🚀 OpenAI's Journey to Human-Level AI

Welcome, AI enthusiasts

Exciting developments are underway at OpenAI as they unveil a new scale to measure AI's progress toward achieving Artificial General Intelligence (AGI). This scale defines stages from Level 1 to Level 5, with OpenAI’s models nearing Level 2. Meanwhile, Google Gemini is breaking new ground by integrating large language models with robots, boosting their capacity to perform real-world tasks while Elon Musk's xAI has announced a delay for Grok 2 until August. Let’s dive in!

In today’s insights:

• OpenAI's Journey to Human-Level AI

• Language Models Powering Real-World Robots

• Elon Musk Delays Grok 2

Read time: 4 minutes

🗞️ LATEST DEVELOPMENTS

OPENAI
🌟 OpenAI's Journey to Human-Level AI

Evolving AI: OpenAI made a system to determine how smart its AI systems are, ranging from Level 1 to Level 5.

Key Points:

• OpenAI's internal scale measures progress toward artificial general intelligence (AGI).

• Current chatbots like ChatGPT are at Level 1; nearing Level 2 indicates AI can solve basic PhD-level problems.

• Level 5 represents AI capable of performing the work of entire organizations.

Details:

OpenAI’s scale marks the journey from simple chatbot abilities to AI handling complex organizational tasks. Level 1 includes today’s chatbots; Level 2 means solving basic problems like a PhD. Level 3 involves taking actions on a user's behalf, Level 4 signifies creating innovations, and Level 5, the pinnacle, involves performing work equivalent to entire organizations. Despite differing timelines, AGI remains a distant goal, requiring immense computing power. The new scale clarifies progress, aiding OpenAI in its AGI mission.

Why It Matters:

Despite the challenges, the future of AI holds immense promise. OpenAI's human intelligence scale provides a structured approach to evaluate progress and identify areas that need further research and development. AGI is still only theoretical, meaning we’re not yet sure if it will be achieved, but we believe it can be. Many believe that we still need one or two breakthroughs before reaching AGI is possible. However, OpenAI introducing this scale gives us some hope that we’re on the right path. If they’re preparing themselves for reaching AGI, it’s probably best for you to do so too.

Become an AI-Powered Finance Decisions to get 10x ROI on your money (free AI masterclass) 🚀

More than 300 million people use AI, but less than 0.03% use it to build investing strategies. And you are probably one of them.

It’s high time we change that. And you have nothing to lose – not even a single $$

Join the most highly rated AI Masterclass! (free for first 100 users only) 🎁

Rated at 9.8/10, this masterclass will teach how you to:

• Do market trend analysis & projections with AI in seconds

• Solve complex problems, research 10x faster & make your simpler & easier

• Generate intensive financial reports with AI in less than 5 minutes

• Build AI assistants & custom bots in minutes

Register & save your seat now! (100 free seats only)

GOOGLE
🦾 Language Models Powering Real-World Robots

Evolving AI: Researchers are integrating large language models with robots, enhancing their ability to perform real-world tasks.

Key Points:

• Google DeepMind’s robot uses Gemini to understand and execute complex commands.

• Gemini’s multimodal capabilities significantly improve human-robot interaction.

• Industry and academic labs are rapidly advancing AI-enhanced robotics.

Details:

In California, a sleek, wheeled robot equipped with Google’s Gemini language model serves as a tour guide and office assistant. This Gemini-enhanced robot processes both text and video, enabling it to navigate and perform tasks based on commands like “Find me somewhere to write,” leading users to a whiteboard.

Introduced in December, Gemini’s multimodal nature—handling video and text—powers this robot’s impressive navigation abilities. Demonstrating up to 90% reliability, it excels even with complex instructions like “Where did I leave my coaster?” The researchers note the system's enhanced naturalness and usability in human-robot interactions.

This development illustrates the potential for large language models in practical applications beyond web browsers and apps. Gemini's ability to interpret visual data marks a significant leap in robot capabilities. Academic labs and startups are racing to leverage AI for robotics, with substantial investments fueling this innovation.

Why This Matters:

Researchers guided robots through real-world environments and showed them important locations. The robots were then able to find these locations again. A simple smartphone video is enough to give the robot an overview of the environment. The ease with which this can be achieved is significant and sets a precedent for robots being trained in the real world. Expect more and more similar headlines in the future with increasing capabilities of the robots, until this becomes a normal thing and we’re all used to it. Dystopian? Maybe. Understandably, not everyone will love these developments, but this will be our future.

GROK
🧠 Elon Musk Delays Grok 2

Evolving AI: Elon Musk's xAI delays Grok 2 to August, while teasing an even bigger leap with Grok 3 by year's end.

Key Points:

• Grok 2 postponed to August; Grok 3 promised by end of year.

• xAI leases 24,000 H100 GPUs; plans to train Grok 3 on 100,000 H100s.

• Moving from Oracle to build world's most powerful training cluster.

• Grok's chatbot lags behind ChatGPT; issues with fake news stories.

Details:

Elon Musk's AI company, xAI, announced a delay in the release of Grok 2, initially scheduled for May, now pushed to August. Musk promises Grok 3, a major upgrade, by year-end. Grok 2 was revealed in March, with Musk boasting it would surpass OpenAI's GPT-4, though specifics were scant. xAI has secured 24,000 H100 GPUs from Oracle for Grok 2's training; the delay is attributed to the meticulous process of purging training data of content from other language models. Musk claims this will yield a substantial improvement. Grok 3, teased in July, will utilize 100,000 H100s and commence training on xAI's in-house infrastructure, aimed at building the most powerful training cluster. This shift from Oracle is part of xAI's strategy to maintain control and outpace competitors. Currently, Grok's chatbot lags behind ChatGPT and Claude, partly due to generating fake news from real-time data on X.

Our Thoughts:

We all know that Elon Musk makes a lot of predictions for his companies that will not always be met. However, xAI has received notable investments in the last few months and is constructing a new AI data center in Memphis with NVIDIA chips. Will he meet his goals? He surely knows that if he wants to catch up and be a serious competitor in the space, Grok 2 needs to be a hit.

🚀 IBM: Teaching the language of your business.

💍 Samsung: Samsung announces their first-ever smart Galaxy Ring for $399.

📈 Trending AI Tools

• 🚀 Evolving AI’s Prompt Hub - The world's #1 ChatGPT Prompt Hub, featuring prompts that consistently produce great results (link)

• 🥗 Salad - Provides you with an affordable GPU cloud for your business (link)

• 📧 Breezemail - Organize your inbox with AI (link)

• 🤝 Pangea - Platform that connects startups with talent (link)

• 📂 Noah - An AI work assistant integrated with Google Drive, Notion, and more (link)

• 📧 Mailogy - Talk to your email with AI (link)