This is version 1.0. Fish detection, species identification, and other features are coming.

Sometimes the best way to understand where technology is going is to build something completely unnecessary.

More to come. Here is the live YouTube stream from the cottage.

This year, 17 engineers gathered around the table. We all have one thing in common. We either studied semiconductors as electrical engineers, built our careers in the semiconductor industry, or, as was true for most of us, both.

In one way or another, silicon brought us together.

For one evening every year, we celebrate not just an industry, but a lineage.

We have one tradition that never changes. Instead of introducing ourselves by our companies or job titles, everyone starts by sharing which edition of Sedra & Smith they studied from.

Benny Lau and Lee Lau were zero edition. They learned directly from Professor Adel Sedra’s class notes before those notes became a world-class textbook used by generations of electrical engineers. Martin Snelgrove completed his PhD under Professor Sedra and later became his colleague. Professor Tony Chan Carusone, another former student, co-authored the eighth edition and is now writing the ninth with Vincent Gaudet. I am second edition and had the privilege of being one of Professor Sedra’s students.

Looking around the table, it struck me that much of the story of Canada’s semiconductor industry can almost be told through one professor, one textbook, and several generations of students.

This year, Professor Sedra joined us as our guest of honour again.

We were also joined by Benny Lau and Lee Lau, co-founders of ATI. Long before GPUs became one of the world’s most strategic technologies, ATI pioneered the graphics chip industry and became the world’s largest graphics chip company. Sorry, NVIDIA, but before ATI was acquired by AMD nearly 20 years ago, ATI was bigger and better!

Even today, many of AMD’s GPUs continue to come from Toronto.

ATI’s influence was visible all around the table. In addition to Benny and Lee, Eva Lau, Jim Seto (CEO of ePIC), Hugh Chow (CEO of ventureLAB), and Amy Lau (now at Hepzibah AI) all worked at ATI. Six of the 17 people around the table were ATI alumni. ATI’s fingerprints are everywhere in Canada’s semiconductor ecosystem.

The next generation of Canadian semiconductor companies was equally well represented.

We had Professor Tony Chan Carusone, former CTO of Alphawave, before its acquisition by Qualcomm for US$2.4 billion.

We also had Nish Sinnadurai, Senior Vice President of Core Software and General Manager, and Val Manohararajah, Chief Product Architect of Cerebras. Cerebras needs no introduction and recently completed one of the biggest technology IPOs in the world this year.

Around the table were also Martin Snelgrove and Raymond Chik of Hepzibah AI, Gem Shoute of Zinite, Niraj Mathur of Blumind, and Professor Joyce Poon of Axiomatic AI. Together, the group spanned multiple generations of Canadian leadership in semiconductors, AI computing, photonics, and other advanced computing systems.

What makes this dinner special is not only who is around the table. It is the conversations that become possible when this particular group comes together.

We talked about the convergence of hardware and software as AI dissolves the traditional boundary between the two. We also discussed why the next bottleneck in AI is all about data movement, not simply computation. Moving data between memory and processors, across chips, and throughout entire systems often consumes more time and energy than the computation itself. Addressing this bottleneck will require innovation across the entire computing stack.

And, of course, we heard insider stories from decades of building the semiconductor industry, including never-before-publicly-shared conversations with TSMC founder Morris Chang. These were not simply historical anecdotes. They offered firsthand lessons on building world-class technology companies, forging enduring partnerships, and navigating industry shifts.

These are conversations that could only happen around this table.

Every year, I leave this dinner with the same feeling.

Canada has been producing world-class semiconductor talent and companies for decades. The people around this table have built category-defining companies, trained generations of engineers, and are now creating the next wave of breakthrough technologies.

The future of semiconductors is being written around the world. Canada has already written some remarkable chapters, and judging by the people around this table, many more are still to come.

In fact, some of the world’s best next-generation semiconductor companies are already being built in research labs and startups across Canada. We are seeing it firsthand. Most people simply have not heard of them yet.

Two Small Fish Ventures is already backing some of these companies and will continue to support the best founders at the earliest stages. We are working hard to ensure that this remarkable Canadian semiconductor lineage not only continues, but becomes even stronger.

This post is a bit more personal.

Long before I became a founder, CEO, or venture capitalist, I was a graduate student in the Department of Electrical and Computer Engineering at the University of Toronto’s Communications Group under the supervision of Professor Elvino Sousa. I spent a significant part of my academic life thinking about one problem: signal processing. More specifically, how do you recover meaningful information from an incredibly weak and noisy signal?

That question consumed years of my life. It eventually led me to writing GSM signal processing software—what we simply called 2G back then—early in my career. That experience launched the rest of my journey in computers and software. I thought I had left signal processing behind for good.

One thing led to another: startups, Wattpad, and eventually Two Small Fish Ventures. Today, we invest in what we believe is the next frontier of computing and its applications: frontier AI, quantum computing, semiconductors, physical AI, smart energy, and other foundational technologies. Through a combination of curiosity, experience, and more than a little serendipity, that journey has somehow brought me back to where it all started.

Back to signal processing.

Why?

Because, at its heart, quantum computing is not just a physics problem. It is also one of the hardest signal processing problems ever tackled. And that is exactly why this investment feels so personal.

Quantum computers operate using extraordinarily fragile signals. Extracting information from those signals without disturbing them is incredibly difficult. Two enemies dominate the problem.

The first is noise. Any unwanted interference can overwhelm the tiny quantum signals carrying the information you are trying to measure.

The second is heat. Every amplifier generates heat, but superconducting quantum computers operate at temperatures only fractions of a degree above absolute zero. Even a very small amount of additional heat can destroy the quantum states you are trying to preserve.

