Skip to main content

State-Of-The-Art

With leaders of the countries citing the influence of AI, now the race is off to private players with high-stakes race to develop Artificial General Intelligence (AGI), several tech giants are vying for the lead.

Russian President Vladimir Putin, "Whoever becomes the leader in artificial intelligence will become the ruler of the world. Chinese President Xi "Space, AI, and quantum computing and communication are China’s top technology priorities".

Artificial General Intelligence (AGI) Race!​

AGI is one of the most disputed concepts in technology. There are several definitions of AGI and more recently researchers from google deepmind have proposed a taxonomy for AGI. AGI may be regarded as form of AI that can understand, learn, and apply its intelligence to solve any problem, is considered the holy grail of AI research. Let's examine the key players: Google (with Gemini/Bard and DeepMind), Meta, OpenAI, Microsoft, and Amazon, and assess their progress and strategies in this ambitious endeavor. AGI Race

Google DeepMind​

Google Deepmind's vision is that in the coming years, AI — and ultimately artificial general intelligence (AGI) — has the potential to drive one of the greatest transformations in history. Google's approach to AGI is two-pronged. With Gemini/Bard, Google is focusing on leveraging its vast data and search capabilities to create responsible AI that can understand and interact with information in a more human-like way. Meanwhile, DeepMind, known for its groundbreaking work in the field of deep reinforcement learning - a combination of deep learning and reinforcement learning - and using games to push the boundaries of AI capabilities, as seen in their development of AlphaGo and AlphaFold So one players approach is to leverage a deep reinforcement based reward and policy learning framework to achieve generalization of AI tasks performed

Meta​

Meta (formerly Facebook) has been investing heavily in AI, particularly in areas that synergize with its social media and virtual reality ventures. Their AI research focuses on creating systems which are open source Llama language models that can understand and interact within complex social and virtual environments. Prof. Dr. Yann LeCun, Chief AI Scientist for Meta AI Research proposes new type of energy-based model architecture called H-JEPA (Hierarchical Joint Embedding Predictive Architecture). H-JEPA learns hierarchical abstract representations of the world that are simultaneously maximally informative and maximally predictable. Yann LeCun proposes there are other neural architectures apart from transformers which will help us enable autonomous intelligence. Yann refrains himself to use the word artificial general intelligence since, there is a misconception that AGI and AI will eventually over-power and take control of humanity hence he, refers to intelligence from machine learning to be autonomous intelligence.

OpenAI​

OpenAI, known for GPT-4 and its ethical approach to AI, is a significant contender. Their models have demonstrated remarkable language understanding and generation capabilities, which are crucial components of AGI. OpenAI's balanced focus on both capability development and safety is a unique aspect of their approach to AGI. The Mixture of Experts (MoE) model architecture, as used in systems like OpenAI's ChatGPT-4, represents a significant advancement in the field of machine learning and natural language processing. In the context of ChatGPT-4, the MoE architecture allows the model to handle a wide range of topics and question types more effectively than previous versions. It can provide more nuanced and accurate responses, especially in areas where specialized knowledge is required. However, the complexity and resource requirements of such a model mean that its development and deployment are significant undertakings, requiring substantial expertise and computational resources. Overall, the Mixture of Experts model represents a powerful approach in the AI field, offering high scalability and specialization, but it also comes with challenges related to its complexity and resource demands.

Microsoft​

Microsoft's strategy in the AGI race involves leveraging its cloud computing prowess and investing in promising AI ventures, including a significant partnership with OpenAI. Their approach is more about creating an ecosystem where AI can evolve towards AGI, supported by robust computing infrastructure and diverse data within Microsoft Azure ecosystem. Microsoft acquiring OpenAI for 10Billion dollars and its partnership to roll out chat-GPT based models are seen as game changer for industry particularly Microsoft().

Amazon​

Amazon's foray into AGI is less publicized but equally significant. Their work in AI is deeply integrated into their vast e-commerce and cloud services, focusing on practical AI applications that gradually push towards more generalizable intelligence. Amazon's strength lies in its ability to deploy AI at scale in real-world scenarios and its recent partnership with Anthropic. Amazon and Anthropic are each engaged across a number of organizations to promote the responsible development and deployment of AI technologies, including the Organization for Economic Cooperation and Development (OECD) AI working groups, the Global Partnership on AI (GPAI), the Partnership on AI, the International Organization for Standardization (ISO), the National Institute of Standards and Technology (NIST), and the Responsible AI Institute. In July, both Amazon and Anthropic joined President Biden and other industry leaders at the White House to show their support for a set of voluntary commitments to foster the safe, secure, responsible, and effective development of AI technology. These commitments are a continuation of work that both Amazon and Anthropic have been doing to support the safety, security, and responsible development and deployment of AI and will continue through this expanded collaboration.

Zarqa​

Zarqa represents an ambitious step in the evolution of LLMs, focusing on grounding AI in reality and ethical considerations. While it shares some similarities with ChatGPT-4, such as advanced language processing capabilities, its unique approach to integrating knowledge graphs and blockchain technology sets it apart. The success of Zarqa will depend on how effectively it can integrate these technologies and scale up to match the performance of established models like ChatGPT-4.

Strengths and Weaknesses of Zarqa Compared to OpenAI 's ChatGPT​

Strengths of Zarqa:​

-Integration with knowledge graphs could provide a more grounded and reality-based response system. -Emphasis on ethical considerations and security, potentially offering more predictable and interpretable AI behavior. -Blockchain integration for resource management and model testing could enhance transparency and decentralization.

Potential Weaknesses:​

As a newer project, it may initially lack the extensive dataset and refinement seen in OpenAI's ChatGPT-4. The complexity of integrating various technologies (neural-symbolic AI, blockchain) might present developmental challenges.

Conclusion​

The race to AGI is not just about technological prowess but also about how these technologies are integrated into our digital and physical worlds. Each contender brings unique strengths to the table: Google's advanced research capabilities and vast data resources. Meta's focus on AI in social and virtual environments. OpenAI's commitment to ethical AI development. Microsoft's strategy of building a supportive AI ecosystem. Amazon's practical application of AI at scale. The winner of the AGI race will likely be the one that not only advances the technology but also addresses the ethical, social, and economic implications of such a trans-formative development. Zarqa is a project by SingularityNET, a next generation of decentralized AI focusing on developing a neural-symbolic large language model (LLM) and hand over the power to control AI to hands of people rather than BigTechs or governments However it would be interesting what regulatory implications one have to face while deploying this exciting technology. It aims to equal or exceed current market tools by leveraging decentralized infrastructure and advanced computing architecture. The concept of whether Artificial General Intelligence (AGI) is a "winner-take-all" game is a subject of debate among experts. While some argue that the first entity to develop AGI could gain a significant, potentially unassailable advantage in various fields, others believe that the nature of AI development is collaborative and incremental, suggesting that multiple players could coexist and contribute to AGI advancements. Additionally, the global impact and ethical and safe considerations surrounding AI and AGI development might necessitate shared governance and oversight, rather than a single entity dominating the field.