Bill Gates thinks GPT-5 will be a bust

Plus: OpenAI takes bold step to counter catastrophic AI risks

Happy Friday! Here is what we got for you today:

  • ⚠️ OpenAI takes bold step to counter catastrophic AI risks

  • 🤖 Nvidia's Eureka AI: teaching robots new tricks with GPT-4 power!

  • 💻 Bill Gates thinks GPT-5 will be a bust

Unlock the power of swift and accurate responses with Answer Pursuit's multiple-choice answering tool, meticulously crafted for students like you who are in pursuit of instant answers during homework sessions or exams. Why get tangled in the web of crafting the perfect input for ChatGPT when AI can seamlessly serve you with precision and speed? Simply copy, paste, and witness the magic unfold.

The best part? It’s absolutely free! Dive in and experience efficiency like never before – try Answer Pursuit now!

In a proactive move towards ensuring safer AI, OpenAI has initiated a dedicated team to tackle the looming "catastrophic risks" tied to Artificial Intelligence (AI). This initiative underscores a significant stride in not just identifying but also fortifying against grave threats that AI could potentially unleash.

Key Points:

  • Preparedness Team Formation: OpenAI has assembled a new team aimed at tracking, evaluating, and safeguarding against severe issues that may emanate from AI, including nuclear threats.

  • Broad Spectrum Risk Mitigation: The team will also focus on thwarting chemical, biological, and radiological threats alongside “autonomous replication” - a scenario where AI replicates itself autonomously.

  • Human Deception and Cybersecurity: Addressing AI's ability to deceive humans and counter cybersecurity threats forms a crucial part of this initiative.

  • Leadership: The helm of this preparedness team is in the hands of Aleksander Madry, who is temporarily on leave from his position as the director of MIT’s Center for Deployable Machine Learning.

We're thrilled to share some groundbreaking news from the world of artificial intelligence and robotics! Nvidia Research has just unveiled Eureka, a cutting-edge AI agent with the mighty capabilities of OpenAI's GPT-4.

Key Points:

  • Autonomous Learning: Eureka is a marvel in AI innovation, autonomously crafting reward algorithms to teach robots a wide array of complex skills.

  • Human-like Dexterity: The AI agent has successfully trained a robotic hand to perform intricate pen-spinning tricks with a finesse comparable to a human!

  • Versatile Skill Set: From opening drawers and tossing balls to deftly manipulating scissors, Eureka has imparted nearly 30 different tasks to its robotic learners.

  • Challenging the Status Quo: Nvidia aims to revolutionize reinforcement learning and tackle existing challenges, making strides with Eureka’s integration of generative and reinforcement learning methods.

  • Open-Source Opportunity: The Eureka library of AI algorithms is now available for public experimentation, powered by Nvidia's Isaac Gym for reinforcement learning research.

Generative AI has taken the tech world by storm, evolving at a breakneck pace and leaving many of us awe-struck. But is this rapid progress sustainable? Bill Gates, the visionary founder of Microsoft, recently sat down with Handelsblatt, a leading German business newspaper, to discuss his take on where generative AI, specifically GPT technology, is headed.

Key Points:

  • Hitting a Plateau: Gates suggests that generative AI might have reached its peak, highlighting the enormous leap from GPT-2 to GPT-4 as a benchmark for significant improvement. Despite this, he acknowledges the possibility of being proven wrong.

  • Skepticism on GPT-5: Although many brilliant minds at OpenAI, including CEO Sam Altman, are confident that GPT-5 will surpass its predecessor, Gates remains skeptical, believing we might have hit a ceiling with the current technology.

  • Efficiency and Reduction in Costs: OpenAI is reportedly focusing on prototypes aimed at boosting model efficiency, which could potentially lead to lower inference costs.

  • A Bright Future, Nonetheless: Gates sees substantial potential in today's AI systems, particularly in medical fields like drug development and health advice, provided development costs decrease and reliability increases—achievements he foresees within the next two to five years.

Quick Hits

  • Databricks Predictive Optimization intelligently optimizes your table data layouts for improved performance and cost-efficiency.

  • OpenAI's new AI, DALL-E 3, creates images from text more accurately than ever before, thanks to its new image labeler.

  • Midjourney, the AI art generator, now has a new website and a major update in the pipeline, bringing image generation to the browser.

  • Forbes' new AI search engine, Adelaide, helps you find information and engage with Forbes stories in new and exciting ways.

  • AI, machine learning, and robotics are transforming retail supply chains, helping retailers meet fast-changing customer expectations and become more agile and resilient.

Tools of the Week

  • 👨‍💻 CodiumAI: AI-powered code testing

  • 📝 Jasper: Write great marketing content

  • 📧 nureply: Send AI-powered cold emails

AI Image of the Week