4 guests 6 episodes 2,503 words

The Mexican Standoff: When Every Role Thinks It Can Do the Others' Job

Are the PM, designer, and engineer roles converging into one?

Marc Andreessen calls it a "Mexican standoff." Every coder believes they can be a PM and designer because they have AI. Every PM thinks they can code and design. Every designer thinks they can be a PM and coder. All three roles are pointing guns at each other, and nobody is entirely wrong.

The most radical version of this idea: a single founder builds an entire company alone, with AI as their team. The question is whether this convergence is real, temporary, or something more nuanced than either extreme suggests. The answer matters urgently because it affects hiring decisions, career investments, and organizational design for every product company alive.

Are the traditional roles of product manager, designer, and software engineer collapsing into a single "builder" role? And if so, what should people in each role do about it right now?

The 3 Positions

1 / 3

Full Convergence Is Happening -- Everyone Becomes a Builder

Marc Andreessen (Co-founder · Andreessen Horowitz) · Boris Cherny (Head of Claude Code · Anthropic)

Advocates: Marc Andreessen (Co-founder, Andreessen Horowitz), Boris Cherny (Head of Claude Code, Anthropic)

As co-founder of the world's largest venture firm and inventor of the web browser, Marc Andreessen has a unique vantage point on how technology reshapes work -- he is both funding the AI companies enabling this convergence and advising the founders living it daily. He sees the convergence as already underway, driven by AI tools that collapse the skill gap between disciplines.

"There's like a Mexican standoff happening between those three roles. Every coder now believes they can also be a product manager and a designer because they have AI."
Marc Andreessen
▶ 00:00:54

The key insight is not just that roles are blurring, but that the combinatorial value of cross-disciplinary skill is superlinear.

"What happens is, the additive effect of being good at two things is more than double. The additive effect of being good at three things is more than triple. You become a super relevant specialist in the combination of the domains."
Marc Andreessen
▶ 00:00:54

This superlinear value is the economic engine driving convergence -- companies will increasingly seek people who combine PM judgment, engineering execution, and design sensibility because the value of that combination exceeds the sum of three separate hires. Andreessen grounds this in the broader AI moment, arguing that coding is now "virtually solved" -- even Linus Torvalds acknowledged over the 2025 holiday break that AI codes better than the world's best programmers.

"Everybody wants to talk about job loss, but really, what you want to look at is task loss. The job persists longer than the individual tasks."
Marc Andreessen
▶ 00:00:43

This reframing matters: while jobs persist, individual tasks within them are being automated or democratized, meaning the boundaries between job descriptions become porous. The most leading-edge founders Andreessen works with are already testing the extreme version of this thesis.

"The most leading edge founders are thinking of, can you have entire companies where the founder does everything?"
Marc Andreessen
▶ 00:00:34

Boris Cherny, as the person leading Claude Code -- the AI coding tool that has doubled its daily active users monthly -- has direct, real-time data on how AI is dissolving the boundary between technical and non-technical roles. He pushes further than Andreessen with a concrete prediction.

"I think by the end of the year everyone is going to be a product manager, and everyone codes. The title software engineer is going to start to go away. It's just going to be replaced by builder."
Boris Cherny
▶ 00:00:44

"In a year or two, it's not going to matter. Coding is virtually solved. I imagine a world where everyone is able to program, anyone can just build software any time."
Boris Cherny
▶ 00:00:22

Cherny observes that the strongest engineers have always been hybrid practitioners -- "product engineers with really great design sense" -- and that AI accelerates what was already true about the best performers. He draws a historical parallel to the printing press: when reading and writing were democratized beyond a small elite, the Renaissance happened. Similarly, when coding becomes accessible to everyone, a new creative explosion becomes possible.

"I have never enjoyed coding as much as I do today, because I don't have to deal with all the minutia. Productivity per engineer has increased 200%."
Boris Cherny
▶ 00:00:13

Anthropic's own Claude Code team is still hiring engineers -- Jevons paradox in action, where more AI capability creates more demand for skilled builders, not less. Claude is now "coming up with ideas, looking at bug reports, looking at telemetry for bug fixes" -- acting more like a coworker than a tool.

When This Applies

Early-stage startups, prototyping, side projects, internal tools -- where speed matters more than polish and one person can hold the full context. Companies like Mercor and the solo founders in a16z's portfolio are already operating this way.