As quantum computers scale from hundreds of qubits to millions, these two challenges become existential. Building better qubits alone is no longer enough. We also need fundamentally better ways to read and control them.

That is exactly what Qubic is building.

The real Qubic quantum amplifier. Sometimes the future of computing doesn’t look like a supercomputer. It looks like a beautifully engineered piece of signal processing.

Its quantum amplifier dissipates approximately 10,000× less heat than today’s cryogenic amplifiers while simultaneously manipulating quantum signals with extraordinarily low added noise. Solving either of those problems would be significant. Solving both at the same time could fundamentally change how quantum computers scale.

Equally important, the technology extends well beyond quantum computing. The same ultra-low-noise signal processing techniques could enable a new generation of sensing, communications, and defence systems that operate using signals far weaker than what is practical today. That opens the door to capabilities that were previously impractical—or even impossible.

One lesson I have learned over the years is that every computing revolution eventually runs into an infrastructure bottleneck.

For classical computing, it was semiconductors. Faster, denser, cheaper, and more power-efficient chips unlocked decades of exponential progress.

For AI, it is compute and energy. As models become larger and more capable, both have become fundamental constraints on what is possible.

For quantum computing, one of the defining bottlenecks is signal processing. Qubits are extraordinarily fragile, and scaling them depends on reading and controlling them with unprecedented precision.

History suggests that solving these infrastructure bottlenecks creates enormous value.

We often celebrate the companies that build the applications.

But just as often, the biggest winners are the companies building the enabling technologies that make those applications possible in the first place.

That is one of the reasons we were so excited to lead Qubic’s seed round.

For me, though, this investment means something more.

It is a reminder that careers are rarely linear. The problems that first fascinated us as young engineers have a funny way of finding us again decades later, often in places we never expected.

I thought I had left signal processing behind.

Through a very long and winding path, it found me again.

Maybe that’s because everything starts with a signal.

We also hosted a School of Fish dinner with a group of researchers, engineers, and founders. Over good food and drinks—including one of the best schnitzels I’ve ever had—the conversation ranged from AI to neuromorphic computing and other areas pushing the boundaries of what is possible.

Berlin reminded us why it is one of the world’s great deep tech hubs. The city brings together exceptional founders, researchers, engineers, world-class universities, and leading research labs, creating an environment where curiosity, technical excellence, and ambition reinforce one another. It is the kind of ecosystem where breakthrough ideas can move from the lab to the marketplace.

As we continue investing globally, we are constantly reminded that breakthrough companies emerge from many different places. Geography can influence a founder’s path, but it is talent, determination, and vision that drive outcomes. We feel fortunate to meet and support remarkable founders around the world who are building the future.

More and more, those founders are reaching out to us as well. That’s a privilege we don’t take for granted.

Canada’s new AI for All strategy is a welcome step forward.

Compared to earlier AI strategies focused primarily on research and talent, AI for All expands the conversation to commercialization, capital, IP retention, sovereign capabilities, and supporting Canadian AI champions. That’s real progress.

In fact, I see it as part of a broader and very positive shift in Canada’s innovation mindset: moving beyond simply supplying talent, research, and ideas to the world and toward owning more of the value those inputs create.

In my previous post, Stop Supplying, Start Owning, I argued that Canada has spent decades becoming exceptionally good at supplying. We supply natural resources, talent, research, inventions, entrepreneurs, and increasingly, AI innovation. But too often, we have not owned a proportional share of the economic value those inputs create.

What I find encouraging about AI for All is that many of its themes point in the opposite direction. The strategy’s focus on commercialization, IP retention, sovereign capabilities, and scaling Canadian firms suggests a growing recognition that creating innovation is only the first part of the challenge. Capturing the value created by that innovation is the next step.

Notably, while the strategy discusses many ownership-related concepts, the word ownership itself does not appear. That may seem like a small detail, but it raises an important question:

What does winning look like?

Because strategy is ultimately about outcomes, not activities.

By many measures, Canada has already been remarkably successful in AI. We helped drive many of the breakthroughs behind today’s AI revolution. We trained world-class researchers and built respected institutions. By those measures, Canada is winning. We are one of the reasons modern AI exists. I would even argue that Canada invented AI.

However, in Stop Supplying, Start Owning, I argued that Canada already lost Game 1. We helped create enormous value through our talent, research, and innovation, but captured only a fraction of the economic upside. If the definition of winning is broad-based prosperity created from the innovations we help create, then our track record is far less impressive.

The opportunity now is not simply to create more AI innovation. It is to capture more of the long-term value created by that innovation.

That requires distinguishing between commercialization and ownership.

Commercialization asks:

Can we build a business?

Ownership asks:

Can we capture a meaningful share of the value that business creates?

The two are closely related, but they are not the same. A company can be successfully commercialized while much of its long-term economic value accrues elsewhere.

If AI is truly going to benefit all Canadians, Canada needs to retain a meaningful share of the value it helps create. Capital, compute infrastructure, talent development, procurement, and commercialization support are all important. They increase the odds of success, but they are not the objective.

Think of it like a hockey game. Scoring goals is necessary, but scoring goals is not the same as winning. A team can score five goals and still lose 10–5. Capital, talent, infrastructure, research, and commercialization are like goals. Ownership is closer to the final score.

If Canada creates the talent, funds the companies, builds the infrastructure, and develops the technology, but someone else ultimately owns most of the economics, then we may have scored impressive goals while repeating the outcome of Game 1.

This is not about excluding foreign capital or avoiding global partnerships. Building globally significant companies from Canada will require global customers, global talent, global investors, and global ambition. Ownership is about ensuring that a meaningful share of the value created by Canadian innovation remains connected to Canada.

