I still remember the first time I got my hands on a first-gen Chromebook.
It was 2011 and I was a computer science PhD student at the idyllic University of Washington. I was coming into my own as a tech guinea pig — Amazon had given my class large-screened Kindle DXs to test their feasibility for textbooks. And, instead of a laptop, I had opted for a desktop plus a Netbook with a screen and keyboard that were comically small for my 6' 4" frame.
Frustrated with the sluggish Kindle and Netbook, I immediately applied when Google announced its Chromebook pilot for its future of portable computing. But as the weeks passed without hearing back, I eventually forgot about it.
Then, one day, a nondescript box arrived. I unboxed the matte black CR-48 pilot laptop, and hurriedly booted it to experience the future: a web browser as an operating system
Alas, it was a product before its time. Web-based software development environments were severely lacking for my work as a computer scientist. On the personal side, the tinny speakers butchered music and Netflix wasn’t yet supported
Now, a decade later, has the promise of Chromebooks been fulfilled? I thought I’d give them another look now that Chromebooks outsold Macs in 2020.
A Chromebook is a laptop that runs Google’s Chrome OS. Unlike Windows and Mac, the primary interface in Chrome OS is the Chrome web browser. You log into a Chromebook with your Google account.
Indeed, Chromebooks outselling Macs in 2020 was likely a result of the COVID-19-induced surge in remote learning.
At PixieBrix, we come across Chromebooks constantly in industries such as finance, the public sector, and health care. In these industries, employees primarily use their computers for business processes and collaboration. The demand for simplicity, security, and cost effectiveness trumps the benefits of flexibility and computing power.
But Chromebooks are also breaking into high-end knowledge work. I have a data scientist friend who uses a Chromebook for his work at Google. His ability to do cutting-edge data science on a Chromebook has been enabled by advances in web-based development environments as well as the seamless provision of cloud resources. You can try it now — Google’s Colaboratory is a hosted data science environment (built on Jupyter) where you run code on a machine in the cloud. You can even train and run your models on custom hardware for machine learning and artificial intelligence.
In the coming years, we’ll continue to see a rapid expansion in the capabilities of web applications. There are three big enabling technologies: (1) WebAssembly means web applications can match the performance of native applications when running code locally, (2) Web Application Programming Interfaces (APIs) mean websites will keep up with native application functionality in interacting with the real world, and (3) infrastructure technologies such as Kubernetes, functions as a service (FaaS), and edge computing mean servers will be able to dynamically adjust to workloads with less latency.
Traditional robotic process automation (RPA) tools were designed for running unattended automation on desktop applications. The focus was on legacy applications — applications where there’s no web alternative and no APIs to program against — and on automating a process completely from end to end.
Companies then began to shift to web applications with rich APIs. In this world, you rely less on unattended RPA to automate a process. Instead, a junior developer scripts a series of API calls. Or, maybe even a business analyst wires up the systems in a no-code interface.
RPA vendors have already started to adjust for this shift toward web apps and APIs. Earlier this year, UiPath — the RPA leader — acquired Cloud Elements, an integration-platform-as-a-service (IPaaS) company, to complement its RPA capabilities.
RPA companies are also shifting focus to attended automation. Attended automation is where there’s a human in the loop, either triggering or partnering with the automation. The business challenge with attended automation is that the scale of processes is smaller and the ROI is harder to measure (especially for knowledge workers).
The rise of Chromebooks reflects the rise of web apps/APIs. The rise also presents a challenge to RPA companies. Traditional RPA tools were designed for business software running on desktops. Many tools can only run on Windows. Or, in cases where they can run on Mac/Linux, their automation capabilities are hamstrung (because of differences in application frameworks and accessibility APIs).
The reality is: To succeed, attended automation must be web — and API — first.
At PixieBrix, we’re embracing this reality. Deploying as a web browser extension, PixieBrix can run on any operating system — from low-end Chromebooks to bleeding edge M1 Macs. And by embracing standards such as JSON and OAuth, you can instantly connect to your software-as-a service (SaaS) or artificial intelligence, or even your unattended automations.
I have many more thoughts on the subject, but would love to hear from you! How are you adapting your automation strategy for a web-first world?