The Unbearable Lightness of Data // Rohit Krishnan // #295

In this section, Rohit and Demetrios discuss the potential of interactive reasoning models in AI. Rohit emphasizes the importance of being able to 'click into' the reasoning steps of AI models, allowing users to direct and adjust prompts more effectively. He believes this level of interaction can enhance the utility of AI, especially in generating reports and scientific inquiries, by allowing users to steer the conversation and refine outputs based on their unique contexts.

The conversation transitions to the implications of AI advancements on the job market. Rohit speculates on how roles, particularly in research and analysis, may change significantly as AI becomes more integrated into day-to-day operations. He argues that many professionals will need to adapt, potentially becoming 'training wheels' for AI as they help capture the necessary data and processes for future automation.

Rohit delves into his work at Bodo AI, particularly focusing on the company's efforts to streamline data processing and analytics within Python. He discusses the significance of efficiency in handling large datasets and the evolution towards open source, aiming to make data engineering and analytics more accessible. The implications of faster processing for AI training and other applications are highlighted, showcasing how efficiency can reshape data workflows.

As the dialogue evolves, Rohit reflects on the challenges and bottlenecks hindering progress towards Artificial General Intelligence (AGI). He emphasizes the need for tangible advancements in chip technology and energy efficiency, as these are critical factors for scaling AI capabilities. Rohit also raises concerns regarding the economic implications of such rapid technological growth, questioning the sustainability of investments in AI and their potential impact on job landscapes.