2020: The Year Everything Changed

I am writing this from my home office, which is to say my dining table, which is to say the same place I have worked every day for the past nine months. It is the last week of 2020, and I am trying to make sense of a year that resists easy summarization.

A year ago, I was commuting to an office, attending in-person meetings, flying to conferences, and operating under a set of assumptions about how work, technology, and society function that turned out to be far more fragile than any of us imagined. Those assumptions did not gradually erode. They collapsed in the span of about two weeks in March.

The Pandemic Acceleration

The defining characteristic of 2020 is not that new trends emerged, but that existing trends accelerated by five to ten years in a matter of months.

Remote work went from a perk to the default operating mode for knowledge workers. My team at a major entertainment company went fully remote in March and has not returned to an office since. Productivity, by every metric we track, has been sustained or improved. Deployments ship. Incidents get resolved. Architecture reviews happen. The work gets done.

The uncomfortable implication is that the office, as we knew it, was in many ways an artifact of inertia rather than necessity. The technology to work remotely has existed for years: Slack, Zoom, cloud infrastructure, Git, CI/CD pipelines. What did not exist was the organizational willingness to use it. The pandemic removed the option to be unwilling.

Cloud computing accelerated from a strategic direction to an urgent necessity. Organizations that had multi-year cloud migration timelines compressed them to months. The math changed overnight: if your workforce is remote, your infrastructure needs to be accessible from anywhere. On-premises systems behind VPN became bottlenecks. Cloud-native systems just worked.

At our organization, cloud adoption that was proceeding at a measured pace suddenly had executive air cover and budget priority. The teams that had been asking for resources to modernize their infrastructure got those resources, because the alternative was operational paralysis.

Digital transformation, a phrase I have always found vaguely irritating, became concrete. Businesses that relied on physical interactions (retail, entertainment, dining, healthcare) had to build digital capabilities overnight or face existential risk. Telemedicine went from a novelty to a necessity. Curbside pickup became standard. Streaming services saw subscriber growth that was not projected for years.

The Technology Stack of a Pandemic

Looking at the technology that made 2020 survivable, a clear stack emerges:

Video conferencing became the connective tissue of organizations. Zoom went from a niche business tool to a household word. The company's daily meeting participants went from roughly 10 million in December 2019 to over 300 million in April 2020. That is a 30x traffic increase in four months. From an infrastructure perspective, that is an extraordinary scaling challenge, and Zoom handled it imperfectly but adequately.

Cloud infrastructure (AWS, Azure, GCP) absorbed an enormous increase in demand as enterprises shifted workloads and consumers increased their digital activity. AWS reported accelerating revenue growth throughout the year. The cloud providers were, in many ways, the critical infrastructure of the pandemic, more so than most people realize.

Collaboration tools like Slack, Microsoft Teams, Notion, and Miro replaced the whiteboard, the hallway conversation, and the conference room. These tools are not perfect substitutes for in-person interaction, but they are good enough for most of the work that needs to happen. The 80% that is coordination, information sharing, and asynchronous decision-making translates well to digital tools. The 20% that is creative collaboration, mentoring, and relationship building does not.

Container orchestration proved its value. Our Kubernetes and ECS workloads scaled without drama because they were designed to be location-independent and horizontally scalable. The systems that struggled were the ones that assumed specific network topologies or required physical access to operate.

Personal Growth in Isolation

On a personal level, 2020 has been a year of uncomfortable growth. Working from home full-time, without the social structure of an office, forced me to be more intentional about how I communicate, how I manage my energy, and how I maintain relationships with colleagues.

I learned that I am more introverted than I thought, and that much of what I found draining about office life was the ambient social overhead rather than the work itself. I also learned that unstructured social interaction, the kind that happens in hallways and break rooms, serves a function that Slack channels do not fully replicate. Connection requires friction, accidental encounters, conversations that go somewhere unexpected. Optimizing away that friction also optimizes away the serendipity.

I spent more time writing this year than any year prior. Architecture documents, design proposals, post-incident reviews, internal blog posts. When you cannot walk to someone's desk and explain your thinking on a whiteboard, you write it down. This is, on balance, a positive change. Written communication is more precise, more reviewable, and more durable than verbal communication. The organizations that thrive in a remote-first world will be the ones that build strong writing cultures.

I also invested time in learning areas outside my immediate role. I went deeper on machine learning fundamentals, because the GPT-3 paper made it clear that AI will intersect with infrastructure engineering sooner than expected. I explored Rust, because it is becoming relevant for systems-level tooling. I read more broadly about distributed systems theory, because the problems I encounter at work keep mapping to concepts from academic papers decades old.

The Year in Releases

The technology releases of 2020 reflect the broader acceleration:

• Kubernetes 1.19 extended its support window to a year, acknowledging enterprise adoption at scale.
• Terraform 0.13 fixed provider management and module composition, addressing pain points that matter at organizational scale.
• GPT-3 demonstrated that language models at sufficient scale develop qualitatively new capabilities, reshaping expectations for what AI can do.
• Apple M1 shipped in November and validated that ARM can compete with x86 at the high end of the performance curve.
• AWS re:Invent announced ECS Anywhere, extending container orchestration to hybrid infrastructure.

Each of these, independently, would be a notable development. Together, they paint a picture of a technology landscape that is maturing in some areas (Kubernetes, Terraform) and leaping forward in others (AI, chip architecture).

What I Got Wrong

Intellectual honesty requires admitting what I misjudged this year.

I underestimated the pandemic. In January, I was aware of the virus in Wuhan and thought it would likely be contained like SARS. By February, I was paying closer attention. By March, I understood that containment had failed globally. My initial assessment was wrong, and I adjusted too slowly.

I underestimated remote work's viability at scale. I was skeptical that a large organization could function fully remotely for an extended period. I thought the loss of in-person collaboration would degrade velocity and quality. I was wrong. The degradation is real but manageable, and the gains in focus time and reduced commute more than compensate for most roles.

I underestimated how quickly cloud migration would accelerate. I thought our organization's migration timeline was aggressive. The pandemic made it look leisurely. When the alternative is operational failure, migration velocity increases dramatically.

Looking Forward

I do not know what 2021 holds. The vaccines are rolling out, which suggests a return to some form of normalcy in the second half of the year. But I do not think we return to the previous normal. Too many assumptions have been invalidated. Too many alternatives have been proven viable.

Remote work will persist in some hybrid form. Cloud adoption will continue accelerating. The tools and practices that organizations adopted under duress will become permanent because they work. The organizations that try to revert will find that their best people have recalibrated their expectations and will choose employers offering the flexibility they now expect.

On the technology side, I expect the AI trajectory defined by GPT-3 to continue, with implications for software development, content generation, and knowledge work broadly. I expect the ARM transition in personal computing to accelerate as Apple ships more M1 machines and the software ecosystem catches up. I expect multi-cloud and hybrid-cloud architectures to become the norm rather than the exception, driven by tools like ECS Anywhere and EKS Anywhere.

2020 was the year everything changed. Not because a single event altered the trajectory of technology, but because a global crisis compressed a decade of change into twelve months. The trends were already there. The pandemic removed the friction slowing their adoption.

We will look back at 2020 as an inflection point: the year the future arrived early, on terms we did not choose, at a cost we are still calculating.

The work continues. The infrastructure holds. We adapt.