What the Dev?

In this episode, David Rubinstein talks with Craig Kerstiens, Snowflake's Director of Software Engineering on Postgres, about the enduring popularity of Postgres, a 30-year-old database technology, due to its reliability and recent innovations like JSON support and vector store databases. He highlighted the challenge of integrating operational (OLTP) and analytical (OLAP) data for AI applications, noting the importance of reducing latency and operational burden. Kerstiens introduced PG Lake, an open-source tool that embeds Iceberg in Postgres, simplifying data movement and querying. He also mentioned Snowflake's mirroring feature and the future of Postgres, emphasizing its continuous improvement and extension ecosystem.

Dave Rubinstein and Jonathan Macoskey discuss the limitations of current AI models in understanding context, particularly in enterprise settings. Macoskey explains that while language models have improved with internet connectivity, they still struggle with outdated information and proprietary data. He highlights Lovelace's context engine, which integrates data from various sources into a knowledge graph, addressing issues like entity resolution and recall. This system ensures AI agents can provide accurate, cited information, enhancing trust and decision-making. The conversation also touches on the challenges of scaling large language models and the importance of grounding AI responses in reliable data.

Dave Rubinstein, Editor in Chief of SD Times, discusses with Kumar Vikesh, Senior Product Owner at Progress Data Direct, the challenges of REST connectivity and the introduction of an AI-driven solution to automate the creation of model files for the Autonomous REST Connector. Progress Data Direct offers a standard connectivity platform with over 60 connectors and 100 data sources. The new AI-driven approach uses three agents: a generator, validator, and launcher, to create, validate, and test REST model files, reducing manual effort from days to hours. The solution is currently in an early access program, with plans for future general availability and integration with Visual Studio Code.

Dave discusses real-time voice AI with Mike Pappas, CEO of Modulate. Pappas explains that voice AI involves three key technologies: speech-to-text, language models, and text-to-speech. This technology is valuable in detecting fraud, ensuring policy adherence, and routing calls efficiently. 

In this episode, Dave speaks with John D'Emic, cofounder and CTO at Revenium, about the challenges organizations face in managing costs and ROI of agentic AI initiatives. John explains that the transition from pre-production to production often reveals gaps in economic visibility, leading to "agent debt" due to unanticipated costs from API calls and tool usage. 

In this episode with Dave Rubinstein, Cameron Etezadi, CTO of LaunchDarkly, discussed the challenges and misconceptions around Vibe coding. He highlighted that while code production has sped up, confidence in the quality of the code has decreased, with 91% of customers being less confident.  He noted that 94% of customers are shipping more code but struggle with validation and quality.

He noted that 94% of customers are shipping more code but struggle with validation and quality.

This week, Dave discusses the issue of Zoom calls freezing despite good Wi-Fi with Shafi Khan, Principal Engineer at Cable Labs. Khan explains that the problem arises from a lack of bi-directional communication between applications and networks. Shafi introduces the  Quality by Design framework based on intent-based APIs, which allows for real-time telemetry and end-to-end metrics. He explains the use of the open-source entity Kamara to provide APIs for developers and application builders.

In this episode, Dave Rubinstein speaks with Jaya Baloo, founder, COO and CISO at cybersecurity provider AISLE, about the implications of AI models like Anthropic's Mythos, which can discover and exploit vulnerabilities. The conversation touches on the existing challenges defenders face in keeping up with the volume of vulnerabilities, suggesting that AI has not made the situation worse but rather highlighted it.

In this episode, Dave Rubinstein, editor-in-chief of SD Times, sits down with  Karthik Sj, General Manager of AI at LogicMonitor, to discuss the challenges and progress of AI in enterprise production. Among the topics discussed are:

* The rapid advancements in AI, particularly in code generation and self-healing

* Emphasizing the importance of context engineering and agent execution engines

* Recommending that organizations start with an automation readiness report to assess their maturity and tailor their AI adoption roadmap

In this episode, Dave speaks with Mike Sirchuk, president, founder and CEO of Swa Technology, about providing organizations with multiple AI models to choose from, and bringing those tools to where people work today, such as in Slack teams, text messaging and Zoom, to name a few. The Swa platform also enables companies to build the security layer to manage who has access to what model.