Chapter 25: Skills for the Future

You are a product leader who has spent a decade perfecting the mechanics of software delivery. You know how to write a product requirements document that covers every edge case. You understand how to manage a scrum team with precision. You pride yourself on your ability to translate customer needs into technical requirements. Suddenly the tools you use have changed. Your engineers are using agents that generate thousands of lines of code in seconds. Your junior product managers are using large language models to draft strategy documents that look indistinguishable from your own. You feel a growing sense of unease. You worry that the skills that made you successful are becoming a commodity. The tension lies between the value of your historical expertise and the new requirements of an automated world. You face a choice between clinging to the mechanics of the past or mastering the unique human skills that artificial intelligence cannot replicate. If you do not evolve your personal talent stack you will become a relic of a slower era. You must decide whether to be a manager of tasks or an architect of outcomes. You must master the skills of the future or watch your impact diminish in a world of abundant intelligence.

CORE SKILL OR PRINCIPLE

The core principle of success in the artificial intelligence era is that human value shifts from execution to judgment, taste, and agency. When the marginal cost of building software drops toward zero the value of knowing what to build becomes paramount. Success depends on your ability to move up the ladder of abstraction. You must transition from being a bricklayer who writes code or documents to an architect who designs systems and models failure. This requires a fundamental shift in how you spend your time and how you train the next generation. You must prioritise soft skills such as empathy, listening, and influence over technical mechanics. You must cultivate a high degree of agency which is the desire to build and solve problems without a defined path. True excellence requires you to develop a signature taste that allows you to identify quality before the results are in. You must become a generative thinker who can produce novel ideas faster than the machine can execute them. You must treat technology as an Iron Man suit that enhances your existing capabilities rather than replaces them.

EVIDENCE FROM THE CONVERSATION

Evidence from leaders at OpenAI and Anthropic suggests that the way humans learn and work is undergoing a fundamental rewiring. Peter Deng observes that his nine year old son built custom agents that perform complex linguistic tasks that would have stumped adults a decade ago. This suggests that education must shift from memorising facts to mastering the ability to ask the right questions and think at higher levels of abstraction. Mike Krieger argues that curiosity and the scientific process of inquiry are the most important traits to nurture in a world where answers are cheap. He notes that the difference between someone who is adept at using these tools and someone who is not is currently vast but will eventually become a baseline requirement.

Leaders at Replit and Cursor believe that the soul of software engineering is shifting from routine implementation to system architecture. Amjad Masad identifies that the return on investment for learning to code is doubling every six months because technology provides more power to those who understand the mental models of how software is built. He argues that the bottleneck in production has shifted from the making of things to the generation of ideas. Michael Truell adds that taste will become the primary differentiator as software becomes more logical and less about manual pixel pushing.

There is a growing recognition that agency is the most undervalued skill in the current market. Varun Mohan observes that many elite graduates are trained to follow well-defined paths but lack the agency to build something from scratch. This desire to take initiative and solve problems independently is what will separate the most successful individuals in an automated workforce. Karina Nguyen identifies that research progress itself is bottlenecked by management and the ability to prioritise paths based on high conviction.

Soft skills are increasingly viewed as hard competencies that machines cannot easily replace. Claire Vo and Karina Nguyen both argue that influence, convincing others, and aligning human systems are the most difficult tasks to automate. Ken Norton suggests that the art of communication and collaboration will be elevated as technical tasks are offloaded. Julie Zhuo emphasises the importance of emotional regulation and introspection as the tools around us become more comfortable and potentially lead to stagnation.

PRACTICAL BREAKDOWN

You must teach yourself and your team how to move up the ladder of abstraction. Start by treating technology as a junior buddy or an intern that requires clear guidance and constant review. Stop focusing on the syntax of code or the formatting of a document and start focusing on the logic and system design. You must develop a mental model of how computers and systems work including concepts like parallel processing and memory. This foundational knowledge allows you to unblock agents when they get stuck.

To develop good taste you must expose yourself to masterpieces and failures alike. Taste is the ability to identify what is good without needing to see the business results first. You must hold yourself to high standards of craft and meticulously care about the details of the user experience. Practice identifying the core logic of a product and use your judgment to decide which features provide the most value.

