The answer to this pivotal question will affect everything from education to venture capital
This past October, Google CEO Sundar Pichai shared that over 25% of all new code at Google was being created by AI, with human engineers reviewing and accepting it. It was a big moment to see the world’s third most valuable company prove that AI can have massive effects on software engineering.
Just a few weeks ago, Anthropic announced the launch of their Economic Index. There’s one key theme that’s hard to miss from their first report: computer programming, software engineering, and data analysis are where the gains are happening. 37.2% of conversations with Claude, Anthropic’s AI assistant, are occurring in this category, yet these roles represent 3.4% of U.S. workers, albeit some of the most highly paid workers.
To give a comparison, 0.3% of Claude conversations are in healthcare support, even though that category represents 4.7% of U.S. workers. The rate at which Claude is supporting software engineers is disproportional to say the least (note: Anthropic rightfully calls out that there is probably some skew in this data because their AI is designed to be great for engineers to create code).
From Google’s update to Anthropic’s data, we know that AI is allowing software engineers to become more productive. A lot more productive. But there’s a big question we need to answer: will we need to increase the number of software engineers from where it stands today?
For years, tech companies both in and out of the Bay Area have complained about how hard it is to find strong fits in software engineers for their company. It felt like there would never be enough software engineers to keep up with the pace of tech growth.
Is that changing?
There’s a question embedded within this broader puzzle: have organizations only been able to hire engineers up to the point that they can get the bare minimum done on their roadmap (due to a lack of supply of engineers)? To speak in more startup language, have companies had a restricted quantity of engineers that only gave them the bandwidth to address the proverbial “hair-on-fire” problems? Are there months worth of “tech debt” on which additional engineers would be able to help?
Perhaps that brings us to the most important question of all: is the excess capacity we are going to unleash in software engineers (as a result of the increased productivity thanks to AI augmentation) going to purely unlock a much longer list of things to do, or are we truly not going to need an increase in the number of engineers going forward?
A lot hinges on these questions, I believe. It’s worth exploring some of the potential implications.
What do the non-top engineers do? If you believe that software engineers are not going to get replaced overnight by AI, then the very top engineers (especially those who are great at utilizing AI’s augmentation) will continue to have jobs. But how about those who aren’t the best-in-class? Or the recent grads who need more training? Due to the undersupply of software engineers over recent years, the unemployment rate has been low. Will that remain the case?
Does this affect the education system? According to the National Student Clearinghouse Research Center, there was a two-fold increase over the last decade in the number of students obtaining bachelor’s degrees in Computer and Information Systems. Colleges and universities around the world have responded by increasing their efforts on building relevant programs. Should they continue to do so if the need for software engineers is going to decline because of AI?
Do we need less of a venture capital ecosystem? For most tech startups, the number one reason to raise money is to be able to hire a team and acquire users. Human capital (namely R&D costs) is expensive, especially engineers. If AI is really making engineers as productive as the data suggests, a company may need to raise less money because it needs fewer engineers to get to every milestone, whether it be a minimum viable product, $1 million in annual recurring revenue (ARR), $10 million ARR, etc. Are the days of $10 million Seed rounds and $50 millions Series A’s over, and will we opt to see founders raise way less in order to give up less dilution? This feels hard to conceptualize in a time when we see mega funds raising bigger funds than ever before, which means they need to either write bigger checks or many more checks than before.
How about the compensation of software engineers? The supply-demand imbalance has led to the role being one of the mostly highly compensated professions. If we reach a saturation point of engineers, it can only be expected that we will see the wage growth for the role slow down. The average base salary for a software engineer in the US is ~$106,000 (~$161,000 in the Bay Area) according to Indeed. It is one of the most highly compensated roles in our current economy and certainly the role that AI is most showing an ability to help, according to Anthropic.
And so many other ripple effects. Contingency recruiting firms often get paid 15-30% of a software engineer’s first year salary. If there’s a decreasing need to hire software engineers, what happens to these recruiters and businesses? Tools that support engineering teams, ranging from GitHub to LaunchDarkly to DataDog, have charged on a per-seat basis. What happens if the total number of seats in the ecosystem stays constant or goes down? The list of ripple effects goes on.
Where do we go from here? Like with many things related to AI, it feels too early to say. The goal of this piece was to open a line of questioning that is important to the future of the tech, venture capital, human capital, and education ecosystems. Whether you’re an engineer yourself, VC, startup founder, recruiter, or in another role, I’d love to hear your perspective. Perhaps I will even do a follow-up piece that shares feedback from those who have a perspective they’d like included.
Chuck Isgar (MBA ’25) loves all things startups. He created and runs Above Board, a weekly newsletter which features Q&A's with investors and founders about startup investment, board management, and corporate governance. Most recently, he served as the Chief of Staff at Scenery, a Series A-stage startup backed by investors such as Greylock. Previously, he was a Schwarzman Scholar where he earned a Master in Global Affairs from Tsinghua University in Beijing, China. Chuck co-founded and was the CEO of Intern From Home, a recruiting technology startup that served students from over 600 colleges and was featured in publications such as The New York Times. Chuck earned his bachelor’s from Brown University, where he served as the Co-President of the Brown Entrepreneurship Program. He loves to golf, cook, and go on long walks.