CareCredit - Women's Health April 2024

This Week in AI: Human-Like Reasoning, Robots and Open Source

artificial intelligence

This week in artificial intelligence (AI) news, a new technique, “Quiet-STaR,” teaches AI to think and chat more like humans by considering various ideas before answering, improving how we interact with technology. Significant investments in AI, including Nvidia’s robot project, GR00T, are rapidly advancing AI’s capabilities. 

Encouraging AI’s Inner Monologue’

Making AI smarter could involve taking cues from how the human brain works. 

A new method called “Quiet-STaR” is being used to teach AI to think before it answers, just like humans often do. This approach asks AI to come up with different reasons or explanations before responding to questions. This idea is part of a push to make AI think and reason more like humans do.

Venky Yerrapotu, the CEO of 4CRisk, a company that makes AI tools for checking rules and risks, shared his thoughts with PYMNTS. He believes AI will perform better if it can think more like a person. Human thinking is unique because it includes understanding the nuances, subtleties and context of conversations. AI that interprets conversations like a person may be more suited to handle complicated tasks.

Most AI chatbots today, like ChatGPT, offer answers without considering different possible conversations that could happen next. But with Quiet-STaR, AI is encouraged to guess what might be discussed next and learn from what’s happening in the conversation. This method has the chatbot generate various respomses, some with explanations and some without. Then, it picks the best one and forgets the rest that don’t make sense. This new approach is still being examined by researchers.

The takeaway goal is to make AI have an inner dialogue that considers different ideas before choosing what to say. This approach could lead to an AI chatbot that may serve as thinking partners, beyond just tools that spit out answers.

Leading Contributors Financing AI Advancements

More institutions are choosing to throw their financial weight behind AI. 

Saudi Arabia is making a bold move in the AI world by planning to invest $40 billion in it, as The New York Times reported. This move could make Saudi Arabia one of the most prominent players, alongside companies like Andreessen Horowitz and Nvidia, as it invests in AI to shape its future.

Businesses and investors see AI as a game-changer because it can transform how industries work, create new economic value, and give companies a boost over competitors. It’s expected to change how businesses operate, make decisions, interact with customers and develop new products. Consulting firm PwC believes AI could add up to $15.7 trillion to the global economy by 2030, thanks to productivity improvements and consumer benefits.

Here’s a look at some of the key players investing in AI:

Saudi Arabia-With plans for a $40 billion investment fund, Saudi Arabia aims to lead in AI. The country’s Public Investment Fund, which has over $900 billion, is talking with top venture capital firms like Andreessen Horowitz about joining forces for this investment.

Softbank: Led by CEO and Chairman Masayoshi Son, investment firm SoftBank Group wants to raise $100 billion for a new venture to compete in the AI chip market, going up against major players like Nvidia.

Google: Google has opted to invest in Anthropic, putting up to $2 billion into the company to help compete with other tech giants like Microsoft and OpenAI.

Meta: Meta CEO Mark Zuckerberg has said that AI is their primary focus for investment, and they are looking to use it to improve their ads and apps.

These investments highlight a global rush to get ahead in AI, showing how important this technology is for the future of business and the economy.

Nvidia’s Humanoid Robot

Nvidia is making waves in humanoid robotics with its latest AI innovation, Project GR00T, unveiled at its developer conference. 

This new foundation model is designed to revolutionize humanoid robots’ capabilities, enabling them to understand natural language and mimic human movements through observation and learning. Project GR00T aims to equip robots with essential skills for interacting within human environments, such as coordination and agility, by analyzing human actions.

Highlighting the significance of this advancement, Nvidia CEO Jensen Huang expressed his excitement at the conference. He said that creating foundation models for general humanoid robots is one of today’s most thrilling challenges in AI, with the necessary technologies aligning to empower roboticists worldwide to make substantial progress in artificial general robotics.

Further enhancing its robotics offerings, Nvidia has expanded its Isaac robotics platform with two key additions: the Isaac Manipulator and the Isaac Perceptor. The Isaac Manipulator introduces a suite of foundation models specifically for robotic arm control, while the Isaac Perceptor enhances robots’ visual processing abilities. 

The Rising Impact of Open Source AI

Elon Musk’s xAI is revolutionizing the AI landscape by open-sourcing its AI chatbot, Grok. This move is poised to democratize AI technology and spur innovation across various sectors.

Grok has been made accessible to X Premium+ subscribers via GitHub and BitTorrent, marking a significant step amidst Musk’s criticisms and legal battles with OpenAI over transparency issues. 

Richard Gardner, CEO of Modulus, weighed in on the implications: “Open source models could lead to more competition and, potentially, lower costs for businesses looking to leverage AI in their operations,” he explained to PYMNTS. Gardner highlighted a shift toward improving data quality and the application expertise of these models as they become more accessible.

Launched in November, Grok offers a ChatGPT-like experience for X Premium+ subscribers at a $16 monthly fee. At its core is Grok-1, a model boasting 314 billion parameters, dwarfing OpenAI’s GPT-3’s 175 billion parameters and hinting at its advanced capabilities.

Gardner said he supports open-source AI because it fosters quicker innovation and enhances transparency compared to proprietary models. 

“Open source supports free speech and reduces the potential for censorship by Big Tech,” he said, highlighting its potential to level the playing field for smaller companies and startups. 

Cloris Chen, CEO of AI firm Cogito, also recognized the power of open-source AI in accelerating development. “AI engineers can go much faster with collaborative improvements if they open up their source code, essentially posing a threat to centralized AI providers,” Chen told PYMNTS. 

This approach not only facilitates rapid advancements in AI but also challenges major tech firms’ centralized control of AI technology, making it a significant stride toward more open and collaborative AI development.