation feature. GPT Trainer sponsors the video, promoting its no-code framework for building chatbots. XAI unveils Grock 1.5 with vision capabilities, and PO introduces multibot chat. Microsoft and Google are in a race to build data centers for AI advancement. Stable Diffusion 3, an AI image generation model, is released, and Leonardo AI is expected to integrate it soon. Microsoft presents Vasa, an emotional talking head generator, and Instant Mesh transforms 2D images into 3D objects. Adobe demonstrates AI capabilities at NAB, including object removal and video extension. Da Vinci Resolve 19 introduces AI color grading and motion tracking. The US Air Force confirms a successful AI dogfight with human-piloted jets. Several AI gadgets are highlighted, including the Humane AI pin, Rabbit R1, Limitless pendant, Nothing earbuds with chat GPT, and Logitech's AI prompt builder for mice. Boston Dynamics' Atlas 001 robot goes viral for its unsettling yet impressive video.
Takeaways
- ๐ Meta has released LLaMa 3, an open-source AI model that integrates real-time knowledge from Google and Bing and offers unique creation features like animation and image generation.
- ๐ LLaMa 3's 8 billion and 70 billion parameter models show comparable performance to existing open-source models, with upcoming releases promising multimodality and larger context windows.
- ๐ง The 400 billion parameter model of LLaMa 3 is highly anticipated for its potential to compete with current models like GP4 and Claude 3 Opus.
- ๐ป Users can access LLaMa 3 via Hugging Face's API and a new website that allows web searches when answering questions.
- ๐จ Meta's AI website features an AI image generator that creates images in real-time as users type, with an additional animation feature.
- ๐ค NVIDIA highlights their role in training LLaMa 3 on their GPUs and the upcoming availability of the model on Grock.
- ๐ฑ GPT Trainer is introduced as a no-code framework for building multi-agent chatbots with function calling capabilities, aimed at enhancing customer support for online businesses.
- ๐ PO's new 'multibot chat' feature allows users to ask questions and have the best model for the task selected automatically, or summon a specific bot by mentioning it.
- ๐ข Microsoft and Google are investing heavily in infrastructure to advance their AI capabilities, with both planning to spend over a hundred billion dollars on data centers.
- ๐จ Adobe demonstrated AI features at the NAB conference, including object removal, video extension, and integration with AI video generation models like Pika and Sora in Adobe Premiere.
- ๐ค Boston Dynamics' new Atlas 001 robot has gone viral, showcasing a more compact and electric design compared to its predecessor.
Q & A
What is the main announcement from Meta this week?
-The main announcement from Meta this week is the release of Llama 3, an open-source, large language model that is expected to be the most intelligent AI assistant available for free use.
What are some of the unique features of Llama 3?
-Llama 3 has integrated real-time knowledge from Google and Bing, creates animations, and generates high-quality images in real-time as users type.
What is the significance of the 400 billion parameter model of Llama 3?
-The 400 billion parameter model of Llama 3 is expected to have significantly better capabilities, including multimodality, the ability to converse in multiple languages, larger context windows, and stronger overall capabilities, potentially competing with current models like GP4 and Claude 3 Opus.
How can users currently access and use Llama 3?
-Users can access Llama 3 via the API on Hugging Face or use it on Meta's platform. It is also available on a new website released by Meta that can search the web when answering questions.
What is the AI image generator feature on the Meta AI website?
-The AI image generator on the Meta AI website is a feature that generates images in real-time as users type in their prompts. It also has an 'animate' feature that can turn generated images into short animations.
How does the GPT Trainer platform assist online businesses?
-GPT Trainer is a no-code framework that allows users to build multi-agent chat GPT-like chatbots with function calling capabilities, enabling 24/7 customer support, and the ability to escalate chats to a real human when needed.
What is the recent update from Xai regarding their AI model?
-Xai announced Grok 1.5 with Vision, which is on par with other models that also have vision capabilities. It can write code from a diagram and has other features showcased on their website.
What is the new feature released by PO chatbots?
-PO chatbots released a new feature called multibot chat, which allows the system to pick the best model to use based on the question asked or lets users tag a specific bot for the question.
What is the significance of the investment by Google and Microsoft in AI infrastructure?
-Both Google and Microsoft are investing at least a hundred billion dollars to build infrastructure to scale up their AI efforts, with the aim of being the first to achieve Artificial General Intelligence (AGI).
What are some of the AI advancements in the art world?
-There have been significant advancements in AI art, with the release of Stable Diffusion 3, which is particularly good with text in images. However, a user interface for it is not yet available.
What is the new feature that Leonardo AI is expected to release soon?
-Leonardo AI is expected to release a style transfer feature soon, which allows users to upload a style reference image and generate a series of images in that same style.
Outlines
๐ Meta Releases Llama 3: A New Milestone in AI
This week's major AI news is Meta's unveiling of Llama 3, an advanced open-source AI model. Llama 3 succeeds Llama 2 and is expected to compete with current models like GP4 and Claude 3 Opus. Meta has released two versions: an 8 billion parameter model and a 70 billion parameter model, with the latter showing comparable performance to existing free AI models. However, the upcoming 400 billion parameter model is anticipated to offer significant improvements in multimodality and language capabilities. The release includes integration of real-time knowledge from Google and Bing, as well as creative features like animation and high-quality image generation. Llama 3 is available via Hugging Face's API and is expected to be hosted on Nvidia's Grock platform soon.
