2 days ago
Six Happenings that Changed the World of AI This Week
From Meta, OpenAI, Google, and Databricks to Nvidia and Samsung, numerous new AI capabilities have been announced
Opinions expressed by Entrepreneur contributors are their own.
You're reading Entrepreneur India, an international franchise of Entrepreneur Media.
Since the global tech giants completed their annual conferences, it seems Artificial Intelligence (AI) developments haven't taken a single day's breath. From Meta, OpenAI, Google, and Databricks to Nvidia and Samsung, numerous new capabilities have been announced, all aimed at pushing AI into its next chapter.
Meta's V-JEPA 2 and the Age of World Models
Meta, long a silent workhorse in the AI research domain, has stepped into the spotlight with its latest offering V-JEPA 2, an open-source AI world model. Unveiled on June 11, V-JEPA 2 isn't just another generative model it's designed to reason about the physical world.
At its core, the model allows AI systems to internally simulate real-world environments in 3D, equipping machines with the cognitive ability to anticipate how objects might move or behave even in unfamiliar settings. In practice, this could dramatically enhance fields like robotics and autonomous driving, where a deep understanding of physical dynamics is vital.
Rather than relying on heavily labelled datasets, V-JEPA 2 is trained on unlabelled video content, making it both scalable and efficient. Meta says the model leverages what's called "latent space" to intuit motion and interaction essentially, AI imagination.
"Today, we're excited to share V-JEPA 2, the first world model trained on video that enables state-of-the-art understanding and prediction, as well as zero-shot planning and robot control in new environments. As we work toward our goal of achieving advanced machine intelligence (AMI), it will be important that we have AI systems that can learn about the world as humans do, plan how to execute unfamiliar tasks, and efficiently adapt to the ever-changing world around us," Meta noted in its official blog.
OpenAI's o3-Pro: A Model That Thinks Before It Speaks
OpenAI's new model, o3-Pro, pushes generative AI beyond fluency and into thoughtful deliberation. Aimed at enterprise users and professionals who care more about accuracy than speed, o3-Pro is engineered for complex reasoning whether it's solving PhD-level science problems or conducting multi-step business analysis.
Notably, o3-Pro reportedly outperforms major competitors like Google Gemini 2.5 and Claude 4 Opus on industry-standard benchmarks such as AIME 2024 and GPQA Diamond. But users looking for casual chats may find it slower than its predecessor, GPT-4o.
The model is now available to ChatGPT Pro and Team users, though temporary chat memory is disabled due to technical issues, and image generation features are still pending.
Databricks Launches Agent Bricks
The company launched Agent Bricks, an automated system that allows enterprises to build, optimize, and evaluate AI agents using their own data without requiring deep ML expertise.
Agent Bricks isn't just another tool in the LLM toolkit. It tackles two of the biggest blockers to AI adoption: cost and quality. Most agent systems today rely on trial-and-error for tuning. Databricks replaces this with synthetic data generation and custom benchmarks that automatically calibrate agents for domain-specific tasks like legal document parsing or extracting information from maintenance manuals.
"Agent Bricks is a whole new way of building and deploying AI agents that can reason on your data," said Ali Ghodsi, CEO and Co-founder of Databricks. "For the first time, businesses can go from idea to production-grade AI on their own data with speed and confidence, with control over quality and cost tradeoffs. No manual tuning, no guesswork and all the security and governance Databricks has to offer. It's the breakthrough that finally makes enterprise AI agents both practical and powerful."
One compelling example is AstraZeneca which used Agent Bricks to transform over 400,000 clinical trial documents into structured data in under an hour without writing a single line of code.
"With Agent Bricks, our teams were able to parse through more than 400,000 clinical trial documents and extract structured data points without writing a single line of code. In just under 60 minutes, we had a working agent that can transform complex unstructured data usable for Analytics," noted Joseph Roemer, Head of Data & AI, Commercial IT, AstraZeneca
Google's New AI Architect
In a quieter but telling move, Google appointed Koray Kavukcuoglu, CTO of its DeepMind AI lab, as its first Chief AI Architect. The appointment, announced via an internal memo from CEO Sundar Pichai.
Kavukcuoglu will now serve as a Senior Vice President reporting directly to Pichai, while continuing in his role as CTO at DeepMind under the leadership of CEO Demis Hassabis.
He is set to relocate from London to California to take on this expanded mandate.
"In this expanded role, Koray will accelerate how we bring our world-leading models into our products, with the goal of more seamless integration, faster iteration, and greater efficiency," Pichai stated in the memo, underscoring the company's push for scalable, AI-first innovation across its ecosystem.
The leadership reshuffle comes at a time when Alphabet is under increasing pressure to demonstrate tangible financial returns from its heavy AI investments, with capital expenditures projected to reach USD 75 billion this year. The tech giant is also balancing regulatory scrutiny and intensifying competition in the AI space.
Meta's Foray Into AI Video Editing
Meanwhile, Meta is also expanding into generative video with its new AI video editing tools. Launched on its Meta AI app and editing platform 'Edits', users can now apply up to 50 preset transformations ranging from comic book aesthetics to sci-fi outfit swaps to short video clips.
While still limited to preset prompts and 10-second clips, the tools point to a future where consumer creativity meets AI augmentation. Meta hasn't confirmed whether its Movie Gen AI models are behind these features, but the trajectory is clear the company is pushing toward broader consumer AI adoption.
"We built this so that everyone can experiment creatively and make fun, interesting videos to share with their friends, family, and followers. Whether you're reimagining a favourite family memory or finding new ways to entertain your audience, our video editing [tools] can help," Meta said in its blog post.
Nvidia and Samsung's Bet on Physical AI
Meanwhile, Nvidia and Samsung are putting their money into robotics.
The two tech giants joined Japan's SoftBank in backing Skild AI, a robotics software startup, with Nvidia investing USD 25 million and Samsung adding USD 10 million. Skild is now reportedly valued at USD 4.5 billion following this Series B round.
The investment underscores a growing belief that the next frontier for AI isn't more virtual assistants but physical AI such as robots, autonomous systems, and machines that act on the world with intelligence.