That is why I view AI for All as such an important evolution in Canada’s innovation strategy. It recognizes that research excellence alone is not enough and reflects a shift from a supplier mindset toward an ownership mindset.

The next step is making the destination explicit.

As currently framed, success can easily be measured by how much capital was deployed, how many companies were funded, how many successful companies were created, how many programs were launched, and how many announcements were made. These are useful metrics, but they do not necessarily mean we are winning.

A decade from now, the more important question will be whether Canadians captured a meaningful share of the value created by the AI revolution. If we want AI to be for all Canadians, then Canadians must own and capture a meaningful share of the value this strategy creates. Otherwise, it becomes very difficult for the benefits of AI to be broadly shared across the country.

AI for All and ownership are not separate objectives. They are deeply connected. Canada already lost Game 1, but we now have a chance to win Game 2.

Canada has scored a lot of goals in AI, and AI for All should help us score many more. Scoring goals is great, but the objective is not to score goals.

The objective is to win the championship.

Earlier this year, GitHub COO Kyle Daigle said GitHub is now seeing 275 million commits per week, versus about 1 billion commits in all of 2025. For non-developers, a commit is basically one saved change to code. It is not a perfect measure, but at GitHub’s scale it is a very good proxy for the amount of software being created, updated, and shipped into the world. Annualized, that implies a pace of roughly 14.3 billion commits this year. That is about 14x last year’s total. It is not a precise year-over-year weekly comparison, and whether the true number is 10x, 14x, or 20x is not really the point. The point is this: software creation is not slowing down. It is exploding by at least an order of magnitude.

There is a growing belief that software is becoming less valuable as AI gets better at writing code.

I believe the opposite is happening. Software is not dying. It is becoming abundant. And when something becomes abundant, value does not disappear. It shifts.

We have seen this movie before. When electricity became cheap and reliable, the world did not use less of it. It reorganized around abundant power. When PCs became widely available and inexpensive, we did not use less computing. We put computers everywhere. When the internet made connectivity cheap and ubiquitous, we did not communicate less. Connectivity became the foundation for entirely new products, behaviours, and industries.

In my Jevons Paradox post, I made exactly this point: when the cost of something foundational collapses, demand usually goes up, not down. More importantly, behaviour changes with it. Cheapness does not just lower cost. It expands the market, unlocks new use cases, and changes what people do.

Now software is going through the same transition.

As the cost of creating software collapses, the amount of software in the world is going to explode.

This should not be surprising. It is just Jevons Paradox playing out again. Efficiency does not simply reduce cost. It expands the market. It unlocks new demand. It changes behaviour. Cheap electricity gave us electrified cities. Cheap computing gave us the digital economy. Cheap connectivity gave us the internet economy. Cheap intelligence will give us a world with vastly more software.

This fits squarely within a thesis we have held at Two Small Fish Ventures for a long time: the collapse in the cost of intelligence will drive massive behavioural change. In my earlier post, Software Is Dead, Long Live Software, I wrote that when the cost of software creation approaches zero, the world does not get less software. It gets flooded with it. GitHub’s number is simply the latest evidence that this thesis is playing out.

For decades, software was constrained by two things. First, by how fast humans could write programs in computer languages. Second, by how few humans could do it well. That scarcity shaped the economics of software.

That constraint is now being lifted.

What AI changes is not just speed. It changes who gets to create. It changes what becomes economically possible. And it creates entirely new problems to solve.

When the cost of creation falls, more things become worth building. Software that was once too niche, too customized, or too small to justify a team suddenly becomes viable. More workflows get software. More edge cases get software. More individuals and companies can justify building exactly what they need.

At the same time, abundance creates new challenges. How do you manage software that is highly customized, rapidly changing, high volume, or even disposable? How do you secure it, govern it, and monitor it when it is being created faster than most organizations know how to handle?

That is where the next opportunity is.

Yes, some parts of software will become harder to defend. Generic wrappers and products whose only moat is “we wrote the code first and faster” may have worked in the past, but they will have a very difficult time building durable businesses going forward. That does not mean the opportunity in software is shrinking. It means the source of value is changing.

Some of the most important companies of the next decade will be built for a world where software is abundant, not scarce.

We have already backed some founders building for that world. We are looking for more.

And the winners will be the ones who understand what a world of abundant software makes possible.

Canada arguably invented modern AI. We trained the talent, helped build the science, and produced many of the key researchers, yet most of the important AI companies ended up being built and owned elsewhere. That was Game 1, and we lost it.

The good news is that Game 2 is now underway in quantum, robotics, physical AI, space, advanced manufacturing, and smart energy. Canada also has world-class scientists in all of these areas, but more importantly, these are not businesses that can be built anywhere with a few smart people and a laptop. They depend on specialized infrastructure, research labs, and ecosystems that are much harder to move.

That is why this moment is so interesting. For Canadian family offices, this is not just about national pride or supporting innovation close to home. It is a generational opportunity to back and own part of the next wave of world-class Canadian companies, right in our own backyard, before the rest of the market fully wakes up to what is forming.

I wrote about this in my latest piece for Canadian Family Offices. You can read it here.

Great to see Financial Post highlight Canada’s next wave of semiconductor opportunities.

As AI adoption accelerates, the demand for more efficient chips is creating room for a new generation of deep tech companies in Canada.

For many years, we at Two Small Fish Ventures have been investing in the next frontier of computing and its applications, including semiconductor companies like Zinite (3D chip technology), Blumind (ultra-low-power, all-analog AI chips), and Hepzibah AI (ARM for AI inference). These founders are building foundational technologies for the next generation of computing, addressing the speed, latency, and energy challenges in AI.

