NVIDIA

🌐 Nvidia at the forefront of AI

The GTC 2024 Conference of NVIDIA was a highlight in the field of artificial intelligence and accelerated computing, with CEO Jensen Huang presenting a series of transformative innovations. The Blackwell platform, a cutting-edge architecture, promises to revolutionize accelerated computing with six disruptive technologies, including an ultra-fast chip connection interface of 10 TB/s. The Blackwell GPUs, with their 208 billion transistors distributed across two chips, represent a remarkable feat in semiconductor design, offering up to 40 TFLOPS in FP64 and an impressive memory bandwidth of 8 TB/s.

The event also highlighted the launch of the DGX SuperPOD, a supercomputer powered by Blackwell GPUs, designed to handle generative AI at a scale of trillions of parameters, setting a new standard for supercomputing. NVIDIA introduced Generative AI microservices, which allow companies to detect and solve software security issues in seconds, an innovation that could transform the maintenance of critical systems. The Omniverse Cloud APIs were announced as essential tools for driving the next wave of industrial digital twin software, while the DRIVE platform promises to fuel the next generation of transportation, from cars and trucks to robot taxis and autonomous delivery vehicles.

Furthermore, NVIDIA solidified strategic partnerships with technology giants such as AWS, Microsoft, Oracle, Google Cloud, and SAP, to accelerate the adoption of generative AI across a variety of industries. These developments underscore NVIDIA's ability to lead the era of generative AI and have the potential to reshape entire industries.

Learn more:

1. nvidianews.nvidia.com | 2. forbes.com

OPENAI

🎧 Interview with Sam Altman resonates on the web

Sam Altman, CEO of OpenAI, participated in a podcast with Lex Fridman, where they discussed OpenAI, AI development, and the future of AGI (General Artificial Intelligence). Altman shared reflections on the organizational challenges faced by the company, including the controversy surrounding the restructuring of the board and the lawsuit initiated by Elon Musk. Altman described the experience as painful but instructive, highlighting the need to build a robust organization capable of withstanding external pressures. He also discussed Sora, OpenAI's video generation model, which he believes has the ability to accurately model the physical world. Despite concerns about the misuse of AI, such as the creation of deepfakes, Altman emphasized the importance of responsible deployment.

Looking to the future, he expressed enthusiasm for the advancements represented by GPT-4, but Altman believes it will seem "unimaginably awful" compared to GPT-5. Altman predicts that computing will become the most valuable resource in the near future, driven by the need to solve energy challenges, such as those presented by nuclear fusion. He concluded with a cautiously optimistic view of the impact of general artificial intelligence (AGI), anticipating that it will accelerate scientific discovery and provoke significant social changes in the coming decades.

Learn more:

1. youtube.com

OPENAI

🌞 GPT-5 release date leaked

GPT-5, the next major update of ChatGPT by OpenAI, could be launched this summer, bringing a series of significant improvements. Although specific details about the improvements have not been officially disclosed yet, GPT-5 is expected to offer advancements in contextual understanding, factual accuracy, computational efficiency, and multimodal capabilities.

The arrival of GPT-5 marks an important moment in the trajectory of artificial intelligence, with expectations that ChatGPT will become even more sophisticated and interactive. OpenAI's continuous innovation in AI development places the company at the forefront of the industry, and GPT-5 is anticipated as a watershed moment in this process.

Learn more:

1. msn.com | 2. arstechnica.com

NEURALINK

🧠 Neuralink: patient plays chess with a brain chip

Neuralink, Elon Musk's company, demonstrated a significant advance in the brain-computer interface with its first patient, Noland Arbaugh, moving a cursor on a computer and playing online chess solely with thought. Arbaugh, who became paralyzed below the shoulders after an accident, received the brain chip implant in January and, during a nine-minute live broadcast, showed his ability to control the cursor to play chess. Neuralink's technology, which involves inserting a coin-sized device into the skull with microscopic wires that read neuron activity, represents hope for individuals with paralysis. The company was authorized to test the chip in humans by the FDA in May 2023 and is among several organizations seeking to commercialize this technology, which could have profound implications for complex neurological conditions.

Learn more:

1. bbc.com | 2. watcher.guru

APPLE

📱 Apple intends to integrate Gemini into the iPhone

Apple is reportedly in negotiations to integrate Google's Gemini artificial intelligence engine into its iPhones. The potential partnership aims to incorporate Gemini to power new AI features in iOS 18, including cloud-based text and image generation. Additionally, Apple is developing its own on-device AI models for the next iOS update. Discussions are still ongoing, and the final implementation and branding of the AI agreement have not been decided yet. If realized, this collaboration could significantly expand the existing relationship between the two companies, beyond the standard search engine partnership. The Gemini technology promises to bring a series of benefits, such as improved Siri performance and new generative AI features for iPhone users.

