Nvidia launches the world’s smallest AI supercomputer

NVIDIA has unveiled Project DIGITS, a $3,000 personal AI supercomputer resembling a Mac Mini, but delivering 1,000x the power of an average laptop. It is powered by the GB10 Grace Blackwell Superchip, enabling it to run AI models with up to 200 billion parameters, and two units can be linked to handle 405 billion parameters.

This system offers 1 petaflop of AI performance, features 128GB of unified memory, up to 4TB NVMe storage, and supports AI development with NVIDIA’s software library and frameworks like PyTorch, Python, and Jupyter notebooks. It runs on the Linux-based Nvidia DGX OS and is designed to empower developers, researchers, and students to work locally on AI models before deploying them to the cloud.

NVIDIA CEO Jensen Huang emphasized its transformative potential to bring data center-class AI computing to individual users. Launching in May, Project DIGITS is set to revolutionize AI accessibility, delivering cutting-edge performance in a compact, energy-efficient design.

Full article

OpenAI vision moves from AGI to ASI

Sam Altman(CEO of OpenAI) reflection explores OpenAI’s journey over nine years, focusing on key milestones, challenges, and lessons learned. It covers the launch and impact of ChatGPT, growth challenges, leadership crises, and OpenAI's evolving vision toward building AGI and superintelligence. Despite setbacks and uncertainties, OpenAI remains committed to its mission of ensuring that AGI benefits humanity.

A major point to focus on in 2025 is the transition to model that can do complex reasoning and as Sam Altman said

Key Points

1. OpenAI's Mission and Growth:

   • Founded nine years ago with the belief that AGI could be humanity's most impactful technology.

   • From a niche research lab in 2022 to a leading AI company with 300+ million users, OpenAI has experienced unprecedented growth.

2. The ChatGPT Launch:

   • ChatGPT, launched on November 30, 2022, became a tipping point for AI adoption, sparking rapid growth and global recognition.

3. Challenges and Lessons:

   • Building a company around new technology involved immense stress, conflicts, and a steep learning curve.

   • Leadership challenges included a sudden firing incident, seen as a governance failure, leading to reforms for better organizational trust.

4. Research, Deployment, and Vision:

   • Emphasis on iterative AI deployment for safety and societal adaptation.

   • OpenAI aims for superintelligence to advance science, innovation, and prosperity.

5. Looking Ahead:

   • Confident about building AGI and preparing for AI agents to transform industries by 2025.

   • Long-term focus remains on maximizing broad benefits and ethical responsibility.

Full article

AI on TVs: Samsung and LG add AI to TVs

At CES 2025, LG and Samsung unveiled new AI-powered features in their upcoming smart TVs, incorporating Microsoft’s Copilot AI assistant and other innovative functionalities.

LG's AI Innovations:

• AI Integration: LG is introducing a dedicated AI section on its TVs, rebranding remotes as “AI Remote,” and integrating Microsoft’s Copilot for contextual information retrieval.

• AI Chatbot: LG showcased an AI chatbot for searching and organizing content but has yet to demonstrate the Copilot feature in action.

Samsung's Vision AI:

• Enhanced Features: Samsung’s Vision AI introduces advancements in AI upscaling, Auto HDR Remastering, and Adaptive Sound Pro, alongside the new Color Booster Pro for richer visuals.

• Interactive AI Experiences:

  • Click to Search: Identify actors, locations, and products in real-time on the screen.

  • Samsung Food: Recognizes dishes in shows, provides recipes, and integrates with grocery services for shopping and delivery.

  • Home Security: AI analyzes connected camera feeds and audio for unusual activity, offering alerts and enhanced home monitoring.

• Accessibility and Security: Live Translate for captions in up to seven languages and Knox Matrix for encrypted data security.

Both manufacturers highlight the potential of AI to enhance user experiences, from personalized content recommendations to innovative interactivity, though detailed demos of Microsoft’s Copilot remain unavailable.

Reference articles -
The verge
Venturebeat

Meta’s terrible AI Instagram and Facebook Profiles

Meta has deleted Facebook and Instagram profiles of AI-powered characters created in 2023 after screenshots of conversations with them went viral. These profiles, initially launched as part of an experiment, included personas like Liv, a Black queer mom, and Carter, a relationship coach. Issues arose when users questioned the bots about their development, revealing a lack of diversity in the creator team. The profiles were removed after a bug prevented users from blocking them.

Meta clarified that these accounts were managed by humans and were part of an early test, not a new product announcement. While Meta-generated profiles are now gone, users can still create their own AI chatbots with various roles, such as a "loyal bestie" or "therapist." Meta provides disclaimers about the accuracy of chatbot messages but does not clearly indicate how content moderation is enforced.

Full article

Amazing Project : Using AI Agent to Automate job application process

"Browser Use" is an open-source tool designed to make websites accessible for AI agents, enabling automated browsing, scraping, and interaction with web pages using large language models (LLMs).GitHub

Use Case: Automated Job Application Process

In a demonstrated use case, "Browser Use" was employed to automate the process of searching for machine learning job opportunities and initiating applications. The AI agent was tasked with the following steps:

1. Resume Analysis: The agent began by reviewing the user's curriculum vitae (CV) to understand their qualifications and experience.

2. Job Search: Utilizing search engines, the agent looked for machine learning job openings that matched the user's profile. For instance, it conducted searches like "Machine Learning Engineer jobs not on job portals" to find unique opportunities.

3. Data Collection: The agent compiled a list of suitable job postings and saved them to a file for record-keeping and further action.

4. Application Initiation: For each identified job, the agent opened new browser tabs and began the application process, filling out necessary information based on the user's CV.

5. User Interaction: If the agent encountered uncertainties or required additional information during the application process, it was programmed to seek assistance from the user, ensuring accuracy and personalization.

This use case exemplifies how "Browser Use" can streamline repetitive tasks, such as job applications, by leveraging AI to navigate web interfaces, extract pertinent information, and perform actions autonomously, thereby enhancing efficiency and productivity.