#285 Getting Depth and Value From Generative AI - In Data Mesh and in General - Zhamak's Corner 33
Jan 12, 2024
auto_awesome
The podcast explores the current impact and challenges of generative AI within data mesh. It emphasizes the need for reliable access to data products for machine learning engineers. The role of generative AI in data discovery is also discussed, including its potential in generating data models and providing context for data products.
Generative AI has not yet proven to be a significant game changer, despite advancements in chatbots.
To fully leverage generative AI, more metadata around data products is needed for improved insights and value.
Deep dives
Generative AI and its Limitations
In this episode, Jean-Mack discusses the limitations of generative AI and its application in delivering deep value. While there have been advancements in chatbots, Jean-Mack points out that there has not been a significant transformation in people's ability to perform deep work. The focus should be on giving data developers easy and reliable access to data products, reducing the time spent on non-value adding tasks. Another challenge is the limited metadata available around data products, which hinders the full leverage of generative AI. Overall, while there is potential for generative AI, more data and improved data access are needed to unlock its value.
Data Mesh and Generative AI
Jean-Mack emphasizes the need for data mesh to support generative AI and machine learning work. Data developers, whether in analytics, ML, or generative AI, should have an easy and reliable path to work with data products. The concept of natively accessible data products allows for interoperability, where the same data product can be used for training ML models and creating dashboards. By implementing data products with the characteristics of data mesh, such as discoverability and value, opportunities arise for advancements in generative AI. Jean-Mack believes that the evolution from manual to co-pilot to autopilot in data product generation can lead to significant possibilities.
The Potential of Generative AI in Data Discovery
Jean-Mack discusses the potential of generative AI in data discovery and its role in providing deep insights. With a multidimensional representation of data and code, generative AI can fundamentally change how people search for data products. However, she acknowledges that the current offerings in generative AI are often shallow, lacking the richness and depth necessary for users to trust the insights provided. Jean-Mack highlights the importance of delivering valuable and insightful answers to prevent users from abandoning generative AI in favor of traditional search methods. Despite the need for further advancements, she sees great potential in leveraging generative AI in data discovery.
Thus far, most of the generative AI stuff Zhamak has seen is not that much of a differentiator. They are doing far better chat bots but that hasn't really changed the game.
When it comes to any ML work - and GenAI is just a subset of ML work - engineers need data products to make their data work easy. Reliable sources of data, ability to version, etc. Data mesh obviously plays well there.
Relatedly, we need to continue to make things easier for people to leverage data products for GenAI. Engineers shouldn't have to spend all their time moving data around and using many systems.
GenAI really could be game changing in data mesh but right now we don't have enough information to really do it well. We need far more metadata around things like data products.
GenAI often gives extremely shallow answers that just aren't that helpful. If we can get better answers, amazing. But right now, it's not there.
Sponsored by NextData, Zhamak's company that is helping ease data product creation.