Tobi Lutke’s Shopify internal memo
Smaller and More Efficient Teams
When should we hire a person versus delegating to AI? Recently I’ve been more reluctant towards hiring people as an attempt to build mid sized projects. Yes, the codebases would get pretty big, and there’s also tasks involved that I wouldn’t say are my forte exactly. Yet, when I think about the meetings I have to sit through communicating what I want to build and just time spent doing filler work, I get more and more inclined towards just doing it myself. It’s not to say that teamwork isn’t good work, some of the most creative product ideas I’ve worked on stemmed from chats, during lunch breaks, exploring tangents, with engineers, journalists. The value of connecting the dots during these conversations is something that is difficult to replace. However, is the assumption that delegating work means higher productivity still valid? After all, the cost of execution is continually decreasing (as long as we have a clear idea of what to build).
Compared to late 2022, I’ve shifted more energy toward building and testing ideas directly, rather than hiring large teams to delegate to. Early-stage startups now need fewer product managers, especially when founders can build and validate ideas over a weekend. The marginal productivity increase associated with an extra headcount might now be lower than that of employing AI agent(s). In practice, this means smaller, more agile teams with outstanding capabilities from ICs. I believe the future belongs to ultra-small teams, for startups, this means any where from 2-5 people, with each member handling a job function while employing multiple AI agents to execute tasks. The benefits are clear: less time spent in meetings, quicker iterations, and the positive ripple effects of tight-knit teams. Rather than building large departments with specialized roles, companies can now assemble small, versatile teams augmented by AI capabilities. These teams can move faster and with greater autonomy than traditional corporate structures allow.
AI Tools Enabling Rapid Prototyping
Recent AI coding tools like Cursor have expedited the prototyping process, allowing developers (and non-technical people too) to quickly build functional MVPs in record time. These tools excel at generating boilerplate code, implementing common patterns, and even troubleshooting errors. A single engineer with Cursor can accomplish what previously required multiple developers working in tandem, or achieve in around one fifth the time they’d spend working on it alone.
Collaboration workflows enabled by Figma, Credit to Kevin Kwok
This acceleration in prototyping parallels what we saw with Figma in design. As Kevin Kwok noted, “Tightening the feedback loop of collaboration allows for non-linear returns on the process” (Kwok, 2020). Just as Figma collapsed the barriers between designers and their collaborators, AI coding tools are now breaking down the technical barriers that previously separated product visionaries from implementation.
While much of the attention has been on AI helping engineers be more efficient, I’m more interested in how its ripple effect on tangential roles - such as product managers and ui/ux designers. Just like how Figma allowed faster iteration cycles between designers and engineers/product managers, would this reduction in prototype testing allow for more innovative workflows where product people (with knowledge of engineering possibilities) could quickly test ideas and engineers focus more on the implementation of production scale systems?
This dynamic raises an interesting question for founders and engineering leaders: when does it make sense to hire additional team members versus investing in improving the AI capabilities of existing team members? The calculus now involves comparing the marginal cost of onboarding and training a new hire against the potential productivity gains from enhancing your current team’s AI workflow mastery.
The Rise of Product Engineers
Software Engineers should become “Product Engineers” as LLMs have commoditized routine coding tasks. What’s truly valuable now are generalists who can code but also have an eye for UI/UX design, good product taste, and deep market understanding. Startups should increasingly seek engineers who talk directly with users, make decisions on what to build, and then build with AI assistance. This approach works because engineers inherently understand both technical constraints and opportunities better than anyone else. The traditional roles of “product manager” and “engineer” are merging into a hybrid that combines technical expertise with product sensibility.
“Design is all of the conversations between designers and PMs about what to build” (Kwok, 2020). Similarly, product development is now larger than just engineers or product managers—it encompasses the entire collaborative process of identifying, designing, and implementing solutions. Too often, people view product management as just a career path rather than a mindset. This limited perspective overlooks the deep curiosity and problem-solving drive that true product managers embody—a drive that goes beyond mere job titles and organizational structures. The best product engineers embody this product mindset, focusing on solving real problems rather than just building features. Super ICs don’t wait for dev support—they build what they need when they need it, combining technical skills with product thinking to deliver complete solutions.
