The Software Wall: Why Robot Programming Is Still Too Hard — and What Olo Robotics Is Doing About It
Robotics is having a hardware moment. Humanoid robots are being ordered by the thousands. Quadrupeds roam factories on three continents. Robotic arms are cheaper and more capable than ever. The industry is rightfully celebrating how far the physical machines have come.
What's getting less attention is the software problem lurking behind all this hardware enthusiasm — and it might be the industry's most significant bottleneck. Eleanor Tang-Smith, COO of Olo Robotics, put it plainly in a recent interview with Robotics & Automation News: getting robots to do what you actually need them to do still requires specialists, and there simply aren't enough of them.
The Expertise Gap Nobody Talks About
Here's the hidden math of modern robotics deployment: a small-to-medium manufacturer can now buy a capable robotic arm or mobile platform at a price point that makes economic sense. What they can't easily buy is the programmer who knows ROS 2, understands motion planning, can configure perception pipelines, and troubleshoot the inevitable integration issues.
ROS 2 (Robot Operating System 2) is the dominant framework for programming most modern robots. It's powerful, flexible, and genuinely impressive as a software ecosystem. It's also notoriously complex for anyone who didn't come up through robotics research. Configuring a ROS 2 stack for a real production environment — with its sensor integrations, safety constraints, and need for reliability — is serious engineering work.
This creates a gap that's widening as robot hardware proliferates. The hardware is becoming a commodity; the expertise to deploy it effectively is not. Companies that can't hire or contract robotics software engineers are either stuck with expensive turnkey systems that don't fit their needs, or they don't automate at all.
Olo Robotics is building toward a different future: one where robot programming is accessible enough that the person who understands the production floor can actually configure the robot to work on it — without needing to understand ROS 2 internals.
The Democratization Thesis
Tang-Smith's framing echoes something the software industry has been working through for decades. When databases were only accessible to people who could write SQL, spreadsheets democratized data for millions. When web development required writing HTML from scratch, drag-and-drop builders and CMS platforms made it accessible to non-coders. Every major technology wave has eventually produced a simplification layer that brings the power to a broader audience.
Robotics hasn't had that layer yet. The current wave of robot deployments is still largely driven by the same specialist-heavy model that characterized industrial automation in the 1990s. What's changed is the hardware cost and capability — the deployment model is largely the same.
What Olo Robotics is building — and Tang-Smith was careful not to over-claim in the interview — is infrastructure to lower that barrier. The specifics involve cloud robotics architecture, simulation, and programming interfaces designed for operators rather than engineers. The goal isn't to eliminate robotics expertise; it's to ensure that expertise isn't required for every routine configuration and adjustment.
Why This Matters Now
The timing matters. The humanoid robot companies racing to commercial deployment are facing a version of this problem at scale. Figure AI, Agility Robotics, Apptronik, and their competitors are building general-purpose platforms — robots that can, in theory, do many different tasks. But programming those tasks, adapting them to new environments, and updating them as workflows change still requires significant expertise.
If humanoid robots are going to fulfill the "versatile worker" promise, they need to be reprogrammable without a robotics PhD every time a customer wants to change the workflow. That's a software infrastructure problem as much as a hardware one.
The parallel to the app store model is instructive. Smartphones became platforms when it became possible for a developer with a laptop and a good idea to build something millions of people could use. Robots will become platforms when it becomes possible for an operations manager with domain knowledge — but not a robotics background — to configure them for new tasks without calling in a specialist.
The Investment Angle
Accessibility-focused robotics software companies are an underappreciated segment of the investment landscape. Most of the public-market excitement is concentrated in hardware makers and AI chip companies. But the middleware and tooling layer — the companies making robots easier to deploy, program, and manage — could capture significant economic value as the installed base of robots grows.
Olo Robotics is currently private. But watching who they partner with, what capabilities they build, and whether they develop a developer ecosystem around their platform will be worth tracking. The companies that solve the programming accessibility problem early will have a distribution advantage that's hard to replicate.
For anyone wanting to understand the software architecture challenges in depth, Programming Robots with ROS remains one of the best introductions to why this problem is genuinely hard. And The Robot Builder's Bonanza offers a more accessible entry point for those approaching from the practitioner side.
For investors tracking the broader automation wave, platforms like Interactive Brokers offer exposure to publicly-traded robotics and automation stocks — and the tooling/software layer is increasingly worth attention as a distinct investment thesis.
The Bottom Line
The robots are ready. The software to deploy them broadly is not — or at least, not accessible to the people who most need to use it. Companies like Olo Robotics are working on that layer, and their success or failure will do more to determine how quickly robotics adoption accelerates than any improvement to hardware specs.
The hardware race is visible and celebrated. The software accessibility race is quieter. But in five years, the companies that won the software battle may matter more than the ones that won on motors and actuators.
---
Source: Robotics & Automation News — Interview with Olo Robotics COO Eleanor Tang-Smith