By: Shan Tong, Principal Tech Lead – Visual Intelligence, Meta
Some of the hardest leadership moments begin when everyone is looking for clarity and no one knows where to find it yet.
At times, product teams and engineering teams are operating with competing priorities. Research is exploring future possibilities while delivery teams are trying to make immediate decisions. The challenge is that leaders still need to choose a direction, defend tradeoffs, and keep the team moving before every answer is known.
If you work in product, this tension might feel familiar to you.
Ambiguity in product leadership is not just about missing information. It is the responsibility to create direction while key pieces are still unresolved. Teams need priorities, decisions, and alignment before certainty arrives. That is why ambiguity is often not a detour from the work. It is the work.
That was exactly the challenge I faced when building Meta’s first generation of AI glasses in 2020. We were defining a product category that did not quite exist yet. Even the basic questions were open: What should the camera’s key use cases be? How do you create a great experience within the limits of battery life and processing power?
I used to think strong leadership meant having the answer early on in the process. In technical environments, I had learned to associate credibility with being prepared, precise, and ready with a recommendation.
Through building zero-to-one products, I learned something more useful: when the path is unclear, leadership is not about eliminating ambiguity on day one. It is about helping a team move through uncertainty with enough clarity, trust, and momentum to discover what works.
The best leaders do not wait for the fog to lift. They learn how to help teams move forward, even before the full picture comes into view.
Here are four practices that have helped me do that: experimenting, mapping, connecting, and multiplying.
1. Experimenting: Turn Uncertainty into Learning
When things feel unclear, teams often oscillate between two instincts: waiting too long for certainty to happen or committing too quickly in order to feel decisive.
Neither of these mindsets creates real progress.
A better response is to treat ambiguity like a lab. Start with a hypothesis, test what you can, and let the learning shape the next move.
I learned this lesson early while stepping into uncharted territory with Meta’s first generation of AI glasses. Instead of treating the challenge like an exam where I needed perfect answers to prove myself, I learned to treat it as a series of experiments: prototyping, testing, and iterating from feedback. This process shifted how I managed the project moving forward. It helped turn ambiguity into learning and laid the groundwork for a category that later gained mainstream traction.
The same mindset helped later when I built Hyperlapse. Video capture on AI glasses was capped at three minutes due to battery constraints. I looked beneath the constraint to the user need: people did not always want long footage; often they wanted a longer activity captured in a lighter form.
So I proposed Hyperlapse, a new mode to record long activities through distilled highlights. I built the prototype, architected the system, and led a cross-functional team for product integration. Hyperlapse extended capture time by 10× and was announced at Meta Connect 2025 as a flagship feature.
These experiences reinforced a lesson I keep coming back to: clarity is not always found before action. Often, it is built through action.
Seen this way, ambiguity stops feeling like a restriction and starts becoming a resource. Instead of predicting the future, we can build it by imagining possibilities that do not exist yet. Innovation begins by testing fast, learning quickly, and iterating forward.
2. Mapping: Make the Terrain Visible
Experimentation helps teams learn. Mapping is the practice of making complexity visible so teams can orient, align, and move forward.
Ambiguity feels heavier when it stays invisible. People use the same words to mean different things. Teams disagree without realizing they are working from different assumptions. Projects feel chaotic not only because they are complex, but because no one has yet made the complexity legible.
That is why mapping is such an important leadership skill.
When I moved from an individual contributor role into management, my scope expanded quickly — from one product line to four. The job did not simply get bigger. It became harder to see clearly. I could no longer go equally deep on everything, so I needed a better learning system.
What helped was learning like a cartographer. I started building a multi-scale view: zooming out to understand the full landscape, then zooming in on the few areas where deeper understanding would most improve decisions.
One of those areas was Mixed Reality Passthrough. It was complex, easy to misread, and important enough that a surface-level view would lead to poor decisions.
I worked with partners to build a multi-scale map: I laid out the foundational structure, then relied on domain experts to fill in the details. Together, we mapped the technical stack, the user experience, the key tradeoffs, and the places where evaluation was too fragmented. That work led to a more systematic Passthrough evaluation framework that connected technical specifications more directly to user experience.