As Eva shared in the article, we are seeing many high-quality Canadian deep tech companies emerging from research breakthroughs. Some are still early. Some are still in stealth. Some are becoming breakout successes.

Exciting time indeed!

As a founder, I attended many founder summits over the years. Now, as an investor, I am hosting our own.

Based on my own experience and from speaking with many founders I know, if I had to pick one thing that allows founders to get the most out of a founder summit, it is peer-to-peer learning, not celebrity speakers.

That is how we designed our Two Small Fish Founder Summit.

We wanted to create an interactive environment where founders could learn from other founders, not through generic advice, but through lived experience: what worked, what did not, what surprised them, and what they are still trying to figure out.

As a thesis-driven VC focused on deep tech, half the battle is won naturally.

Many of our founders come from similar technical backgrounds. They are building in adjacent areas, exploring similar frontiers, and facing similar company-building questions. More importantly, they have learnings that other founders in the room will find directly relevant because they are on a similar journey.

This is also where our own experiences make a difference. Our whole investment team comes from backgrounds similar to many of our founders. We have been engineers, operators, and company builders. We know how hard it is to turn technology into a product, a product into a company, and a company into something enduring.

We are not outsiders looking in. We have been there. We understand the company-building journey because we have lived it ourselves.

But relevant context is only part of it. For founders to share openly, they also need trust. We wanted to create a safe space for founders to ask hard questions, share what they have learned, and help each other out.

To make that possible, our team also shared some of our own painful war scars from building companies and working with founders over the years. Many of these are lessons we have never shared externally. But in a confidential setting, they became part of the conversation.

That is what makes the conversation different. It is focused enough to be relevant, and trusted enough to be honest.

For our Founder Summit, we also focused on unique topics that are particularly relevant and useful for founders coming from deep tech backgrounds. Building world-class companies requires more than great technology. It requires founders to keep building new muscles in many other areas.

We believe we delivered something uniquely Two Small Fish Ventures: focused, relevant, interactive, high-trust, and founder supportive for deep tech founders.

Thank you to all the founders who joined us, especially those who travelled from out of town.

Thank you to Matt Clancey, Program Director for COVE’s NATO DIANA Accelerator Program, for joining us as our guest speaker and leading such an informative session. As a veteran and also an engineer, Matt brought a perspective that was especially relevant to many of the founders in the room. COVE’s work is especially important at this moment.

May is Asian Heritage Month in North America, and one topic that often surfaces is the “bamboo ceiling,” the observation that Asians are often underrepresented in leadership positions relative to their educational attainment and other professional successes.

Based on my experience and observation, one reason is the tension between Eastern and Western value systems. Many Asians are taught to:

• Respect authority vs. embrace disagreement

• Maintain harmony vs. embrace pushing back

• Be humble vs. embrace self-promotion

• Listen first vs. embrace speaking up

• Let your work speak for itself vs. embrace advocating for your work

• Accept the answer vs. embrace finding another path

The list goes on. Inherently, there is no right or wrong in these value systems. But they can be incompatible because they are optimized for different paths to outcomes. One is optimized for harmony. The other is optimized for constructive tension. You cannot optimize for both at the same time, just as you cannot turn left and turn right at the same time.

That tension helps explain the bamboo ceiling. I have a unique perspective on this. I was born and raised in the East and have spent my entire adulthood in the West. I am culturally bilingual in both systems.

Although the bamboo ceiling is real, I do not think this system incompatibility is always the root cause. Often, it is the symptom. The deeper issue is discomfort avoidance.

Discomfort avoidance is universal. It is human nature. We avoid difficult conversations, disagreement, asking for what we want, and situations where we might be rejected, criticized, or misunderstood. This is not unique to Asians. But when you are operating across incompatible systems, you naturally have to face more discomfort. Many Asian cultural values can also amplify the instinct to avoid it. I know they did for me.

For a long period of my career, I believed that if I worked hard, delivered results, and treated people well, everything else would take care of itself. Until I realized that leadership in the West requires something different. As an extreme introvert, I was further handicapped by my discomfort avoidance.

The breakthrough came when I became more comfortable with discomfort. It unleashed the full potential of what I could do. I started doing things I never thought I could do as a young adult.

Today, if I disagree with you, you will hear it from me in a respectful and diplomatic way. If I believe something is important, I will speak up. If I think a decision is wrong, I will say so. If someone tells me no, I may respectfully challenge that answer and look for another path forward, all in a Western-compatible way. Not because I enjoy debate, but because I have learned to embrace healthy debate.

Growth, leadership, and influence live on the other side of discomfort. Breaking through that mindset was liberating.

I did not have to abandon the values I grew up with. I simply had to reinterpret them. More importantly, I leveraged my cultural bilingualism and turned it into one of my biggest unfair advantages. I can leverage the best of both worlds and thrive in both.

This Asian Heritage Month, I am grateful for the values my culture gave me. Keep them. Leverage them. They are incredible tools in your toolbox. I am equally grateful for learning when to push beyond them.

For many of us, the most important barrier we need to break through is the discomfort inside our own heads.

This year, my main stage keynote was a masterclass called Calculating Risk, Maximizing Upside.

Many people are perceived to be risk averse. Over the years, I have discovered that they are not really risk averse. They just do not have a good framework to help them understand how to take risk in order to achieve outsized outcomes.

Risk shows up everywhere. Should I start a company? Should I join a startup? Should I stay on the “safe” path? Should I bet my career on AI? Should I work for a big company or a small one? Should I move faster, or wait until the path is more obvious?

Underneath all these questions is the same concern: it feels risky because the outcome is uncertain.

