Force #3 -- Role Consolidation
The future does not belong to the deepest specialist or the broadest generalist. It belongs to people who can execute across the functions where there is the most friction.
In my last post, I explored how specialization gets democratized and how AI empowers novices to execute. Today, we examine how roles are consolidating around whoever can execute, not whoever holds the title.
How the world changed under our noses
Between 2022 and 2024, tech companies laid off over 580,000 people. In 2024 alone, Instagram eliminated its entire technical program manager function, consolidating sixty roles into product management. By the second quarter of 2023, UX research and design job listings had dropped below their 2021 levels and never recovered. The World Economic Forum projects that 23% of all jobs will change by 2027, with 83 million positions eliminated as roles consolidate and redistribute across organizations.
I watched this transformation happen in real time, not from the outside looking in, but from the center of it. What I experienced at Uber and Amplitude was not unique; it was a microcosm of a structural shift that reshaped how knowledge work gets distributed across the technology industry.
The moment I saw consolidation in real time
In May 2020, two months into the job as a Sr. Research Manager, I was laid off from Uber, along with 25% of the company, at the height of the pandemic. Two months later, they called me back, and I ultimately stayed at Uber for five years. But what I found when I returned changed how I understood the future of work.
The research work was still happening. User studies were being conducted, insights were informing product decisions, teams were building with customer empathy. But researchers were not doing all of it anymore. Designers were running usability tests, product managers were synthesizing insights, engineers were talking directly to users. The roles were gone, but the work had consolidated into whoever could execute it fastest.
This pattern, which I thought was specific to Uber’s circumstances, was actually unfolding across the entire industry. The data confirm what I witnessed: roles were not eliminated because the work was unnecessary, but because it could be redistributed to people who could execute across multiple functions rather than defending a single specialized domain.
Why consolidation happens
When I returned to Uber, I kept asking: Why was my team hit harder than others? Were we not adding value?
The answer had nothing to do with value. It was about distribution and scale. The research function still mattered, but Uber realized it did not need as many dedicated researchers if designers could run basic usability tests, and if product managers could validate concepts with users. This was not “doing more with less.” This was consolidating functions into whoever could execute them fastest.
I saw the same pattern at Amplitude years later. The company went through what I would call a management purge, driven by a belief that product teams need doers rather than team builders and shapers. They decided it was better for their founder to lead directly, with one layer of management rather than layers of professional managers delegating work. This showed me that everyone should be building and figuring it out, not creating layers between ideas and execution.
The three patterns of consolidation
There are three ways consolidation happens in practice.
First, your work gets absorbed by adjacent roles. This happened to research at Uber when designers and PMs started conducting usability tests. My initial reaction was defensive. Then I realized that if we enabled this well through training programs and templates, we would free capacity for strategic work that only we could do. The designers running their own studies were not replacing us; they were extending our impact.
Second, other work gets absorbed into your role. This occurred to me at Amplitude when we began prototyping in Cursor and other tools, such as Bolt and Figma Make, to demonstrate to stakeholders how the features would feel. My team and I were not trying to become engineers; we were removing the handoff between design intent and stakeholder alignment. When we could show the CEO what Chart Chat would feel like, not just how it would look, decisions that used to take weeks happened in a single session.
Third, entire roles disappear. This is the concerning one. I have watched companies eliminate product management roles entirely and distribute the function to designers and engineers. I have seen research teams cut to zero with the expectation that designers will validate their own work. The harsh truth: if the function can be distributed and the role is not executed at a level others cannot match, the role disappears.
How I navigated consolidation
At Uber, I scaled research by enabling others to do research. We built training programs, created templates, and held office hours. The research team did not shrink because of this; instead, we increased in size because leadership recognized that research was becoming embedded in every product decision rather than serving as a bottleneck.
At Amplitude, I removed blockers by learning to prototype using AI. Design velocity increased, and our team’s influence grew. Every time I resisted consolidation, I lost ground. Every time I embraced it by learning new skills and enabling others, I gained leverage.
The approach I developed has four parts.
First, identify your biggest blocker. What slows you down because you are waiting for another function? For me at Amplitude, it was showing stakeholders what a design would feel like without waiting for engineering. For me at Uber, it was scaling insights without conducting every study myself. That blocker is your expansion opportunity.
Second, learn just enough to remove it. You do not need to become an expert in the adjacent discipline. You need to close the execution gap. I am not an engineer, but my team and I learned enough Cursor and other vibe coding tools to build interactive prototypes. Use AI to accelerate this; the technical barriers are collapsing.
Third, enable others to do your work. The more I taught designers to conduct research, the more valuable research became to the organization. The more I documented design processes, the more leadership understood design’s strategic value. Your expertise is amplified when you make it accessible.
Fourth, double down on what only you can do. At Uber after layoffs, designers ran usability tests while research increasingly focused on strategic discovery. At Amplitude and now at Evisort, when we could prototype our own designs, we focused on vision and alignment. Consolidation clarifies your unique value.
