Lenny's PodcastThe future of AI in software development | Inbal Shani (CPO of GitHub)
At a glance
WHAT IT’S REALLY ABOUT
GitHub’s CPO on AI copilots, developer happiness, and software’s future
- Inbal Shani, Chief Product Officer at GitHub, explains how AI tools like GitHub Copilot are reshaping software development by augmenting, not replacing, engineers. She argues the real shift is in developers’ thinking: from writing code to designing systems, architectures, and connected experiences, with AI handling more of the rote work. Shani highlights under-explored opportunities like AI-driven testing and discusses how GitHub measures Copilot’s impact through productivity, quality, security, and, ultimately, developer happiness. She also shares lessons on adopting AI in organizations, GitHub’s design philosophy, its research engine (GitHub Next), and what it takes to lead modern product teams.
IDEAS WORTH REMEMBERING
5 ideasAI will augment developers, not replace them, for the foreseeable future.
Shani emphasizes that Copilot is a “copilot, not a pilot” and that human creativity, innovation, and judgment remain irreplaceable. AI needs humans both to generate data and to provide the creative spark that drives meaningful software.
Developers must shift from code-centric to system- and architecture-centric thinking.
As AI handles more low-level coding, developers—especially juniors—can focus earlier on understanding systems, environments, and products. This accelerates their growth toward senior, big-picture thinking rather than just syntax and boilerplate.
AI-driven testing and quality assurance are massively under-utilized opportunities.
Shani argues that as AI increases code volume, automated unit, integration, load, performance, infrastructure, and security testing become critical. AI can generate and expand test suites at scales that manual effort can’t match.
Adopting AI requires intentional change management, not just handing teams a tool.
Many companies drop AI tools in and expect magic. GitHub’s experience is that you must start from the customer problem, integrate AI where it truly reduces friction, and actively guide organizational behavior change around new workflows.
Productivity metrics must move beyond speed and lines of code to time-to-value and happiness.
GitHub looks at a mix of measures: faster code reviews, better security, fewer leaked secrets, and developer sentiment. Pure time or volume metrics are misleading because you can write bad code very quickly; the goal is valuable outcomes and satisfied engineers.
WORDS WORTH SAVING
5 quotes“Copilot is a copilot, it’s not a pilot. You still need the human in the loop.”
— Inbal Shani
“AI cannot replace innovation, that creative spark that is the center of humanity.”
— Inbal Shani
“You can write really bad code really fast. Time alone is not a success metric.”
— Inbal Shani
“If you don’t take risks, you cannot create a future.”
— Inbal Shani
“If we cannot use our own tools, our customers cannot use them for sure.”
— Inbal Shani
High quality AI-generated summary created from speaker-labeled transcript.
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome