Why AI Makes Us More Likely to Hire, Not Less

This is how we're thinking about AI at Alpha Theory including how it's changing how we build, hire, and invest.

We use AI across all parts of our business. Counter-intuitively, the work that is easiest to augment with AI is where the investment case is strongest.

By far, AI has impacted our software development group the most. It has improved, often by an order of magnitude, the time and quality of output.

A few examples: we cross-check code across multiple models before merging. We feed error logs directly to AI and debug instantly inside the codebase. We build working prototypes in hours that used to take weeks. Documentation that used to be an afterthought now informs both developers and the AI itself. And the deep work that used to require immense focus just to navigate a codebase? That is now delegated, freeing developers to focus on problems that actually require human judgment.

Said another way, software development we used to do in months now takes weeks. Weeks become days. Days become hours. And most of the code is higher quality, with higher fidelity to the original design.

So far, the biggest wins have been eliminating technical debt—modernizing systems, cleaning up legacy code, and strengthening our foundation. Faster feature delivery is coming, but we're building on solid ground first. (More on how we're modernizing our tech stack in a future post.)

AI has also changed what work is possible. For most of our history, Alpha Theory has been a software company that works with financial data. We're becoming something different: a data company that builds software. Projects that would have been too expensive—complex data pipelines, AI-powered analytics, intelligent assistants—are now within reach.

We're also planning to use AI internally as a project management catalyst and best practices repository—an internal "answer engine" that will help our team move faster by surfacing what we've learned, past decisions, and documentation when it's needed. This isn't just about efficiency; it's R&D. What we learn building this for ourselves will inform the answer engine we're building for customers.

So what do we do with all this? Three options:

1. Reduce headcount: Do the same with fewer resources
2. Flat headcount: Do more with the same resources
3. Increase capacity: Do much more with more resources

Now let’s assign some hypothetical ROIs (Return on Investment) to different roles before and after AI (We said hypothetical, so no one get upset with us for the ROI we assign to your group😊).

First note. Every role has higher ROI post-AI. And as we shift toward being a data company, data roles see gains similar to development. So where would you invest the next dollar? If you're thinking like a portfolio manager using Alpha Theory, you allocate to the highest expected return. The logic points towards going long on software and data.

There's something deeper happening, too. The nature of the development work is shifting. Skills that once distinguished senior engineers including systems thinking, managing complexity, and seeing the big picture are now expected earlier. The bottleneck isn't writing code anymore. It's clarity of intent, breaking work into pieces that can run in parallel, and steering AI effectively. We're still figuring out what this means, but it's changing how we think about what we need.

The world fears AI-driven job replacement. That's not what we're seeing. We see a chance to invest in the roles AI has supercharged thoughtfully, in ways that compound. The firms that figure this out early won't just be more efficient. They'll be building something their competitors can't easily copy.