Blab AI
OpenAIs Greg Brockman Highlights Shift to Real-World AI Use at DataAI Summit
On June 16 in San Francisco, OpenAI president Greg Brockman turned the spotlight from model bragging to real‑world application at the DataAI Summit. He said the industry’s focus is moving from competing on model performance to deploying AI in everyday work.
The Databricks‑hosted summit drew more than 30,000 attendees and featured keynotes from Microsoft’s Satya Nadella, Databricks co‑founders, and other leaders. Brockman’s remarks were part of a broader conversation about how AI can be integrated into business workflows. He emphasized that model improvement is rapid, with a new GPT release each year, and urged companies to align their AI strategies around this continuous development.
Brockman explained that OpenAI’s role has shifted from building foundational models and infrastructure to applying those models in practical contexts. He said the organization is "focused on creating a way to build an agent that provides value to people and systems," and added that the company’s recent work on AI agents is aimed at making the models useful for a wide range of tasks.
A key point of the presentation was the potential for AI to handle up to 80 % of software‑development work. Brockman noted that earlier tools only assisted with simple tasks, but newer agents can now write code, debug, and manage documentation. He suggested that similar levels of automation could extend to other knowledge‑based activities such as presentation creation, document editing, and data analysis.
Data quality was highlighted as a critical factor for AI performance. Brockman compared data to the "bottom of a cake called a model," stressing that clean, well‑managed data can improve results beyond what model size alone can achieve. He noted that understanding how users interact with ChatGPT—currently the most widely used AI service—provides valuable feedback for product improvement.
The summit also covered OpenAI’s partnership with Databricks. The two companies announced a collaboration that makes OpenAI models available within Databricks’ data‑intelligence platform and its Agent Bricks workspace. The partnership is intended to help enterprises securely combine their data with OpenAI’s models at scale, with governance and performance controls.
Regarding artificial general intelligence, Brockman said that AGI is a spectrum rather than a single milestone. He added that the most important measure is the capability of AI to help people, not whether a model has reached a theoretical AGI threshold. He described the current state of AI as "jagged intelligence," noting that it excels in some areas while still showing limitations in others.
Brockman concluded that the present era offers a unique opportunity for creators and developers. He urged that AI should be directed toward expanding human capabilities and that the industry should maintain human control and goal alignment.
The DataAI Summit’s discussions underscore a growing consensus that the next phase of AI development will be defined by practical applications, data integration, and enterprise partnerships. OpenAI’s focus on agentic tools and its collaboration with Databricks signal a shift toward embedding AI more deeply in business processes, while the company’s ongoing model releases continue to push the boundaries of what AI can do.
The Databricks‑hosted summit drew more than 30,000 attendees and featured keynotes from Microsoft’s Satya Nadella, Databricks co‑founders, and other leaders. Brockman’s remarks were part of a broader conversation about how AI can be integrated into business workflows. He emphasized that model improvement is rapid, with a new GPT release each year, and urged companies to align their AI strategies around this continuous development.
Brockman explained that OpenAI’s role has shifted from building foundational models and infrastructure to applying those models in practical contexts. He said the organization is "focused on creating a way to build an agent that provides value to people and systems," and added that the company’s recent work on AI agents is aimed at making the models useful for a wide range of tasks.
A key point of the presentation was the potential for AI to handle up to 80 % of software‑development work. Brockman noted that earlier tools only assisted with simple tasks, but newer agents can now write code, debug, and manage documentation. He suggested that similar levels of automation could extend to other knowledge‑based activities such as presentation creation, document editing, and data analysis.
Data quality was highlighted as a critical factor for AI performance. Brockman compared data to the "bottom of a cake called a model," stressing that clean, well‑managed data can improve results beyond what model size alone can achieve. He noted that understanding how users interact with ChatGPT—currently the most widely used AI service—provides valuable feedback for product improvement.
The summit also covered OpenAI’s partnership with Databricks. The two companies announced a collaboration that makes OpenAI models available within Databricks’ data‑intelligence platform and its Agent Bricks workspace. The partnership is intended to help enterprises securely combine their data with OpenAI’s models at scale, with governance and performance controls.
Regarding artificial general intelligence, Brockman said that AGI is a spectrum rather than a single milestone. He added that the most important measure is the capability of AI to help people, not whether a model has reached a theoretical AGI threshold. He described the current state of AI as "jagged intelligence," noting that it excels in some areas while still showing limitations in others.
Brockman concluded that the present era offers a unique opportunity for creators and developers. He urged that AI should be directed toward expanding human capabilities and that the industry should maintain human control and goal alignment.
The DataAI Summit’s discussions underscore a growing consensus that the next phase of AI development will be defined by practical applications, data integration, and enterprise partnerships. OpenAI’s focus on agentic tools and its collaboration with Databricks signal a shift toward embedding AI more deeply in business processes, while the company’s ongoing model releases continue to push the boundaries of what AI can do.
