The Full Stack Builder: The End of the Design Process as We Know It
The double diamond is a liability. Engineers ship faster than designers can explore. The PM role is dissolving and the three profiles that will survive this era look nothing like who we've been hiring
In the early 2000s, we called them webmasters. One person who designed the layout, wrote the HTML and CSS, figured out the JavaScript, managed the server, and sometimes even wrote the copy. There was no “handoff” because there was no one to hand off to. You owned the whole thing - from Photoshop (or Fireworks) slice to FTP upload. It wasn’t glamorous. It wasn’t specialized. But it was end-to-end, and it worked.
Then the industry professionalized. We split into UX designers and UI designers and front-end developers and back-end developers and product managers and project managers and QA engineers. Every year, the org chart got more granular, the handoffs got more elaborate, and the distance between “having an idea” and “shipping it” got longer. We traded speed and ownership for scale and specialization. For a while, that was the right trade.
I even remembered a short moment in time where I used to remove photos from the templates I was designing, desaturate them and ship them as “Wireframes”. It felt weird.
I never fully made it. Even after founding Alma Creativa in 2011 and eventually joining Accenture, I kept one hand in design and the other in code - motion, web, interaction, strategy. People called it being a multipotentialite, a generalist, sometimes as a compliment, sometimes not. But what I was actually doing was refusing to let go of the thing that made the webmaster powerful: the ability to think about a problem and build the solution in the same breath.
Now, twenty years later, LinkedIn’s CPO is telling us that the future belongs to exactly that kind of person. He just gave it a better name.
Tomer Cohen, Chief Product Officer at LinkedIn, recently shared something that should make every product leader, designer, and engineer sit up straight: LinkedIn has eliminated its Associate Product Manager program. In its place, they’ve created a new role - the Full Stack Builder - with a formal title, a dedicated career ladder, and a philosophy that fundamentally reimagines how products get made.
This isn’t a rebrand. It’s a structural bet on a future where AI collapses the distance between having an idea and shipping it. And it carries profound implications for how we organize talent, design teams, and think about creative work - especially at big enterprises, where the intersection of strategy, design, and technology is our daily bread.
Meanwhile, Jenny Arden - former design director at Anthropic, who voluntarily stepped back from managing 12–15 reports to become an individual contributor during one of the most turbulent periods in product design - is living the consequences of this shift in real time. Her perspective from the trenches adds a crucial layer to Cohen’s structural argument: the design process itself is fracturing, and the designers who thrive will look nothing like the ones we’ve been hiring for the past decade.
The 70% Problem
Let’s start with the number that frames everything. According to LinkedIn’s own Economic Graph data, 70% of the skills required for most jobs will change by 2030. Not evolve. Change. That’s not a slow drift; it’s a tectonic shift.
And yet, most organizations are still structured around assumptions from the pre-AI era: rigid functional silos, sequential handoffs from PM to Design to Engineering to QA, teams of fifteen people spending six months to ship a feature that a motivated individual with the right AI tools could prototype in a week.
Cohen calls this “organizational bloat,” and he’s not wrong. The traditional stage-gate process - where every discipline operates in its lane, throwing deliverables over the wall to the next group - was designed for a world where each step required deep, irreplaceable specialization. AI is dissolving that premise faster than most org charts can adapt.
So What Exactly Is a Full Stack Builder?
The definition is deceptively simple: a Full Stack Builder is someone who can take an idea from concept to launch, regardless of their starting discipline. They combine three capabilities:
Coding - not necessarily at a senior engineer’s depth, but enough to prototype rapidly with AI coding agents, ship functional software, and speak the language of implementation.
Design - UX thinking, visual craft, the ability to translate user needs into interfaces that work. Not pixel-pushing, but experience architecture.
Product - the strategic layer. Vision, prioritization, go-to-market thinking, and the judgment to know what’s worth building in the first place.