In other words, the goal is not to avoid risk. The goal is to understand the risk you are taking and make sure the potential upside is worth the leap. That does not eliminate uncertainty, but it turns uncertainty into something you can work with.

This simple framework can be described in three pillars.

1. Set your bottom line

Not every risk is worth taking. Not every loss is acceptable.

Before emotion takes over, define the line. How much are you willing to lose? What is the downside you can survive? Where is the point where ambition turns into recklessness?

A bottom line gives you the freedom to act. Once you know the boundary, you can fight like crazy and swing for the fences.

2. Know your edge

The same risk is not the same risk for everyone.

One person may look at an opportunity and see danger. Another person may see something very different because they have technical depth, domain expertise, data, distribution, lived experience, credibility, network, or speed.

Your edge changes the risk equation.

This is especially true in deep tech. Some opportunities look irrational from the outside because the underlying technology is not yet obvious to the general public. But if you understand the technology, the product, the disruption, and the inflection point, the “crazy” bet may actually be calculated.

3. Chase the upside

If the downside is real, the upside must be extraordinary.

This is especially true for venture-backed startups. They are not built for 10 percent improvement. They are built for 100x outcomes.

In other words, small upside does not justify massive sacrifice.

This applies beyond founders too. If you are choosing a career path, joining a startup, or betting on an emerging technology, ask the same question: is the upside big enough to justify the downside?

The risk-reward profile has to be asymmetric.

This framework can be summarized into three simple points:

1. Downside defines the risk.

2. Edge improves the odds.

3. Upside justifies the leap.

Or even better, three simple words:

Downside. Edge. Upside.

I used numerous real-life examples, including many from my own founder journey, to show how the framework applies in practice.

The best bets are not fearless. They are bounded, advantaged, and asymmetric. That was the heart of the masterclass.

That is the difference between taking risks and taking risks thoughtfully and methodically.

P.S. I was also on the Winning with Responsible AI panel at Upper Bound. I have more thoughts on that topic, but I will save them for a separate post.

P.P.S. This talk was closely related to the framework Eva Lau discussed in her talk when she was conferred the degree of Doctor of Laws, honoris causa. This is the framework we have been using for a long time, with a lot of success.

In the past few years, I have been spending much of my time with researchers, professors, and deep tech founders sitting on remarkable breakthroughs and asking a difficult question:

What does it take to turn this innovation into a world-class company?

That question has become central to how I think about this chapter of my life.

This month marks the fifth anniversary of the Wattpad acquisition in 2021. That milestone naturally invites reflection. I am deeply grateful for that chapter and everything I learned from it, but I have 100% moved on.

Why?

Because I have even more fire in the belly for what comes next.

People sometimes ask me: Do I miss operating? Why am I doing this now? Why not just start another company and do it all over again?

Those questions all lead to a deeper one:

What does winning look like for me now?

For me, winning has always been about impact. Of course, many other things matter too. But without impact, I am not interested.

That has not changed.

What has changed is where I believe my experience is most useful, where I can have the highest-leverage impact, and what kind of challenge I feel most drawn to now.

I have no interest in simply repeating the last chapter. I have already co-founded and led one iconic company that reached 100 million users globally and became a cultural phenomenon. Fewer than two hundred companies in the world have achieved that 100-million-user milestone. Even fewer founders have stayed for the full marathon, from the first user all the way to a home run.

Rather than taking it easy, I want a new and even more ambitious challenge.

I have come to believe that one of the most important challenges in deep tech today is helping scientists and deep tech founders cross the very difficult bridge from scientific and engineering breakthrough to building a world-class company.

That is the work I feel most drawn to now. It is also work I feel unusually prepared for.

I started three companies. Between them, I have seen almost every major founder outcome: one VC-backed failure as CTO with tons of very stupid mistakes, one bootstrapped company as CEO with a small exit, and one VC-backed home run in Wattpad as CEO.

From the outside, Wattpad looked deceptively simple. In reality, it became an incredibly complex global business spanning consumer, enterprise, subscriptions, virtual goods, social media, frontier technology, traditional publishing, entertainment, and even geopolitics.

We were one of the pioneers in UGC and mobile, before the iPhone even existed. Over time, Wattpad also became a deep tech company. We started building our first of many proprietary machine learning models around 2013, long before most people had even heard of AI. That shaped how I think about technology disruption and how real technical innovation becomes enduring advantage.

I also lived through strategic investors, IPO preparation, multi-bidder acquisition processes, the constant need to disrupt and reinvent ourselves, moments when tech giants were clearly trying to kill us, and a few phoenix-from-the-ashes moments.

I do not say any of this to dwell on the past. I say it because seeing that many startup permutations from the inside gives you a massive toolbox of practical judgment and wisdom.

Those tools were built and tested through real-world operating experience from day one, scaling a business into a world-class company while enduring the full roller-coaster ride.

You learn them by starting a company when the market is not ready. You learn them by doing things no one has done before. You learn them by making painful mistakes. You learn them by carrying a company through long periods when it is working, but not yet obvious. You learn them by navigating the transition from prototype to product, from product to business, from business to company, and from startup to an enduring world-class venture.

That sequence matters deeply in what I do now, because very few people have lived it end to end: a breakthrough is not yet a product. A product is not yet a business. And a business is not yet an enduring, world-class company.

The bridge between those stages is turning innovation into product and then commercializing it. That is where many promising breakthroughs fail. Not because the innovation is weak, but because turning groundbreaking innovation into a world-class company requires a different set of skills and judgments: timing, technical moat, customer need, business model, smart capital, team building, endurance, and grit.

You need to understand how all of the pieces connect and work together to build a world-class company.

I now know what I did not know.