Crossover Advantage -- The Hybrid Professional Wins

Howie Liu (CEO · Airtable)

Advocates: Howie Liu (CEO, Airtable)

As a 13-year founder-CEO who restructured Airtable's entire organization for AI and personally returned to IC-style building, Howie Liu has lived the convergence thesis at both the individual and organizational level. He is more measured than the full convergence camp -- he does not predict that roles disappear, but that the advantage goes to anyone who can cross the boundaries.

"There's a strong advantage to any of those three roles who can kind of cross over into the other two. As a PM, you need to start looking more like a hybrid PM prototyper, who has some good design sensibilities."
Howie Liu
▶ 00:00:33

This is not "everyone does everything." It is "the best people in each role can do 30% of the adjacent roles, and that 30% makes them dramatically more effective." A PM who can prototype in code gets feedback faster. An engineer who understands design ships better UX. A designer who understands technical constraints designs buildable interfaces.

Liu's reasoning goes deeper than just skill acquisition. He argues that the nature of product work has fundamentally changed -- product value is now often in the interaction, not the visual design. He uses ChatGPT as an example: it has perhaps the most basic design imaginable, yet the "real design is happening underneath the hood in how it responds to different queries." This means designers who cannot understand models and tool-calling are designing blind, and PMs who cannot prototype are specifying abstractions.

"It really does become more about individual attitude and maybe some polymathism."
Howie Liu
▶ 00:51:07

Liu himself exemplifies the "IC CEO" trend -- founders and CEOs returning to hands-on building, personally writing code and leading product initiatives. He encourages his team to cancel all meetings for a day or week to explore AI tools, treating curiosity as a core job requirement. His restructuring of Airtable into fast-thinking (rapid AI experimentation) and slow-thinking (foundational platform work) groups reflects his belief that execution now looks more like an experimentation playground than deterministic roadmap planning.

"Everyone can learn how to be a versatile kind of unicorn product engineer/designer hybrid in the AI-native era. The only thing stopping you is just going out and doing it."
Howie Liu
▶ 00:51:07

The job market reflects the crossover trend. "Product engineer" roles -- a term that barely existed five years ago -- now show 415 open positions according to recent job market data. The market is pricing in the convergence.

When This Applies

Growth-stage companies like Airtable where team sizes are 3-8 people and the overhead of handoffs between PM/design/engineering is the bottleneck. The 415 open "product engineer" roles signal that the market is pricing in this model.

Non-Technical People Can Now Build -- Convergence From the PM Side

Zevi Arnovitz (PM · Meta)

Advocates: Zevi Arnovitz (PM, Meta)

As a non-technical PM with zero coding background who now builds significant products using AI tools -- and whose Meta engineers ask him to teach them his workflow -- Zevi Arnovitz is living proof that the PM-to-builder convergence is real and accessible. He studied music in high school, was not in a tech unit in the Israeli army, and had zero technical background before AI tools arrived.

"I have zero technical background, did music in high school ... when Sonnet 3.5 came out. I remember watching a YouTube video building apps using Bolt or Lovable. It basically felt like someone came up to me and said, 'You have superpowers now.'"
Zevi Arnovitz
▶ 00:00:06

Within days of returning from a trip to Japan, he opened his laptop and started building. What makes his approach notable is the sophisticated workflow he developed to compensate for his lack of technical knowledge. He created a "CTO" persona in his AI tools -- a complete technical owner of the project with instructions to challenge him and not be a people pleaser. He recognized early that AI coding agents are "super eager to write code" but that planning is essential for complex implementations, especially around databases and payments.

His code review process is particularly clever -- since he cannot catch mistakes himself, he runs "/review" to have Claude review its own code, then has Codex review Cursor's output and vice versa, creating a multi-model verification system. His engineers at Meta actually ask him to teach them his workflow, inverting the traditional mentor-mentee dynamic between PMs and engineers.

