The Birth of AGI: From Rough Consensus to Running Code

The Birth of AGI: From Rough Consensus to Running Code

In 1992, computer scientist David D. Clark made a bold statement: "We reject: kings, presidents, and voting. We believe in: rough consensus and running code." This sentiment encapsulates the essence of technological progress - instead of relying on hierarchical decision-making processes, Clark advocated for a collaborative approach where ideas are tested and refined through practical implementation.

Fast forward to the present day, and we find ourselves on the cusp of a groundbreaking development in artificial intelligence - the birth of the Artificial General Intelligence (AGI). This monumental achievement brings together the power of language models and the ability to generate text that follows specific instructions. The latest iteration, the GPT-3.5 series, has surpassed its predecessors in terms of zero-shot generation and long-term memory capabilities.

With a long-term memory of up to 8192 tokens, the GPT-3.5 series can process and generate text that is twice as long as the previous GPT3 model. However, it still falls short in certain areas. It struggles with mathematical calculations, generates false information about the real world, and even writes faulty code. These limitations prevent it from passing Turing, SAT, or IQ tests.

Despite these shortcomings, the GPT-3.5 series finds its greatest utility in areas where creativity and innovation take precedence over precision. In brainstorming sessions, drafting documents, and presenting information in creative ways, this AI model shines. Its ability to generate text based on instructions provides a valuable tool for those seeking inspiration and novel approaches to problem-solving.

One of the ongoing debates surrounding AGI is the extent to which a ChatGPT-like experience can replace traditional search engines like Google. On one hand, ChatGPT excels at providing direct and legible answers to questions, surpassing the sometimes convoluted results of a typical Google search. However, it must be noted that these answers are not always accurate and lack proper sourcing. This raises questions about the reliability and trustworthiness of information provided by AI models.

Nevertheless, the birth of AGI marks a significant milestone in the field of artificial intelligence. It represents a leap forward in our understanding of Reinforcement Learning via Human Feedback. The pace at which we have advanced in this area is truly astonishing and has far exceeded expectations.

As we navigate this new era of AGI, it's crucial to consider the implications and harness its potential responsibly. Here are three actionable pieces of advice to guide us in this exciting journey:

1. Embrace collaboration and consensus-building: Just as David D. Clark advocated for rough consensus and running code, we must prioritize collaboration and cooperation in the development and deployment of AGI. Bringing together diverse perspectives and expertise will ensure a more well-rounded and balanced approach.

2. Augment AI with external assets: While AGI models may fall short in certain areas, such as accuracy and fact-checking, we can compensate for these limitations by combining them with external assets. Incorporating reliable sources and verification mechanisms can enhance the quality and reliability of the information generated by AI models.

3. Foster transparency and accountability: As AGI becomes more integrated into our daily lives, it's crucial to establish transparency and accountability frameworks. This includes ensuring that AI-generated content is clearly identified as such and providing mechanisms for users to verify and fact-check the information presented.

In conclusion, the birth of AGI represents a significant milestone in the field of artificial intelligence. It embodies the principles of rough consensus and running code, highlighting the importance of collaborative development and practical implementation. While AGI models have their limitations, they offer immense potential for creativity and innovation. By embracing collaboration, augmenting AI with external assets, and fostering transparency and accountability, we can navigate this new era responsibly and reap the benefits of AGI's remarkable capabilities.