The magic isn’t in mastering all three at expert level. It’s in having enough fluency across all three that AI can fill the gaps. A designer who can code well enough to prompt an AI coding agent effectively. An engineer who understands enough about user research to validate an idea before building it. A product thinker who can sketch a prototype instead of writing a 40-page requirements document.
Engineer ≠ Builder
This distinction matters, and it’s where most people get confused. A Full Stack Engineer masters front-end and back-end development. They’re focused on technical execution - React, APIs, databases, CI/CD. Their core question is: “How do I build it?”
A Full Stack Builder includes all of that, but adds three layers that transform the role from executor to owner:
Research & Discovery - talking to users, validating hypotheses, generating insights that inform what gets built.
Design & Experience - prototyping, UX architecture, visual craft, brand coherence.
Product Strategy - vision, roadmap, go-to-market, impact measurement.
The Builder’s core question is fundamentally different: “What is worth building?” This isn’t an upgrade in technical skill. It’s a shift in ownership - from problem-solving to problem-finding, from execution to end-to-end accountability.
The Design Process Is Breaking
Here’s something that doesn’t get said enough: the classic design process - discover, diverge, converge, deliver - is breaking down. Not because it was wrong, but because the tempo of building has changed underneath it.
When an engineer can spin up seven coding agents and ship a working version before a designer finishes exploring options in a diverge phase, the carefully sequenced double diamond becomes a liability. You’re still mapping the problem space while someone else has already solved it - imperfectly, maybe, but functionally. And in a world where iteration is cheap, a functional imperfect thing beats a perfect plan every time.
Jenny Arden sees this clearly. She describes design work splitting into two distinct modes:
The first is supporting execution: consulting with engineers as they build, giving feedback in real time, polishing directly in code rather than producing handoff specs that are obsolete before the Figma file is closed.
The second mode is setting short-range vision - but scoped to three to six months, not the multi-year roadmaps designers used to traffic in. This vision work is still critical, maybe more critical than ever, because when everyone can build anything fast, someone needs to point the team in a coherent direction. Speed without direction is just expensive chaos.
This doesn’t mean spatial design tools are dead. Figma remains essential, but for different reasons than before. It’s still the best environment for rapidly exploring eight to ten different design directions on a canvas - something that coding tools handle poorly because they’re too linear and create investment bias toward whichever direction you started building first. For micro-level visual and interaction decisions, spatial exploration still beats sequential iteration. But the output of that exploration increasingly feeds directly into code, not into a spec document.
Human + Machine: The Division of Labor
Cohen presents a framework that I find genuinely clarifying. Picture a timeline from Insight to Launch, with eight phases: Insight, Research, Solution, Roadmap, Design, Code, Test, Launch. Now draw two axes: Human contribution above the line, Machine contribution below.
The pattern that emerges is counterintuitive. The early phases - Insight, Research, Solution - are heavily human. This is where Vision, Empathy, and Communication dominate. The middle phases - Roadmap and Design - are where Creativity peaks. And the later phases - Code, Test, Launch - are increasingly automated, though Judgment remains essential throughout.
The five skills that AI cannot automate, according to Cohen, are Vision, Empathy, Communication, Creativity, and Judgment. As he puts it: “I’m working hard to automate everything else.”
But What About Taste?
There’s a tempting narrative in the design world right now that “taste” is the ultimate moat - the one thing AI will never replicate. Jenny Arden pushes back on this, and I think she’s right to. AI will likely get better at taste and judgment over time. The models are already producing work that passes the taste test for many applications. Designers who are clinging to taste as their sole differentiator may be holding onto a shrinking island.
“I have no technical ability. I know nothing about music. I can’t play an instrument. I can’t operate a board. All I know is what I like and what I don’t like, and I’m decisive about it.”
— Rick Rubin
But here’s the nuance: someone still has to be accountable for what ships. The same way an engineer is accountable for AI-generated code - reviewing it, testing it, owning the consequences - a designer or builder needs to be accountable for AI-generated design decisions. Taste may become less rare, but accountability never will. The builder’s real moat isn’t having good taste. It’s having the judgment to know when the AI’s taste is wrong and the authority to override it.
