In this episode, Brandon Jung, VP Ecosystem at Tabnine discusses the potential of generative AI (GenAI) in the enterprise space, emphasizing its augmentation of human productivity rather than replacement. Jung also discusses the current state and future prospects of GenAI in software development, highlighting the importance of starting with use cases, leveraging cloud infrastructure, and balancing control with accuracy in AI model development. Jung stresses the need for data stewardship, while emphasizing the platform’s flexibility to work with various models and cloud providers.

AI/GenAI in the enterprise space, with a focus on Tabnine’s experience and views on the technology

• Tabnine has been using GenAI for code for 10 years, adopting early models like GPT-2 and continuing to evolve rapidly.
• The company views big singular models as a way to go, but also uses tailored models for code use cases, being agnostic and looking at all fun models.

AI-assisted coding, its benefits, and adoption in enterprises

• AI-assisted tools will become essential for developers, starting with open-source projects and customized models.
• Jung points out that developers will need to adopt these tools to make new hires productive, while also instituting best practices across teams.
• GenAI for code is seeing most adoption due to existing controls in the software development lifecycle.
• Jung notes that enterprises should leverage GenAI for coding tasks that require customization, control, and personalization.

Using GenAI in software development, with focus on security and data quality

• Jung suggests starting with internal use cases like code, as it’s structured and has built-in security checks.
• Companies can learn data practices and stewardship by using GenAI, leading to better output.
• Jung points out that Tabnineoffers highly secure, fully permissive open source models for AI, with optionality to use other models or build custom ones.

AI technology adoption and future-proofing in software development

• Jung discusses the impact of Tabnine technologies on developers, time, resources, and cost.
• Developers are using Tabnine for code completion, unit testing, and other use cases, with mid-50 to 60% of their code coming from Tabnine.
• Jung highlights the importance of future-proofing investments in the rapidly evolving AI space.
• Tabnine focuses on enabling customers to pick up innovation as it emerges, rather than being locked into a single model or platform.

AI for code development, integrations with various tools, and customized suggestions for developers

• Customers want choice in running AI models, including on-premises or SaaS options.
• Jung states that Tabnine is open and flexible, enabling customers to choose their platform and models.
• Jung goes on to add that the evolution of AI for code will involve adding new data sets across the whole software development lifecycle.

AI adoption, control, and personalization in software development

• Companies want to know what data goes into AI models for transparency and accountability.
• Some companies have concerns about data privacy and security when using AI models.
• Jung discusses the importance of control in AI model adoption, highlighting the need for personalization and transparency.
• Jung mentions various ways to make AI models adoptable and build trust in their outputs, including giving users control over what goes into the model and how it’s used.

Guest: Brandon Jung (LinkedIn)
Company: Tabnine (Twitter)
Show: Let’s Talk