The biggest lesson was not that I needed to know everything. It was that I needed to make the terrain visible enough for better judgment.
Good maps do not remove complexity. They make it navigable. Like cartography, leadership begins by quickly capturing the complexities of the terrain and distilling them into simple representations. Mapping helps you understand the terrain through a multi-scale learning system: zoom out for the big picture, then zoom in for more texture.
3. Connecting: Build Shared Direction across Functions
Different functions within an organization are often responding to the same problem from very different vantage points. Leadership may be asking a strategic question while the team is still solving an operational one. Everyone is usually right about something. No one is right about everything.
That is why leading through ambiguity is also an act of connection.
When I stepped into a broader tech lead role across imaging algorithms, I saw how much intelligence already existed across the organization and how easily it could stay fragmented.
To help bridge that gap, I organized the inaugural Camera Algorithm Roadmap Summit in 2024, bringing together more than 400 people across Product, Engineering, and Research. Together we built a clearer direction for camera experience. The shared direction translated into record-speed execution in 2025, when Meta shipped four new devices with dozens of new features.
The vision also laid the groundwork for unifying camera control systems. Previously, our camera control systems were fragmented across devices. By creating stronger alignment early, we mapped vague product requests into concrete architectures and built a more unified platform approach. Our in‑house solutions are now the standard across all Meta products—improving quality while cutting latency.
That experience changed how I think about cross-functional leadership. Sometimes progress does not depend on a better answer. It depends on a better conversation.
At its best, connection does more than reduce friction. It creates energy. It helps people see a future they want to build together and gives them a way to move toward it. That is why leaders in ambiguity have to be connectors who bridge worlds and help partners see how their work fits together.
The strongest teams do not only align, but they have a clear point of view on synthesizing ideas and balancing constraints to make complex systems work as one. The future belongs to teams that learn from each other faster than the problem evolves.
4. Multiplying: Scale Leadership Through Others
In leadership, multiplication means creating more people who can use good judgment and move the work forward. I learned this especially when managing engineers with very different styles. Some were naturally vocal. Others were quieter, equally capable, and easier to overlook in fast-moving environments. I realized that if I only rewarded the loudest voices, I would miss a lot of the team’s intelligence.
That risk grows in ambiguity. When ambiguity rises, many leaders respond by tightening control. It can feel reassuring, even responsible, to review more, decide more, and hold tighter. But over time, that instinct makes teams slower, quieter, and more dependent.
The strongest leaders I know do something different. They use ambiguity as a reason to grow more decision-makers, not fewer.
To grow my quiet team members, I became more deliberate about creating different pathways for influence. I assigned presentation opportunities, coached people on technical storytelling, helped them sharpen written narratives, and looked for ways to make underheard strengths more visible.
What I learned is that multiplication is not only a people principle. It is an operating principle. Teams move faster when leadership is distributed, trust is stronger, and more people are equipped to use their judgment well.
That is what real multiplication looks like. It is not just about delegating more. It is about bringing out and extending the intelligence of the team. Leadership matures when success is measured not only by what you deliver, but by who grows because of you.
Wrapping Up: Moving from Ambiguity to Agency
Ambiguity is uncomfortable because it removes the emotional comfort of certainty. It can make even experienced leaders hesitate, overcontrol, or cling to old assumptions.
It is where new categories are built. It is where strategy gets sharpened. It is where teams learn whether they can merely coordinate or truly think together. It is where leadership deepens from having answers to creating conditions where better answers can emerge.
That is what these four practices have taught me:
- Experiment so uncertainty becomes learning.
- Map so complexity becomes visible.
- Connect so fragmented insight becomes a shared direction.
- Multiply so leadership grows beyond one person.
Together, they offer a different model of leadership, one grounded less in certainty than in curiosity, clarity, collaboration, and trust. Ambiguity is not going away. In product, it rarely does. But that is not a reason to retreat. It is an invitation to lead differently.