print(“Hello AI World”): Evolution of the developer economy in the age of AI

A new chapter begins for the software development industry as AI elevates coding to new heights of abstraction and democratization, but classical programming will not be going extinct any time soon.

I’ve spent most of my venture career focused on developer tools and data infrastructure where I’ve had the immense privilege of partnering with highly technical founders. Within these spaces, I’ve made investments based on a few core guiding principles. Two tenets that I’ve found particularly compelling have been:

1. Abstraction: I view this as tools that automate low-level tasks or reduce the complexity of workflow completion so that developers can focus on higher-level responsibilities. Examples in this category could include platforms that remove the tedium of DevOps, hosting, and deployment.

2. Democratization: This refers to technologies that empower all employees in an organization, not just engineering talent, to contribute to software development. Examples in this category could include low code/no code offerings to build websites or create database queries.

I believe that it is important to be thesis-driven when investing in the developer economy as the landscape is constantly evolving — frameworks go in and out of favor, languages rise and fall in popularity, end user evangelism and word of mouth virality can make or break platforms, the stack is constantly bundling and unbundling, architecture and deployment models change… all of this dynamism makes the developer realm one of the most exciting areas to focus on as an investor.

Source: Goldman Sachs, “The Potentially Large Effects of Artificial Intelligence on Economic Growth”, (3/26/23)

The developer economy is currently going through one of its biggest transformations yet with the disruptive force of AI blazing through the ecosystem. As I’ve previously written, AI will impact the scope and skills of professions across numerous industries, but perhaps none more so than those of the developer since coding is a particularly well-suited use case for AI (chart above). This is due to a slew of reasons including the availability of training data and benchmarks, as well as coding being akin to predictive problem-solving through a progression of steps which bodes well for algorithmic functions.

Source: Morgan Stanley, “AI: Adopters Are Adopting” (2/11/24)

AI coding is not a distant or aspirational concept. Over the past few years, we’ve not only seen AI displace incumbent tooling (chart above), but we’ve also seen an explosion of new AI-powered coding tools across co-pilots, agents, and foundation models (a comprehensive market map below from my friend Corinne Riley at Greylock). In fact, in our recent State of the Cloud 2024 report, my colleagues and I at Bessemer highlighted how $3.9 billion VC dollars have been funneled into this area just in 2023. 

Source: Greylock, “Code Smarter, Not Harder” (5/22/24)

There is very real usage and dollars being spent on AI for coding use cases. For instance, GitHub Copilot’s revenue has reportedly exceeded $100MM ARR, with Satya Nadella commenting in Microsoft’s recent earnings call that “ Copilot accounted for over 40% of GitHub's revenue growth this year and is already a larger business than all of GitHub was when we acquired it”.

As AI adoption proliferates amongst developers, we’re starting to see signals of tangible productivity uplift and cost-savings:

• Research from Microsoft in May 2024 found that developers who used GitHub Copilot completed tasks ~55% faster than the developers who did not leverage Copilot.

• An early productivity study from Amazon found that  CodeWhisperer helped developers complete tasks 57% faster and with a 27% higher likelihood of success compared to developers who did not use CodeWhisperer.

• PayPal CEO Dan Schulman commented in a Fortune panel that “I do think that we are likely underestimating the impact that it [AI] will have. It will be pretty massive. For instance, we’re doing a lot of experimentation in code development right now, and we’re seeing a 30% improvement in productivity. Across all industry, front office, back office, you’re going to see 30% to 40% productivity increases.”

• In BP’s 1Q24 earnings call, CEO Murray Auchincloss highlighted “We have done an awful lot to digitize many parts of our business, and we are now applying Gen AI to it. The places that we are seeing tremendous results on are coding. We need 70% less coders from third parties to code as the AI handles most of the coding, the human only needs to look at the final 30% to validate it. That's a big savings for the company moving forward.”

While increased productivity and cost-savings are certainly key value drivers for AI adoption within the engineering domain, I’m more excited by another vector of value — AI’s impact on expanding the boundaries of the developer universe. Returning to my opening comments on democratization: AI has pushed this concept to new heights by making it easy for anyone to use natural language to develop code. This has significantly reduced the barriers to entry and enabled programming to become more accessible. In fact, Andrej Karpathy has made a provocative claim that “the hottest new programming language is English” as programming is starting to look more like conversational prompting as opposed to traditional coding in this age of AI. He reinvigorated discussions around this topic last week with another viral tweet about how fast the world of programming is changing in the current AI paradigm:

Source: Tweet from @karpahthy

These AI coding tools are not just a boon to non-technical stakeholders or junior devs. Going back to the concept of abstraction, because such tools augment developer workflows across all domains from SecOps to SRE to testing to code review… experienced developers can also benefit from the AI revolution by leveraging AI tools to relieve coding load on more low-level or mundane tasks such as bug fixing, refactoring, or code maintenance. This abstraction frees up precious time for them to concentrate on more intricate, creative, or meaningful tasks, such as new feature launches or complex integrations to drive competitive advantage and product differentiation. In his LinkedIn post from last week, Andy Jassy of Amazon provided some compelling anecdotes supporting this assertion:

Source: LinkedIn post from Andy Jassy

A new chapter begins

The levels of abstraction and democratization ushered in by AI coding have been so unprecedented that my colleagues and I at Bessemer predict that “by the end of this decade, development capability will be an accessible skill to most of the global population” and that “by 2030, a majority of corporate software developers will become something more akin to software reviewers. The cost of development will fall and as experienced developers become more productive their salaries will rise.”

Some observers have gone so far as to purport that it’s “The End of Programming” as software will be replaced by AI systems that are trained rather than programmed. While I do believe that AI will rapidly become entrenched in the backbone of software development, I don’t think that classical programming will being going extinct any time soon. For one, the accuracy of AI developer tools are not yet perceived to be at enterprise-grade reliability. Furthermore, enterprise requirements for AI-generated code go way beyond quality and involve aspects of privacy, security, compliance, scalability, and copyright.

Additionally, software development is a multi-faceted process that is as much an art as it is a science, involving judgement decisions and aspects such as ideation, design, systems, and delivery, that are oftentimes greater than the sum of the parts. To view programming as a rules-based process that can be completely replaced by algorithms would be a gross and dangerous oversimplification. Rather, AI-augmented engineering workflows can serve to elevate the role of the human programmer, moving them away from easily commoditized coding tasks into the realm of specialty skills, creativity, and esteemed expertise.

As I highlighted earlier in this piece, the developer economy is constantly adapting to various paradigm changes and potential impacts to job security. Yet the developer legion has remained extremely resilient, with devs being busier than ever and seeing increasing demand for their skills. Rather than rendering the human developer obsolete, AI can be a channel to amplify their impact. In a signal of confidence, Stack Overflow’s recent survey of ~65k coders globally revealed that ~62% of respondents are currently using AI tools in their development processes and ~68% of respondents do not view AI as an existential threat to their job:

Source: Stack Overflow, “2024 Developer Survey” (May 2024)

This post and the information presented are intended for informational purposes only. The views expressed herein are the author’s alone and do not constitute an offer to sell, or a recommendation to purchase, or a solicitation of an offer to buy, any security, nor a recommendation for any investment product or service. While certain information contained herein has been obtained from sources believed to be reliable, neither the author nor any of her employers or their affiliates have independently verified this information, and its accuracy and completeness cannot be guaranteed. Accordingly, no representation or warranty, express or implied, is made as to, and no reliance should be placed on the fairness, accuracy, timeliness or completeness of this information. The author and all employers and their affiliated persons assume no liability for this information and no obligation to update the information or analysis contained herein in the future.