You must follow your curiosity as a strategic advantage. Pursue every curiosity and play with new technology to see what is possible. Use agents as a curiosity engine to ask anything and learn about any topic in depth. Pick a specific goal or a weird project and use new tools to solve it rather than just reading about them. This active play rebuilds your intuition for what technology can and cannot do.

Increase your agency by stepping out of your functional lane and taking ownership of whole problems. You must be a doer of things who builds prototypes and ships features regardless of your title. Avoid waiting for a defined path or a set of instructions from your manager. Instead you should be the one telling your leader what needs to be done based on your understanding of the customer and the business.

Become more generative by training your muscle for idea creation. When you find a task is easy to execute you must use that freed up time to explore more ambitious ideas. Use tools to automate your energy drains such as routine administrative tasks so you can focus on the work that truly drives growth. Conduct a regular energy audit to identify what fuels your work and what drains you.

Master the skill of debugging automated output. You must learn to take a holistic view of the system to understand where the source of an issue might come from. Do not simply trust that the output is correct but spend time reviewing and validating the work. Use technology to generate test suites and unit tests to help you validate the code more efficiently.

SKILL APPLICATION

Apply these future skills to your daily discovery work by using technology to expand the frontier of hypotheses you can test. Use the eigenquestions framework to simplify complex problems down to the one or two questions that truly matter. This higher level thinking allows you to avoid getting bogged down in the surface level details of a solution.

Operationalise your taste by requiring a specific rubric for quality in every product review. You must be the chief tastemaker who ensures that every user journey is seamless and reflects meticulous craft. Use tools to push your taste out to the edge of the organisation by codifying your standards into prompts that writers and builders can use.

Manage your team velocity by imbuing them with a sense of urgency and a culture of play. Encourage them to use new tools for their day-to-day workflows and assess their performance based on their ability to leverage these tools effectively. Replace traditional status meetings with demo Fridays where everyone shows working prototypes and shares what they have learned from playing with the technology.

Build trust with your users by navigating the agency and control tradeoff in your products. Start with high control and low agency for your automated agents and then slowly move up the rung as you build confidence in their performance. Use rigorous evaluation metrics to measure how well the technology automates the core value chain of your business.

Establish a mentor and intern mindset where you are simultaneously a student of your craft and a teacher to your team. Avoid becoming a domain specialist in an exhausted field and instead seek out roles that put you close to the technical frontier. You must manage your personal growth curve as it is the primary constraint on your professional impact.

ACTION CHECKLIST

• Block out two hours this week to play with a new software tool you have never used before.
• Audit your calendar and identify one energy-draining task that you can automate using an agent.
• Ask your manager what one skill they believe you need to build to be ready for the next technological shift.
• Write a one-page rubric for what good looks like for your current product and share it with your team.
• Identify a complex problem on your roadmap and spend an hour reverse engineering its core logic.
• Schedule a demo Friday where your team shows working prototypes rather than slide decks.
• Interview one customer who recently stopped using your product to find the struggling moment that triggered their decision.
• Draft a return on investment model for your product that includes cost savings and opportunity costs for the user.
• Perform a walk the store review of your onboarding journey on a mobile device and write a friction log.
• Use tools to generate a list of 50 failure modes for your current product to help you build better evaluations.
• Disclose one mistake you made recently to your team to build deeper trust and connection.
• Ask a junior member of your team to teach you a skill that they are more native in than you are.
• Set a personal service level agreement to respond to all team dependencies within four hours this week.
• Replace one recurring status report with an asynchronous update in a shared document.
• Identify one high conviction low consensus bet and document the logic behind it.
• Create a gratitude house list of everyone who has helped you reach your current role.
• Spend 30 minutes reading the documentation for a system you do not fully understand to build your mental model.
• Use the What So What Now What structure for every update you send this week.
• Ask yourself if you would fully fund your own team if you were the chief executive today.
• Commit to a six week execution cycle for your next major project and use a circuit breaker to manage scope.