The enterprise is not ready for "the rise of the developer"

Key Concepts

Vibe Coding: Rapid software development often utilizing AI assistance, potentially bypassing traditional coding expertise.
Observability: The ability to understand the internal state of a system based on its external outputs (logs, metrics, traces). Expanding beyond traditional IT Ops to include developers.
Shift Left (and Everything to Developers): Moving responsibility for security, reliability, and observability earlier in the development lifecycle, ultimately placing ownership with developers.
AI-Assisted Development: Utilizing AI tools like GitHub Copilot and LLMs to generate and modify code.
Feature Flags: Mechanisms to enable or disable features in production, facilitating experimentation and reducing risk.
Agentic AI: AI systems capable of autonomous action and interaction, potentially automating tasks previously performed by humans.
Developer Empowerment: Providing developers with the tools and ownership to manage the entire SDLC.

The Evolving Role of Developers and the Rise of Vibe Coding

The conversation centers around the significant shift occurring in software development, driven by the emergence of “vibe coding” and the increasing role of AI. Sean Odell emphasizes that while development is expanding outside traditional observability boundaries – with citizen developers leveraging AI tools – observability remains crucial. The core issue isn’t whether code is written traditionally, but ensuring all code, regardless of origin, is understood, safe, and observable. He states, “Vibe coding needs observability… a developer, a professional developer, a citizen developer needs to understand what that test or what that experiment or what this latest release has caused.”

This necessitates a fundamental change in who needs observability. Historically focused on SREs, platform engineers, and IT Ops, observability is now becoming a critical requirement for developers themselves. Datrace is responding to this shift by integrating observability directly into the developer workflow, providing intelligent log analysis and contextual data. “It’s definitely a blending or an expansion of observability and contextual data at the same time.”

Reaching the New “Developers” and Datrace’s Strategy

Odell defines a “developer” as a “builder,” differentiating them from platform engineers. The challenge lies in reaching these new developers – including citizen developers – where they are. Datrace is actively engaging with this audience by participating in events like We Developers in Berlin and GitHub Universe, and evolving its presence at events like KubeCon to focus on software delivery first, rather than observability.

This strategy is exemplified by the launch of “Developers that Perform,” a practitioner-focused event featuring lightning talks, hackathons (with Microsoft and GitHub Copilot), and even presentations from individuals with non-traditional development backgrounds (e.g., an English major using vibe coding). The goal is to showcase the innovation happening in the development world and demonstrate how Datrace can support it, rather than simply pushing observability as a separate solution. “It’s about the citizen developer, the professional developer, and better yet… we want to give those SREs, those platform engineers… something in addition to what they’re doing on a daily basis.”

Addressing Concerns About Code Legitimacy and Security

A key concern with vibe coding is the legitimacy and security of AI-generated code. Odell acknowledges this, referencing a developer advocate’s point that “it’s not vibe coding if you know what you’re doing.” However, he frames this as an opportunity for learning and expansion. Even if developers don’t initially understand the underlying code (e.g., the language chosen by an AI), they can leverage observability tools to understand its behavior and impact.

He draws a parallel to the art of music, highlighting the creative aspect of development and the importance of embracing experimentation. “It is the art… it’s a beauty. And I think that’s one thing we have to embrace.” He emphasizes that feature flags are crucial for enabling experimentation in production while mitigating risk, allowing developers to own the entire SDLC, including reliability and SLOs.

Bottlenecks and the Future of AI in Development

The conversation addresses the potential for AI to shift bottlenecks in the development process. While AI accelerates code generation, the bottleneck may shift to deployment. Datrace advocates for embedding observability and security checks earlier in the pipeline, even during code generation, to prevent these issues from becoming roadblocks. This involves leveraging AI-powered agents to automate tasks like code validation and security scanning.

Odell points out that enterprises are surprisingly quick to adopt AI, faster than previous technological shifts like cloud native or virtualization. He believes the key to successful AI integration is empowering developers to own the entire SDLC and providing them with the tools to manage the associated risks. “The rise of the developer… enterprises have been slow to embrace, but when I have large Fortune you know 10 organizations look at me and go we have AI assisted development we have productions that are doing agent to agent communication no human involved already.”

Data and Statistics

95% of AI projects don’t make it to production: This statistic, sourced from Gardner (though the speaker couldn’t recall the exact publication), highlights the challenges of moving AI initiatives from experimentation to real-world deployment. Odell reframes this not as a failure rate, but as a reflection of the early stages of AI adoption and the learning process involved.

Technical Terms & Concepts

APM (Application Performance Monitoring): Traditional observability focused on monitoring the performance of applications.
SRE (Site Reliability Engineer): Engineers responsible for the reliability, scalability, and performance of systems.
CI/CD (Continuous Integration/Continuous Delivery): A set of practices for automating the software delivery process.
LLMs (Large Language Models): AI models capable of understanding and generating human language (e.g., used in GitHub Copilot).
MCP (Machine Learning Control Plane): A platform for managing and deploying machine learning models.
Agentic AI: AI systems capable of autonomous action and interaction.
SLO (Service Level Objective): A target level of performance for a service.

Conclusion

The conversation paints a picture of a rapidly evolving software development landscape, driven by AI and the rise of “vibe coding.” The key takeaway is that observability is no longer solely the domain of IT Ops; it’s becoming a critical requirement for developers themselves. Datrace is positioning itself to support this shift by integrating observability directly into the developer workflow, empowering developers to own the entire SDLC, and embracing the opportunities presented by AI-assisted development. The future of software delivery hinges on embracing the developer, fostering experimentation, and ensuring that all code – regardless of its origin – is observable, secure, and reliable.