That belief sits at the center of how I now spend my time. My mission is to unleash the full potential of scientists and their innovations by helping them become entrepreneurs and build world-class companies.

That, more than anything else, is what this chapter is about for me.

That is also part of what makes TSF distinctly TSF.

We are not trying to be a generic early-stage fund. Across our whole investment team, without exception, we bring together a combination that is very rare in venture globally: deep technical chops, real-world operating experience, scar tissue from world-class company building, and a mission-driven commitment to helping scientists and technical founders build enduring, world-class companies. We are also proven full-cycle investors with the experience of backing multiple companies of exceptional scale.

To me, TSF is not simply an investment platform. It is one of the highest-leverage ways I can apply what I have learned to make the greatest impact.

Building one iconic world-class company is hard. Leveraging that experience to help build multiple world-class companies may be even harder.

That is the challenge I want.

The next chapter started four years ago when I stepped down as Wattpad CEO, and I have been building into it ever since.

I am entering this next stretch with a lot of fire in the belly and 100% commitment. For me, that is what winning looks like.

My last post — Stop Supplying. Start Owning. — was one of the most engaging posts I have published in recent weeks. One part of that talk that I did not go deep enough on was the banana analogy.

I have been using this framework in talks for some time now — most recently at the Engineering Deans Canada annual meeting in Winnipeg, and before that in a keynote at York University’s Schulich School of Business on a related but different topic: “Think like an owner”. I think this analogy deserves its own post.

So for this Sunday morning, something a little lighter. Just five banana lessons that I think apply whether you are a student figuring out your first job, a founder building a company, or an investor trying to understand where value actually lives.

Yes, I went bananas — but not in a way you might think. It all starts with a monkey.

The Setup

Jack Ma once said: if you put money and a banana in front of a monkey, the monkey takes the banana. It does not know that money can buy many bananas.

I love this analogy. On the surface, it is a simple observation about short-term versus long-term thinking. But the more I have sat with it — through three companies with well over a thousand employees across three companies, two exits, and now as an investor in over 60 companies — the more I think it contains an entire philosophy of value creation.

Here are the five lessons I have drawn from it.

Lesson 1 — Cash Is Better Than a Banana

Jobs are bananas.

People grab them because a steady salary and job security feel safe. And they are not wrong to. A banana feeds you. It is real. It matters.

But a banana feeds you once. Cash — real ownership, real equity, real stakes in what you are building — feeds you many times. The problem is not that people value jobs. The problem is that too many people never stop to ask whether there is something better in front of them.

Recognizing that alternatives exist is the first step toward creation and ownership. Most people never take that step — not because they lack ambition, but because nobody ever put the alternative clearly in front of them.

Lesson 2 — Some Bananas Are Better Than Others

Not all jobs are created equal.

Some opportunities teach you, compound your skills, put you inside exceptional teams, and give you proximity to how great companies are actually built. Others keep you comfortable but stuck — including, surprisingly, many high-paying jobs at large, slow-moving organizations.

The number of jobs created is not a useful measure of prosperity. The quality of those jobs — and what they teach, and what they lead to — is what matters.

When I was early in my career, I worked at Delrina, one of the most successful Canadian tech companies of the 1990s. The company grew from 20 people to 800 in four years when Eva and I worked there. We learned more in that rocketship environment than I could have anywhere else. That was not a banana. That was a greenhouse banana — rare, valuable, and 100% worth seeking out.

The lesson: be deliberate about which banana you grab. Not all of them are the same.

Lesson 3 — A World-Class Banana Tree Is Better Than a Banana

A banana feeds you once. A tree feeds you forever.

This is where the shift from employee thinking to founder thinking begins. One banana is a salary. A tree is equity, ownership, and compounding returns on something you built.

Ownership compounds. Wages do not.

The number of jobs created is still the wrong KPI. A single world-class company — owned, scaled, and defended — creates more durable economic value than a thousand comfortable jobs at organizations that will be restructured, acquired or hollowed out over time.

Do not just own the banana. Own the tree.

Lesson 4 — A Banana Farm Is Better Than a Tree

A tree is better than a banana. But a farm is better than a tree.

Dole, a company valued at just over $1 billion and one of the most recognizable fruit brands in the world, does not just grow bananas. It operates a vertically integrated business — owning farmland, managing logistics, controlling supply. It works with over 8,000 independent farmers who supply it. Those farmers are good at what they do. But they are suppliers.

Participation as a supplier is not enough. Ownership of the platform — the farm, the infrastructure, the system — is where the compounding really begins.

Even “the number of trees” is the wrong KPI. It is the farm that matters.

Lesson 5 — A Store Is Better Than a Farm

This is the one that tends to land hardest in a room.

Even Dole — a billion-dollar company, one of the most recognized brands in its category — is a tiny supplier to the giant retailers that actually own the customer relationship. Walmart. Costco. Amazon. These are the stores. They do not grow bananas. They sell them — at scale, with leverage, owning the customer from end to end.

The store owns the customer. The store sets the rules. The store captures the value that flows through the entire chain. Needless to say, many of these stores are an order of magnitude more valuable than Dole.

This is the lesson that I applied at Wattpad. We were not just a reading platform. We owned the direct relationship with five million writers and one hundred million readers, with virtually no external dependencies. That end-to-end ownership is what made us defensible.

Amazon Kindle tried to kill us — not once, but multiple times. They launched product after product specifically designed to compete with Wattpad. Not only did we win every battle, we won the war. A clean sweep. It is rare for a company our size to take on Amazon Kindle directly and come out on top. Owning the full chain — the writers, the readers, and the relationship between them — is what made that possible. I will save the full story for another post.

