Meta’s new LLM launched, “Llama 3.1” 2024
Trending AI News
Meta’s new LLM, “Llama 3.1”
Meta is constantly working on developing open-source LLMs by launching its Llama family of AI models. Mark Zuckerberg, the CEO of Meta, released a video and announced about the Meta’s most capable AI model, “Llama 3.1”. These new LLMs have outperformed well-known LLMs such as GPT-4, GPT-4o, Claude 3.5 Sonnet, etc, on several benchmarks such as code generation, mathematics, and reasoning.
Llama 3.1 has over 405 billion parameters and is trained on over 16000 GPUs of Nvidia H100. Meta is working with many companies to integrate this model into their system. Meta also claimed that the cost of Llama 3.1 is roughly half the cost of OpenAI’s GPT-4o in production. Soon, this LLM will also be available for the general public as the Meta AI assistant in their apps. Mark Zuckerberg also stated that Llama 3.1 would play a role similar to Linux in making an open-source operating system for the entire world.
Also Read- 7 Best FREE AI Chatbots That Will Blow Your Mind
Meta new AI face swap
Meta also released new AI features for its image generation model, “Imagine Me.” Now, users can scan their faces using the phone’s camera and imagine themselves in different avatars. Meta also added AI to its AR/VR headset Quest and Ray-Ban glasses, which allows users to identify and learn about what they are looking for.
Also Read- Best AI Humanoid Robots
Mistral launched a new AI model, “LARGE 2”
The Paris-based AI startup launched its new open-source large language model(LLM) just after the one-day launch of Llama3.1, “Large 2.” Large 2 has close to 123 billion parameters and a 128000 token window, which means it can take input of approximately equal to the 300-page book. Large 2 is multilingual, as it supports more than 80 languages. Mistral also claims that their new LLM outperforms Llama 3.1 regarding code generation and math performance.
OpenAI’s new experiment, “Serach GPT”
OpenAI is the company that created a rush in the decade-old companies. The company recently announced the “Search GPT,” which is similar to Perplexity AI. It is an AI-powered search engine with real-time access to the information. It works like when you search for something on Google, giving a handful of links related to that keyword. Rather than this, this LLM organizes the information with a summary and the attributed links. Users can also ask follow-up questions. This new experiment is currently in the prototype stage and will soon be available to the public. This could be there for the Google search business as it is the main source of revenue for Google.
Also Read- Top 30+ AI Robotics Projects: Step Into the Future
Google AI scored silver in the Maths Olympiad
The Google Deepmind LLM, AlphaProof, and AlphaGeometry 2 have achieved the silver-level medal in solving the International Maths Olympiad questions. During the test, the AlphaProff solved two algebra problems and one number theory problem. It also managed to solve the hardest problems of the competition, which can only solved by the five contestants. While the AlphaGeometry 2 can solve complex geometric problems in just 19 seconds. This new advancement is closing the gap between the human mind and machine capabilities.
Also Read- Open Source Takes Over AI: New Wave of LLMs
Google AI’s new weather model
Google Deepmind released a weather forecasting AI model named “NeuralGCM.” This new AI model has trained over the weather data of past years and combines the theory of machine learning and physics in a very subtle. It is currently in testing mode as it predicts the weather for the next 15 days, mainly due to the atmospheric changes in the area over 25km, like the cloud formations or regional microclimates.
MIT’s new achievement in cancer detection
MIT researchers have been working on a breast cancer detection model for many years, and the initial results are mind-blowing. The researchers have developed a deep-learning model to predict breast cancer from mammograms. A mammogram is an X-ray screening used to detect breast changes in women who have no signs or symptoms of breast cancer. This AI model shows accurate results based on various demographics and clinical settings. This AI model also predicts the risk factors in the future time based on family history and age.
Also Read- 4 Best AI Tools For Web Development: Try Now
Adobe’s new AI features for Photoshop
Adobe has launched new AI features for Adobe Photoshop and Adobe Illustrator. These features are planned to simplify the user design process, and they include Generative Shape Fill, Style reference, Retype, Mockups, a new Taskbar, and Adobe Firefly powers.
Trending AI tools
- Granola (Click here to try)
An AI tool that helps users enhance their meeting notes and convert them into a transformational record. - Octolens (Click here to try)
An AI tool that helps the business for social listing for lead generation. - XspaceGPT (Click here to try)
An AI tool that helps users analyze conversations in Twitter spaces and generate key points and summaries using well-known LLM. - Tidyread (Click here to try)
An AI tool that created the summaries for the user’s feed. - Llama Tutor (Click here to try)
An open-source AI-powered tutor that is developed using the Llama LLM.
Prompt of the Week
Many times, we need to get some specific information from a document or PDF. However, it is hard to find information from the bulk of digital pages. So, to solve this problem, we will use the computer vision of well-known LLMs such as Claude 3.5 or GPT-4o.
Here is the prompt:
Act as a computer vision expert whose expertise in scanning and locate all the mentions of “context” in the given document. Create a comprehensive report that organises these findings, which includes the findings, including exact locations, summaries of the surrounding text, and an in-depth analysis of the role each mention plays in the document’s overall message or purpose. Aim for a clear and thorough explanation to enhance understanding of the document’s treatment of the specific topic.
Job Listing
- Research Scientist, Waymo (Click here to apply)
Waymo is an autonomous driving technology company developing AI and computer vision technology for autonomous vehicles. The company also developed Google’s Self-Driving Car Project in 2009. Waymo seeks a Research Scientist role to work on the foundation model and compelling experiments by training and evaluating large deep learning models. The candidate must have a Master’s or PhD in deep learning or equivalent years of experience and have expertise in Python and deep learning frameworks such as Tensorflow, Pytorch, or JAX. - Research Scientist, Google Deepmind (Click here to apply)
Google Deepmind is a segment of Google that mainly develops artificial intelligence and machine learning for education, learning, and development. The company is looking for a research scientist whose main work is to solve new problems and improve the performance of existing models. The candidate must have expertise in Python and deep learning frameworks and a PhD in machine learning. - Mobile Engineer, Abridge (Click here to apply)
Abridge is a company that is developing AI solutions for healthcare. It involves chatbots for medical conversations, clinical documentation, etc. The company is looking for a Mobile Engineer to improve their existing applications. The candidates must have expertise in mobile tech stack, which includes both Android/iOS native.
This concludes our exploration of AI updates for this week. Stay tuned for more exciting updates next week! 🔥❤️⚡🤖📷
About The Author