This maps onto the three macro-phases I use in my own work: Product (Research), Design, and Build. The human contribution is heaviest in the first two; AI’s contribution grows as you move toward implementation. But the builder’s judgment - the ability to evaluate, redirect, and curate - acts as a throughline across the entire cycle.
The Agent Layer
Where this gets operationally interesting is in the agent architecture. Along that same Insight-to-Launch timeline, LinkedIn has deployed specialized AI agents:
Research Agent - synthesizes data, surfaces competitor insights, analyzes market trends.
Growth Agent - identifies opportunities, sizes markets, generates business cases.
Trust Agent - validates solutions, finds vulnerabilities, stress-tests assumptions.
Design Agent - generates prototypes, explores UI patterns, maintains visual systems.
Coding Agents - write implementation code, handle boilerplate, accelerate development.
QA Agent - runs tests, catches regressions, validates quality.
Maintenance Agent - monitors post-launch, flags issues, suggests improvements.
The critical insight: the builder orchestrates the agents. The agents execute. Judgment stays human. This is not about replacing people with AI. It’s about compressing the team needed to ship something meaningful from fifteen people to three - or even one.
Chat as the Connective Tissue
One thing worth noting about the interface layer: chat as an interaction paradigm isn’t going away. Despite early expectations that chatbots were a temporary stop on the way to richer, more structured UIs, the reality is that chat offers infinite flexibility. It’s the universal fallback for any task that doesn’t yet have a dedicated interface.
But the future isn’t chat-only. It’s hybrid. Models are increasingly generating UI elements on the fly for specific tasks - the interactive widgets that Claude recently shipped are a good example - while chat remains the connective tissue between them. For the Full Stack Builder, this means designing not just screens but conversation architectures: systems where the AI knows when to talk and when to show, when to ask and when to act.
What LinkedIn Actually Did
The transformation at LinkedIn wasn’t theoretical. It happened in four concrete steps.
First, they eliminated the APM program entirely. The traditional Associate Product Manager track - a prestige pipeline at most big tech companies - was retired. Second, they launched the Associate Product Builder program, which teaches coding, design, and product management together from day one. Third, they created a formal “Full Stack Builder” title with a dedicated career ladder. And fourth, they integrated AI proficiency into their performance review criteria.
That last point is perhaps the most consequential. When AI fluency becomes a performance metric - not an optional nice-to-have but a criterion by which you’re evaluated - adoption accelerates dramatically.
The Counterintuitive Truth About AI Adoption
Here’s the finding that surprised Cohen’s team the most. The expectation was that AI would be a great equalizer: juniors would benefit most because AI would close their skill gaps, while seniors would resist the change.
The reality was the opposite. Top performers adopted AI fastest and derived the most value from it. Why? Because they had the judgment and experience to know what to ask for, how to evaluate the output, and where to apply it for maximum leverage. A senior designer doesn’t just use an AI design agent - they direct it, critique it, and iterate with it in ways that produce fundamentally better outcomes.
The implication for organizations is clear: invest in your top performers as AI champions. Don’t democratize mediocre AI training across the entire org. Find the people who are already great and give them the tools to become extraordinary. Their success stories will pull the rest of the organization forward faster than any mandatory training program.
Four Levers for Change (Plus One)
Cohen’s playbook for driving adoption is refreshingly practical. It’s not about technology - it’s about culture.
Celebrate wins. Make early adopter successes visible. Internal storytelling - showing how a builder shipped something remarkable using AI in a fraction of the usual time - is the most powerful adoption engine. People don’t change because you tell them to. They change because they see someone they respect doing something they want to be able to do.
Create exclusivity. Counterintuitively, making AI tools available as a privilege rather than mandating their use generates more interest. The FOMO effect is real. When people hear that the best performers have access to something special, they want in.