True prosperity means owning the whole chain — from innovation to commercialization, from suppliers to customers. The number one KPI is ownership. Jobs follow capital, innovation, and commercialization. Not the other way around.

Why This Matters Beyond Business

I have shared this framework in many different rooms — with founders, with students, with engineering deans — and it lands every time. I think it is because the banana lessons are not really about business. They apply broadly to how we think about our lives.

Yes, the framework is useful in a business context. And yes, it is useful in an investment context. But it also applies to personal decisions, career choices, relationships, and how we spend our time. Are you grabbing the banana in front of you because it is comfortable and familiar? Or are you asking what the tree looks like? What the farm looks like? Who owns the store?

The mental model you build early about what success looks like shapes every decision that follows. Most people never stop to interrogate it.

Do not just grab the banana. Ask yourself what the tree looks like. Then ask what the farm looks like. Then ask who owns the store.

That is where the real value lives — in business, in investing, and in life.

If you missed the post this is a follow-up to, you can read Stop Supplying. Start Owning. here: [Link]

The post Five Banana Lessons appeared first on Allen's Thoughts....

Canada has a paradox that I have been talking about for a long time.

We are home to most of the world’s AI godfathers. We have Nobel Prize winners, world-class researchers, and some of the most respected engineering schools on the planet. Those institutions attract exceptional students from around the world. And in turn, Canada trains some of the best engineering talent anywhere.

Yet an increasing number of those graduates — now approaching 80% — leave after convocation.

With them go the startups they might have built, the economic value they might have created, and the wealth that could have stayed here. We all know the largest and most valuable companies are technology companies, and they are based in the United States. Many of them were co-founded or led by Canadians.

It is a story of could have been, would have been, and should have been.

I was recently invited to speak at the Deans of Engineering Conference in Winnipeg. I want to share the core of what I said, because I think this conversation matters well beyond that room.

The root cause is not what most people think

The easy explanations are talent, capital, and policy. We hear them constantly. They are not wrong, but they are not the root cause.

The root cause is a mindset.

Clearly, we do not have a shortage of ambition or ability. One fundamental issue stands out — yet very few people talk about it, let alone address it. We have collectively learned to think like suppliers, not owners.

Without the owner mindset, we unintentionally and subconsciously optimize for producing great talent for other countries, rather than building a stake in where that talent goes and what it creates.

This is the supplier mindset made visible.

Let me illustrate with a banana

I use a banana analogy in my talks because it makes the concept concrete.

If you put money and a banana in front of a monkey, the monkey takes the banana. It does not know that money can buy many bananas. That insight comes from Jack Ma, who used this analogy to compare the mindset of someone who grabs a job versus someone who builds a company.

Jobs are bananas. They are real, they matter, and they feed people. A banana feeds you once. Cash feeds you many times.

But the lessons go further.

A banana tree is better than a banana because it can feed you forever. Ownership compounds; wages do not. A banana farm is better than a tree — because participation as a supplier is not enough; ownership of the platform is the real prize. And a store is better than a farm, because the store owns the customer relationship and captures the value that flows through the entire chain.

This is the progression from employee thinking to owner thinking. From banana to store. From grabbing to owning.

Canada has been grabbing bananas

Let me make this concrete with two examples — one personal, one national.

The Wattpad calculation. When Wattpad was acquired for US$660 million, the headline was a Canadian success story. And in many ways, it was. But here is the number nobody talks about. By the time of the acquisition, roughly half of the company was owned by Canadians. When the deal closed, about US$330 million in economic value left the country — because we had raised capital from outside Canada to build it.

Wattpad’s annual payroll was roughly US$30 million. Not small. But compared to the acquisition price, it is a fraction. Ownership creates far more value than employment. Jobs matter. Entrepreneurship matters. But nothing compares to owning world-class companies.

The auto industry analogy. Many people say Canada has a strong auto industry. We do not. We have a strong auto supplier industry. That is not the same thing. Our auto suppliers — collectively — are worth a fraction of GM, Ford, or Toyota. They build the factories, employ the workers, and take on the operational risk. When the EV transition stalled, the suppliers’ brand new facilities went quiet. When the majors slowed production, the layoffs rippled through.

The supplier bears the downside. The owner captures the upside and sets the rules.

And when the Canadian government went to attract EV investment, what did we do? We signed deals to become suppliers again — subsidized by Canadian taxpayers, while the ownership, brand, and margin stayed elsewhere. We took the risk but not the profit.

This is the supplier mindset at a national scale.

The question nobody asks clearly enough

Here is the hinge question: what does winning actually look like?

The supplier mindset and the owner mindset do not just lead to different outcomes. They lead to completely different definitions of winning.

If you are a supplier, sending your best researchers to OpenAI is a win. You produced world-class talent. Mission accomplished. That belongs in the annual report.

If you are an owner, that is a loss. You invested in that person for years, and you ended up owning nothing. The outcome looks identical from the outside — a brilliant Canadian thriving on the world stage — but the two mindsets score it completely differently.

Until we agree on what winning actually means, we will keep celebrating losses as victories.

Where the mindset gets formed

Here is what I have come to believe: the supplier mindset is not learned on the job. It is learned in school.

The mental model a student builds about what success looks like — a FAANG job offer, a US grad school acceptance, a signing bonus from a company they can brag about — is set before they ever enter the workforce.

In the US, building a unicorn startup is Plan A. Getting a job at Google is Plan B. In Canada, getting a job at Google — or going to the US — is Plan A. Building a startup, let alone a unicorn, is often not even in the equation.

That quote is from a world-class Canadian AI scientist who is now at a US company. It landed hard when I first heard it, because it is accurate.

The deans are the front line

I said something direct to the room in Winnipeg that I want to say here too.