"In the next coming years, I think everyone's going to become a builder. Titles are going to collapse and responsibilities are going to collapse."
Zevi Arnovitz
▶ 00:00:22

"It's the best time to be a junior, contrary to what a lot of people are saying. When else in history could you get out of school and just build a startup on your own?"
Zevi Arnovitz
▶ 00:00:59

Consider what this means for the typical product review meeting. Previously, a PM would present a specification, an engineer would estimate feasibility, and a designer would show mockups. Now, a PM can walk in with a working prototype built over a weekend. The conversation shifts from "here is what I think we should build" to "here is what I built -- let's evaluate it." The power dynamics of the product triad change fundamentally when any member can produce tangible output, not just artifacts about future output.

When This Applies

Large-scale systems at companies like Stripe, Meta, and Anthropic, or any domain where depth of expertise (security, performance, accessibility, brand) cannot be replicated by AI tools.

Evidence from the Archive

His journey from seeing a YouTube video about Bolt/Lovable in Japan to becoming a proficient builder within a year

Arnovitz built a complete multi-model code review workflow: /review in Claude, then cross-reviewing with Codex and Cursor checking each other

Zevi Arnovitz · PM ▶ Watch

Andreessen Horowitz

Leading-edge founders in the a16z portfolio are testing whether a single founder can run an entire company with AI

Linus Torvalds publicly acknowledged that AI codes better than the world's best human programmers over the 2025 holiday break

Marc Andreessen · Co-founder

Anthropic

Claude Code at Anthropic has seen daily active users double monthly, demonstrating massive demand for...

Claude Code at Anthropic has seen daily active users double monthly, demonstrating massive demand for builder-accessible coding

Boris Cherny · Head of Claude Code

Airtable

Airtable restructured into fast-thinking and slow-thinking groups to accelerate AI investment

Howie Liu himself became an IC CEO, personally getting into code and leading product initiatives

Howie Liu · CEO ▶ 00:00:33

The Synthesis

The convergence is real but asymmetric. It is much easier for a PM to learn to build prototypes with AI than it is for an engineer to develop deep user empathy, or for a designer to develop systems thinking. The tools are collapsing the technical barriers, not the judgment barriers.

Asymmetric Convergence

Is the convergence of PM, design, and engineering roles symmetric?

Mechanical vs. Judgment

Which parts of each role are converging and which are diverging?

Organizational Pacing

How fast should product organizations embrace convergence?

20% Adjacent Investment

What is the career advice for individuals in the convergence era?

The convergence is real but asymmetric. It is much easier for a PM to learn to build prototypes with AI than for an engineer to develop deep user empathy, or for a designer to develop systems thinking. Technical barriers are collapsing, but judgment barriers are not.

The mechanical parts of each role are converging. The judgment parts are diverging in value. Engineers whose value is writing clean code face commoditization. Engineers whose value is systems architecture become more important. The pattern is consistent across all three disciplines.

Small teams building new products should lean toward convergence; large teams maintaining complex systems should preserve specialization while encouraging crossover. The organizational structure should follow the product architecture, not an abstract theory about what roles should exist.

Whatever your primary role, invest 20% of your learning time in an adjacent discipline. PMs should build with AI tools this quarter. Engineers should spend time with customers. Designers should learn basic queries. The convergence rewards generalists who can go deep when needed.

Which Approach Fits You?

Answer 3 questions about your situation. We'll match you to the right approach.

Question 1

How large is your product team?

Question 2

What is slowing your team down most?

Question 3

How are AI tools changing your team's workflow?

Notable Absences

The Bottom Line

The practical question for product leaders is not whether to embrace convergence but how fast. If you reorganize around "builders" tomorrow, you lose the deep expertise that specialists provide. If you cling to strict role separation, you lose the speed and flexibility that crossover skills enable. The answer, as usual, is conditional: small teams building new products should lean toward convergence; large teams maintaining complex systems should preserve specialization while encouraging crossover.

One more dimension worth noting: at Figma, removing the "black box" between design, engineering, and PM processes has created better outcomes long before AI entered the picture. The convergence is not purely an AI phenomenon -- it is an acceleration of a trend that high-performing teams have been pursuing for years. AI is the catalyst, but the underlying principle -- that tighter feedback loops between disciplines produce better products -- has always been true.

Sources

Howie Liu — "How we restructured Airtable’s entire org for AI | Howie Liu (co-founder and CEO)" — Lenny's Podcast, August 31, 2025
Zevi Arnovitz — "The non-technical PM’s guide to building with Cursor | Zevi Arnovitz (Meta)" — Lenny's Podcast, January 18, 2026