Update performance reviews. What gets measured gets improved. When AI proficiency shows up in how people are evaluated, it moves from “interesting” to “essential.”
Train by doing. Not theoretical AI courses, but real pilots on real projects with real stakes. LinkedIn’s APB program doesn’t teach coding, design, and PM as abstract subjects - participants build and ship actual products.
Build trust through speed, not perfection. This one comes from Anthropic’s playbook rather than LinkedIn’s, but it’s equally important. Anthropic ships products early, labels them research previews, and iterates publicly based on real feedback. The conventional wisdom is that shipping something rough damages your brand. Jenny Arden argues the opposite: what actually degrades trust isn’t launching something rough - it’s launching something rough and then going silent. If you ship fast, respond to feedback visibly, and keep improving, users will trust you more, not less. This is a profound shift for organizations that have been trained to equate quality with polish rather than with responsiveness.
Three Designers Who Will Thrive
If the Full Stack Builder model reshapes how products get made, it also reshapes who gets hired to make them. I see three designer profiles that are uniquely suited to this moment - and none of them look like the “senior UX designer with 8+ years of experience” that dominates most job descriptions today.
The Block-Shaped Generalist
Not just broad - consistently strong across multiple domains. Think 80th percentile in research, visual design, interaction, prototyping, and front-end. They’re not the best at any one thing, but they’re dangerous at everything. In a world where AI handles the bottom 60% of execution across every discipline, the person who can operate at 80% across five disciplines and orchestrate AI to cover the remaining 20% is extraordinarily valuable. This profile is rare, and it’s the natural home of the Full Stack Builder.
The Long T-Shaped Specialist
Deep craft in a core area - top 10%, the kind of designer where the bar is simply different. The master typographer, the motion designer who makes you feel something in 0.8 seconds, the researcher who sees patterns no one else catches. AI makes generalists more capable, but it also makes the gap between “good enough” and “exceptional” more visible. When everyone can produce competent work with AI, the specialist who produces transcendent work becomes more valuable, not less. Also rare. Also hard to hire.
The Cracked New Grad
This is the most overlooked profile, and Jenny Arden makes a compelling case for it. Most companies are hiring experienced designers with deep portfolios. But early-career people with blank slates, fast learning curves, and no attachment to legacy processes may be uniquely suited to this moment. They don’t carry baked-in rituals that are now obsolete. They’ve never known a world where you spend three weeks in a discovery phase before opening a design tool. Their lack of expectations isn’t a weakness - it’s a superpower. They’re curious, ambitious, adaptive, and comfortable with constant change. In a field that has transformed so radically that years of experience alone are no longer a reliable proxy for impact, the new grad who can learn fast and ship faster may outperform the seasoned veteran who’s still trying to run the old playbook.
Jenny herself made a version of this bet: she stepped down from a design director role managing 12–15 reports to go back to being an IC. She questioned whether middle management had a safe future in a world of compressed teams and AI-augmented builders, and she wanted hands-on time during a period of rapid change. The IC work is giving her hard skills she wouldn’t have developed while managing - and a front-row seat to how the craft itself is evolving. That takes intellectual honesty and courage. It’s also a signal of where the value is shifting.
The Full Stack Creative
Everything I’ve described so far comes from Cohen’s perspective on product and engineering, filtered through Arden’s perspective on design. But the Full Stack Builder model has a natural mirror in the creative world - and it’s one I think is even more urgent for companies.
I call it the Full Stack Creative: someone who owns the entire chain from insight to concept to content to distribution, with AI agents as their production team.
Think about how creative work has traditionally been organized. You have a copywriter or an art director (rarely both). A social media manager separate from a brand strategist. Production outsourced to agencies or freelancers. And a process that moves from brief to concept to production in a slow, sequential chain.