๐จ Meta's AI Image Generator: Real-time Image Creation and Animation
Meta has introduced a new AI image generator under the Imagine tab on their website. This tool creates images in real-time as users type their prompts. It also includes an 'animate' feature, allowing users to transform still images into short animations. The tool is free to use and offers a playful and interactive way to generate images in various styles, including realistic, Simpsons-style, and 8-bit video game styles.
๐ค Multibot Chat and the Future of AI Language Models
PO has launched a new feature called multibot chat, which allows users to ask questions and have the system select the best model to answer. Users can also summon specific bots by mentioning them. This approach suggests a future where chatbots interact with various large language models to provide the best response to a query. Additionally, both Microsoft and Google are investing heavily in infrastructure to advance their AI capabilities, with a focus on achieving AGI (Artificial General Intelligence).
๐จ Stable Diffusion 3 and AI Art Innovations
Stable Diffusion 3 has been released, offering improved text incorporation in images. Although there's no user-friendly interface yet, the API is available for integration into software products. The AI art world is also abuzz with Leonardo AI's anticipated release of a style transfer feature, which uses a style reference to generate images in a consistent style. Microsoft has also made strides with Vasa, an AI research project that creates talking videos from headshots and audio clips, with advanced emotive expressions.
๐ ๏ธ AI Tools for 3D Modeling and Video Editing
New AI research called Instant Mesh allows 2D images to be transformed into 3D objects. Adobe showcased AI capabilities at the NAB conference, with features like object removal, AI-driven color grading, and motion tracking. Spline Tool added text-to-3D image generation within their app, and Da Vinci Resolve 19 introduced AI-powered features. These advancements are set to revolutionize content creation and editing.
๐ค AI Gadgets and the US Air Force's AI Dogfight
The US Air Force has confirmed the first successful AI dogfight with a jet controlled by AI, which did not require human intervention. In the consumer space, AI gadgets are gaining attention. Rabbit R1 is now shipping, allowing users to train it for specific tasks. The Limitless pendant, formerly Rewind, records conversations with consent and provides transcripts. Nothing's earbuds are integrating with chat GPT, and Logitech is announcing an AI prompt builder for their mice. Lastly, Boston Dynamics' new Atlas 001 robot has gone viral for its human-like movements.
๐ฐ Wrapping Up AI News and Future Tools
The host summarizes the AI news covered in the video and encourages viewers to check out Future Tools for more AI news and tools. He also promotes the NextWave podcast for deeper discussions on AI topics. The host expresses gratitude to the viewers and the sponsor, GPT Trainer, for their support.
Mindmap
Keywords
๐กLlama 3
๐กOpen Source
๐กMultimodality
๐กHugging Face
๐กAI Image Generator
๐กGPT Trainer
๐กGrock 1.5 with Vision
๐กStable Diffusion 3
๐กAdobe Premiere
๐กAI Dogfight
๐กAI Gadgets
Highlights
Meta has released Llama 3, an open-source large language model that is set to compete with current models like GP4 and Claude 3 Opus.
Llama 3 integrates real-time knowledge from Google and Bing, enhancing its AI capabilities.
The model features unique creation tools, enabling it to generate animations and high-quality images in real-time.
Two versions of Llama 3 have been released: an 8 billion parameter model and a 70 billion parameter model.
The 400 billion parameter model of Llama 3 is expected to have advanced capabilities such as multimodality and larger context windows.
Llama 3 is available for use via the API on Hugging Face and is expected to be on Grock soon.
Meta's AI website allows users to ask questions that the model will search the web to answer, citing sources.
The Imagine tab on Meta's AI website features an AI image generator that creates images in real-time as users type.
GPT Trainer is a no-code framework for building multi-agent chatbots with function calling capabilities.
XAI announced Grock 1.5 with Vision, which can write code from a diagram and is comparable to other vision-equipped models.
PO has introduced multibot chat, allowing the platform to select the best model for the user's question or let the user choose.
Google and Microsoft are both investing heavily in infrastructure to advance their AI efforts, aiming for AGI.
Stable Diffusion 3 has been released, excelling at text within images, but lacks a user-friendly interface for now.
Leonardo AI is expected to soon integrate Stable Diffusion 3 and will release a style transfer feature.
Microsoft's VasaOne research allows the creation of talking videos from headshots and audio clips, with advanced emotion and detail.
Instant Mesh is an open-source tool that converts 2D images into 3D objects, offering a rough draft for further refinement.
Adobe demonstrated AI-powered features at NAB, including object removal, video extension, and integration with AI video generation models like Pika and Sora.
Da Vinci Resolve 19 introduces AI color grading and AI-powered motion tracking, enhancing post-production capabilities.
The US Air Force confirmed the first successful AI dogfight using an AI-controlled jet against a human-controlled jet.
AI-enabled gadgets like the Rabbit R1, Limitless Pendant, and Logitech's AI prompt builder for mice are gaining attention.
Boston Dynamics' new Atlas 001 robot showcases significant advancements in size, noise reduction, and electric operation.