The Shift Down of Software Development
AI is the thing. It is hard to go an hour without learning about a new foundational model, large cloud providers releasing their suite of AI tooling or AI girlfriends generating significant income. AI is definitely the thing, but it will not be the outcome.
The hype is significant, but the second order effects of AI will be its lasting legacy rather than the foundational models themselves. Like many others, we feel strongly that AI will impact every industry and job function in some way, for better or for worse, including developers and the software development process.
Software development has undergone huge platform shifts over the past three decades with the creation of the internet, the migration from on-premises to cloud infrastructure and now, AI. With each wave, software release cycles have accelerated which has led to massive shifts in development practices. This has allowed for further iteration and releases of applications, but has also driven more responsibility towards developers including the shift left of necessary evils like testing and security.
Every platform shift has also changed the roles and responsibilities of the developer, and AI already has as well. In a recent conversation with engineering leaders from companies including Adobe, Navan, LiveRamp, Rippling, and others, the feeling was unanimous that we’re headed towards a future where code will be auto-generated, using platforms like ChatGPT, GitHub Copilot, Replit, Hugging Face and many other emerging open-source projects, but developers won’t disappear because of this. Instead, developers get to spend more of their time on important and higher-level decision making like technical specifications, design and infrastructure choices rather than the actual redundant process involved with executing on their vision.
Here’s my view of the platform evolution:
Pre-Cloud: Developers were primarily responsible for code, including the use of open-source libraries. Given the large monolithic architectures, coding was a massive undertaking across engineering teams which culminated in massive releases (for example, the launch of Windows XP). Infrastructure, including hardware, was managed by non-developers in broader IT, database administration or other roles.
Post-Cloud: The advent of distributed architectures, including microservices and cloud applications, gave way to developers being responsible for more of the development process including the shift left of processes like testing (from QA teams), security and infrastructure provisioning (databases, containers, Kubernetes) in order to get applications to production more quickly. This was all necessary to keep up with release speed expectations, but at the cost of developer productivity.
Post-AI: The prior waves have shifted more burden to developers, but AI is the wave that has already started to alleviate the pressure with the shift down of development operations away from humans. Coding has been the lifeblood of developers for decades, but the first one in will also be the first one out as developers will be able to focus on other, more impactful parts of the development process while still being responsible for critical parts of the production workloads including monitoring and security. Rather than spending the numerous hours to actually write the code and provision their infrastructure, developers will gain freedom to focus on what is being built versus how it’s being built.
The increased load on developers has led to major issues within enterprises such as ramping tech debt and environment complexity across distributed microservices, and AI is becoming a major driver to bring the joy back to developers by the shift down of many of their responsibilities.
As early-stage investors, CRV gets excited by major technological shifts that pave the way for emerging platforms to displace incumbents that have been the beneficiary of prior waves of technology. The AI advancements that we read about every day are exciting, but the outcomes that they will drive, buoyed by new technology, are what we can’t wait to see.