Shifting Emphasis: From Implementation to Idea Generation
As AI handles more of the routine implementation work, the traits we look for in product engineers should also evolve. The most valuable skills now center around workflow generation and new idea creation rather than implementation details. The ability to conceptualize solutions, design effective workflows, and identify the right problems to solve becomes paramount.
Product engineers who excel in this new environment should demonstrate:
- Strong systems thinking across the entire product lifecycle
- The ability to rapidly iterate and test hypotheses
- Comfort with ambiguity and exploration over rigid planning
- Skill in designing human-AI collaboration workflows
- An understanding of when to leverage AI and when to apply human judgment
This shift means that strong ICs with AI fluency can now accomplish what previously required teams of specialists. For early-stage startups, this dramatically changes the calculus around hiring, especially for traditional product management roles.
Shopify’s AI-first Approach
Shopify’s CEO Tobi Lütke’s internal memo illustrates this shift perfectly. He writes:
“We are entering a time where more merchants and entrepreneurs could be created than any other in history… Having AI alongside the journey and increasingly doing not just the consultation, but also doing the work for our merchants is a mindblowing step function change here.”
Lütke emphasizes that “reflexive AI usage is now a baseline expectation at Shopify.” He notes that using AI well is a skill that needs to be learned through frequent use, and that AI acts as a multiplier for already high-performing individuals. He compares Shopify to a “red queen race” from Alice in Wonderland—you must keep running just to stay still. In a company growing 20-40% year over year, everyone must improve at that rate just to re-qualify for their position. With AI tools, this previously daunting expectation now seems achievable.
The Critical Role of Taste in Product Development
As AI handles more of the execution, the differentiating factor for product engineers becomes “taste” - the ability to discern what makes for excellent product design, user experience, and strategic direction. Product taste is what separates adequate solutions from exceptional ones. In a world where AI can generate competent designs and functional code, the human with superior taste becomes indispensable. “Just like how the constraints on design at companies is often not a problem of pixels, but of people” (Kwok, 2020). As technical constraints dissolve through AI assistance, the human factors around judgment, taste, and strategic direction become the primary differentiators.
Taste involves the ability to:
- Identify which problems are worth solving
- Determine the appropriate level of complexity for solutions
- Recognize when simplicity serves users better than feature-richness
- Balance aesthetic appeal with functional needs
- Anticipate user needs before they’re explicitly requested
These skills can’t be easily replicated by AI systems. While AI can execute with increasing competence, it still lacks the intuitive understanding of human needs and experiences that informs good taste.
What This Means for Organizations
Shopify’s approach includes several key principles organizations should consider:
- AI as a fundamental expectation - Not an optional tool but a core competency
- AI integration from the prototype phase - Using AI throughout the development process
- Headcount requests must demonstrate why AI can’t do the job - Teams must explore AI solutions first
- Universal application - This applies to all levels, including executives
Organizations following these principles will likely outperform those treating AI as merely an optional tool. The principles represent a fundamental rethinking of how work gets done, not just a technological upgrade.
While some companies are hiring for AI product managers nowadays, this might be a transitionary position, just as prompting is an intermediary step in human-LLM interaction—similar to the arrival of GUIs before the personal computer age became mainstream. Product managers now need to learn when to dive into detailed work and when to step back for strategic oversight, a balance that will continue to evolve as AI capabilities expand.
The Path Forward
The future belongs to smaller, more nimble teams of highly capable individuals who leverage AI to achieve what previously required entire departments. This might be the new normal of work. For ICs, learning to effectively collaborate with AI tools is no longer optional—it’s essential for remaining competitive. For organizations, the challenge is creating environments where these super ICs can thrive, with processes and cultures that maximize the human+AI partnership rather than treating them as separate domains. Most importantly, this shift will enable more creative exploration and novel problem-solving, as routine tasks become increasingly automated. The most exciting products and services of the coming years will likely emerge from these small, AI-augmented teams combining human taste with AI-powered execution.
Works Cited
Goodspeed, Elizabeth. “Design Taste vs. Technical Skills in the Era of AI.” Nielsen Norman Group, 10 May 2024.
Kwok, K. (2020, June 19). Why Figma Wins. https://kwokchain.com/2020/06/19/why-figma-wins/
Lütke, Tobi. “Internal Company Memo on AI Usage.” Shopify, 2025.