The Generalista position
Role consolidation is happening, and the data confirms what I saw firsthand at Uber and Amplitude is accelerating across the industry. McKinsey estimates that between 400 and 800 million workers globally could need to find new jobs by 2030 as automation and role consolidation reshape how organizations distribute work. But the people who expanded their capabilities thrived while those who tried to gatekeep struggled.
Generalistas do not resist consolidation. They lead it.
The specialists who defend titles will find themselves managing smaller territories. The Generalistas who expand capabilities will find themselves leading larger initiatives. Forces one and two created the conditions: democratization made expertise accessible, and AI made novices capable. Force three is the result: roles consolidate around whoever can execute fastest.
What to do this week
Identify your biggest blocker. What work are you waiting for someone else to do that you could learn yourself?
That is not just a frustration. That is your opportunity to expand before consolidation forces the expansion on you.
Start with one skill. Learn just enough to remove the dependency. Use AI to close the gap. Ship something this week that used to require a handoff.
The technical barriers are lower than ever. The question is whether you will expand voluntarily or wait for consolidation to force it.
What’s next
You have now seen all three forces. Force one showed how specialization gets democratized. Force two showed how AI empowers novices. Force three showed how roles consolidate around execution.
Next week, I will share the complete Generalista framework for navigating all three forces simultaneously and positioning yourself not just to survive but to thrive.
Until then: what is your biggest blocker right now? Reply and tell me. I read every response.
—John
---
**P.S.** The future does not belong to the deepest specialist or the broadest generalist. It belongs to people who can execute across the functions where you experience the most friction.
John — Excellent article. I learned a great deal about your background I didn’t know, although you shared much of it in our neighborly interactions.
I really liked your “description of work.” It left me thinking the AI had less to do with the job losses in the technology sector since the pandemic than a fuller realization of the Internet and its impact on work. My first job out of grad school (MS in Management) was as business manager for an R&D project between Carnegie Mellon University and IBM. A small team of about 20 computer scientists designed, build and deployed the first fully scalable network. As I understand it, it’s still in operation on campus and dorms and in one form or another around the world. The director of the program asked me my first day if I understood the Internet. I told him I didn’t, but I understood how to use it. He described it as “transportation innovation recursing on itself.” Huh? WTF!
He went onto explain that transportation technology was all about moving people and things faster and more efficiently from one place to another. That it began with “fixed networks” (train tracks) from fixed designations to other fixed designations. But when computing technology merged with communications technology, it became possible move experiences to people, instead of people to experiences. Of course, long before computers were networked, transportation had already succeeded “individualizing” transportation. Maximizing personal choices around “what”, “when” and “where” people wanted to travel. My own family migrated from the east coast in 1949 (before I was born) with my mother ant two older brothers taking the train to Sacramento and my father driving a relatively new car across country — only 80 years after the completion of the trans-continental railroad and before the completion of the national highway system.
Your description also caused me to reflect on the evolution of management. My studies went as far back as Peter the Great of Russia. He lived and ruled in the latter part of the 17th and early part of the 18th centuries. He is most known for modernizing Prussia, the Russian Navy and Russian Army. His success was due to being highly disciplined and extremely well organized. One of the realities had to deal with was his empire was comprised of many different cultures and languages. In fact, no common language. That didn’t stop him. He compartmentalized and routinized to great success. This necessarily required strict hierarchies. Sound familiar? I know from my own experience in high tech that designers, programmers, product marketers, etc. often think inside their own silos — their own cultures. Those techniques were used by those who brought us the Industrial Revolution. Think about how structured your work was until the Pandemic turned everything upside down.
I was stunned in 1988 when I joined Sun Microsystems and had the opportunity to move back to the Bay Area from Western Pennsylvania. I was part of a focused marketing and sales support group for the Higher Education Market. This group was one of many focused on specific market segments (e.g., Geographic Information Systems, Engineering, Finance, etc.). They only hired people who were experts in their particular market segment. But it didn’t matter how much knowledge or experience we had, we still had to bring “special deals” back to our managers for approval (the hierarchy). I designed a model that provided framework around the company’s target profit margin and how far each marketing segment manager could go with discounts and “special deals” to both grow the business AND as a “market segment” stay within the company’s profit targets. Wow, did that threaten the entrenched management team. I was invited to even present my model to the company CFO, Bill Riduchel (sp?), the CEOs former Harvard B School professor. He said he was impressed with the model and stated that this kind of analysis was not done anywhere in the company. The more important question was: “Why the F not?” I left the company shortly after that. What was suppose to be one of the leading companies in tech if not American business was so laden with 17th century thinking, it was clearly not for me.
My point is this, I think what you’re describing has less to do with AI and more to do with some external event (global pandemic) forcing changes in behavior that were already available to society, but due to our own habits we never really explored them until forced to. My small professional services firm was routinely engaging employees and clients via teleconferencing and email in the 1990s, well ahead of video conferencing. We learned that we could hire the best talent and NOT have to move them to where I lived and worked to have a successful company — 20 years before it became fashionable because of the pandemic.
I’m not denying the important impact of technology like AI on the today’s work. But I think it’s important to understand the true nature of these changes and not scapegoat an amazing technology like AI to hide the entrenched thinking of most of American business management.
I like the way you write and will continue to follow your ideas.