Learn more:

1. reuters.com | 2. arstechnica.com

RESEARCH

👁️‍🗨️ MindEye2: reconstruction of images from brain activity

Researchers presented a significant advance in reconstructing visual perceptions from brain activities with the MindEye2 project. Using just 1 hour of fMRI data, the model demonstrated high-quality results. They used an MRI machine to capture what happens in a person's brain while they look at something. Then, a generative model, trained to understand these brain images, tries to recreate the original image the person saw. This study is an advance that could lead to the creation of mind-reading technologies to help people with disabilities or to better understand how our brain works, representing a milestone in brain-computer interface research.

Learn more:

1. arxiv.org

⚠️ Stability AI, the company behind the image generator Stable Diffusion, faces challenges after the departure of key researchers. The loss of talent and financial difficulties raise uncertainties about the company's future in the AI market. Learn more.

🤖 During a conference, Nvidia's CEO Jensen Huang discussed the possibility of achieving General Artificial Intelligence (AGI) within 5 years, defining it through specific tests. He also addressed AI "hallucinations," explaining they are solvable problems and that AI model predictions should be treated with caution due to inherent uncertainty. Learn more.

🎮 At the 2024 Game Developers Conference (GDC), Ubisoft unveiled a prototype of NPCs called Neo, utilizing Nvidia's AI technology to create characters that dynamically respond to players' commands. These Neo NPCs are capable of conversing with players and reacting with contextual awareness, representing a significant advancement in creating more immersive and interactive game worlds. Learn more.

🔍 GitHub introduced CodeQL, an analysis tool that allows users to query code as if it were data. With CodeQL, it is possible to identify vulnerabilities within a codebase by writing queries to find all variants of a vulnerability and eradicate them permanently. Learn more.

⚽ DeepMind introduced TactiCAI, an AI assistant for soccer tactics, which helps coaches analyze and optimize plays, especially in corners. Developed in collaboration with Liverpool FC, TactiCAI uses deep learning to predict players' actions and suggest tactical adjustments, being validated by experts as an effective tool indistinguishable from real strategies. Learn more.

🌍 Nvidia launched Earth-2, a digital platform for advanced climate simulation, offering quick and detailed forecasts and alerts, using AI to improve accuracy and efficiency in predicting extreme weather events and their economic consequences. Learn more.

🔬 Samsung established a research lab to develop advanced chips aimed at the next phase of artificial intelligence, focusing on language models and energy efficiency. The initiative aims to push progress towards General Artificial Intelligence (AGI), led by experts in semiconductors and software engineering. Learn more.

🇯🇵 Japan is utilizing AI in matchmaking programs to boost its low birth rate, subsidizing local projects that employ advanced analytics to match people based on interests and values. The measure aims to address population decline and ensure a sustainable workforce for the country's future. Learn more.

🧠 Researchers at Stanford developed a method that improves the performance of AI systems by training them to perform an “inner monologue” before responding. The QuietSTaR algorithm applied to the Mistral 7B model increased its scoring in reasoning tests and doubled its effectiveness in mathematics, addressing deficiencies in contextual reasoning and common sense. Learn more.

Neuralink impressed the industry last week with a demonstration of the first patient to have a company chip implanted in the brain using the technology to play a chess game solely with the mind. Click here or on the image above to watch.

🎶Hydra: Rightsify's AI music generation tool, which uses an extensive dataset and advanced technology to create custom instrumental music and sound effects. Ideal for content creators, advertising, entertainment, and personal focus, Hydra offers plans ranging from a free version to professional and premium options, allowing commercial use of the generated music. Link.

📱Glowbom: an intuitive platform that allows anyone to create apps and games without needing to code. With a virtual assistant for guidance, users can customize their projects and export them to various platforms, making app development accessible and encouraging innovation. Link.

🧑🏽‍🦱Character.AI: a platform that enables the creation and interaction with personalized AI characters, offering realistic conversations and a wide range of historical, fictional, or original figures. Ideal for entertainment, education, and creative development, Character.AI uses cutting-edge technology to provide unique dialogue experiences. Link.

🎮 Buildbox: a platform that allows anyone to create amazing games, regardless of technical skill. With Buildbox, users can develop games without needing programming or software design experience, making the game creation process smooth and accessible. Link.

🚀 AdCreative.ai: an advertising tool that uses artificial intelligence to generate effective ads. With the ability to create ad creatives optimized for conversion, AdCreative.ai helps users significantly improve click-through and conversion rates, offering a 7-day free trial to test the platform. Link.

⏹️ Create diagrams with the free version of ChatGPT

Did you know that with the free 3.5 version of ChatGPT, you can create detailed diagrams of your project? Here's how to do it:

• Open ChatGPT 3.5 and ask it to create a step-by-step roadmap with the activities of your project.

  

• Ask it to transform the steps into a flow diagram. Detail any actions you want it to consider, such as placing independent tasks in parallel and considering the prerequisites for each step.

  

• Ask it to generate a mermaid code of the diagram. Copy the generated code.

  

• Go to Draw.io, click on Insert > Advanced > Mermaid, and paste the copied code.

  

Voilà! Now your diagram can be customized to your preferences.