Today, we’re somewhere in the middle. T-shaped creatives who can handle copy and visual and some strategy. AI to accelerate concept development and drafting. But the pipeline is still the pipeline. AI is a tool, not yet a workflow.
The Full Stack Creative is what comes next. It’s someone who thinks in terms of campaigns, not assets. Who moves from audience insight to concept to finished content to distribution and measurement - orchestrating AI agents for copy, visual generation, video, and analytics along the way. Their deliverable isn’t a piece of content. It’s measurable impact.
The Six Competencies
The Full Stack Creative’s skill map looks like this:
Research & Insight - audience analysis, trend scouting, data-informed briefs. AI synthesizes social listening and market reports so the creative can focus on pattern recognition and strategic interpretation.
Copy & Narrative - brand voice, storytelling, microcopy. AI serves as a co-writer, enabling one person to produce volume without sacrificing quality or consistency.
Visual & Motion - art direction, image generation, video editing. From Midjourney to Runway, the creative doesn’t execute - they direct. The AI agents are the production team.
Strategy & Planning - content calendar, channel strategy, media mix. The Full Stack Creative thinks campaign-first, not asset-first.
Performance & Analytics - measure, iterate, optimize. AI-powered dashboards make it possible to understand what’s working and why in near real-time.
AI Orchestration - prompt engineering, agentic workflows, tool chaining. This is the meta-skill that amplifies everything else. The ability to compose and direct multiple AI agents into a coherent creative workflow.
I want to be clear about what this doesn’t mean. The Full Stack Creative doesn’t replace specialists. It makes them rarer and more valuable where deep expertise genuinely matters. You still need a master typographer for a luxury brand identity. You still need a cinematographer for a hero film. But for the vast middle ground of creative production - the campaign content, the social assets, the performance marketing - a Full Stack Creative with the right AI tools can deliver what used to require a team of eight.
What This Means for Companies
The implications for how you organize, hire, and serve clients are significant. You need to move from rigid functional silos to cross-functional Full Stack Builders. From sequential PM-to-Design-to-Dev-to-QA handoffs to parallel idea-to-prototype-to-launch cycles. From large teams and long cycles to lean pods and rapid iteration. From AI as an optional add-on to AI as foundational infrastructure. From static vertical skills to T-shaped competencies with AI fluency as a baseline.
And we need to rethink who we hire. The three profiles that will define the next era of design - the block-shaped generalist, the deep specialist, and the fearless new grad - don’t map neatly onto traditional job descriptions or leveling frameworks. You need to create space for all three, with career paths that reward breadth as much as depth and speed as much as seniority.
This isn’t a change of tools. It’s a change of operating model.
Start Now, Not Tomorrow
The window for experimentation is open, but it won’t stay open forever. Every month that passes, the gap between organizations that have embraced the Full Stack Builder model and those still debating it grows wider.
Six things you can do immediately in your company:
Launch a Full Stack Builder pilot on a real project with a short feedback cycle.
Build an internal AI platform customized to our enterprise context - off-the-shelf tools aren’t enough.
Identify top performers as AI champions and early adopters.
Integrate AI proficiency into performance evaluation criteria.
Pilot the Full Stack Creative model on an end-to-end Song campaign.
Hire at least one “cracked new grad” per studio and give them a real project, not an internship.
Cohen’s insight that top performers adopt AI fastest should guide our sequencing. Arden’s insight that the design process itself is fracturing should guide our humility. Don’t try to train everyone at once. Don’t try to preserve processes that were built for a different era. Find the builders who are already great, give them permission, tools, and visibility. Their wins will create the gravitational pull that moves the entire organization.
The future of work isn’t about AI replacing people. It’s about AI enabling a different kind of person - one who can see the whole picture, orchestrate the right tools, and ship things that matter. The Full Stack Builder is that person. The question is whether we’ll build the culture and infrastructure to let them thrive.
How do you split between delieverance and experimentation with new workflows as a consequense of new products being released weekly?
You have the ability to destillate chaos into action, great read as always.