The founders of most of Canada’s future tech giants are sitting in engineering classrooms right now. The deans who lead those schools are the single most underleveraged force in Canada’s innovation economy.

Here is why this is also in the deans’ own interest. If their students build world-class companies and keep them here, those companies will forever be associated with that school. That is a legacy that compounds for decades. The next crop of students is inspired by the tech giants that exist. Right now, leading universities outside of Canada are winning that recruitment battle — not because their engineering programs are better, but because the companies their graduates built are more visible, more celebrated, and more aspirational.

Think about OpenAI. It was co-founded by a University of Toronto alumnus. Most people associate it with Silicon Valley.

That association is not fixed. It is a choice, made one graduating class at a time.

This is not only about encouraging entrepreneurship

I want to be precise here, because there is a version of this argument that deans hear all the time and that I think misses the point.

Many engineering schools already encourage entrepreneurship. Hackathons. Incubators. Pitch competitions. These are necessary. But they do not define what success looks like. And in a strange way, encouraging entrepreneurship is still a supplier mindset — we are producing entrepreneurs for the ecosystem and hoping something sticks.

The real call to action is different. It is to start and scale world-class companies here in Canada.

That is a higher bar. A different ambition. A fundamentally different culture to build. It means celebrating the founder who builds a billion-dollar Canadian company with the same institutional pride as the researcher who wins a Nobel Prize. It means changing what the school defines as a win — not only placements, publications, and patents, but also companies that stay, scale, and own their category.

The window is now

I have saved the most important point for last.

We lost Game 1. Canada invented modern AI. The most important AI companies are almost all based in the US. That window has closed.

But Game 2 is underway. Quantum computing. Robotics. Physical AI. Space. Advanced manufacturing. Smart energy. Just to name a few. Canada has deep roots in all of these — world-class labs, exceptional researchers, and early-stage companies that are genuinely competitive.

Here is what is different about Game 2: you cannot pack up a quantum computing facility or a robotics lab and move it to San Francisco. Unlike software, the physical infrastructure is sticky. The talent clusters around it. The companies that emerge will be rooted where the labs are.

And the ground-level signal I am seeing is genuinely encouraging. I have never met more professors and researchers who want to start companies — and who want to do it in Canada. That is new. That is meaningful.

The conditions are finally aligned to address the root problem, not just the symptoms. But the only trophy that ultimately matters is homegrown, world-class companies. And we can only win Game 2 — and ultimately the championship — if we build the owner mindset now, starting with the people who shape how the next generation of engineers think about what success looks like.

Addressing the supplier mindset and turning it into an owner mindset can create the domino effect that turns Canada’s bragging rights into lasting economic wins.

That is the game we can win.

The post Stop Supplying. Start Owning. appeared first on Allen's Thoughts....

We are super excited to share our investment in Tiptree Systems!

Founded in Montreal by two Mila AI researchers, Dr. Martin Weiss and Nasim Rahaman, Tiptree is building an AI-native researcher and knowledge network: infrastructure for how AI researchers share, exchange, and advance knowledge.

There is an irony here. AI has advanced remarkably quickly, yet until now, there has not really been an AI-native way for AI researchers themselves to share knowledge and build on one another’s work. Tiptree is solving that problem.

This is exactly the kind of company that fits our thesis at Two Small Fish. We invest in technologies that can reshape large-scale behaviour, enabled by foundational shifts in computing. The collapsing cost of intelligence is changing not only how software is built but also how work gets done. It is also changing how knowledge can be organized, explored, and shared.

That is what makes Tiptree so interesting to us.

My partner, Brandon, himself an AI scientist, and I wrote more about why we invested. Here is our blog post on Two Small Fish’s website.

Today, Apple turns 50.

I cannot tell you how much my first computer, an Apple II, impacted me.

It was my first computer. It had a green monochrome monitor. It had 16KB of memory, which is roughly 2 million times less than the 32GB MacBook I am using to write this post. By any modern standard, it was primitive. To boot it, I had to use a cassette tape. Yes, cassette tape!

Thankfully, it had a few expansion slots, so the machine was quickly upgraded to 64KB, with a floppy drive and later two of them to facilitate copying disks, plus a sound card to make some noise. Still no hard drive, of course. That would have been too expensive to ask my parents to buy for me.

What made that machine special was not the hardware. It was what it unlocked for me.

For the first time, I could be creative through writing software. I started with the programming language BASIC, which felt magical back then. Occasionally, I had to write in assembly language, which meant dealing with 0s and 1s directly and pulling hair along the way. It was clunky. It was frustrating. It was also incredibly fun and deeply fulfilling when the program finally worked!

So I wrote games and other fun programs to keep myself entertained. Looking back, that was the real gift. My late father did not just buy me the best toy I could have imagined. He gave me a machine that allowed me to become a permissionless innovator, a tinkerer, a self-taught learner, and someone who learned early to think differently.

That mindset stayed with me.

It shaped how I see the world as an engineer, as a founder, and now as an investor. Long before I had the language for it, the Apple II taught me that technology is at its best when it gives people leverage, creativity, and the freedom to build.

Apple’s anniversary post mentions the Apple II among the products that helped define the company’s first 50 years. For me, it did something even more personal. It helped define mine.

Maybe that is where it all began. Thinking differently is not just about being creative. It is about seeing what others do not see yet. In many ways, that idea has stayed with me ever since and still shapes how I think about technology, innovation, and investing today.

Looking back, I think learning to think differently was where the idea that seeing the future is our superpower first took shape.

And thank you to one little Apple II with 16KB of memory, a cassette tape, and just enough magic to change a kid’s life.

The post Happy 50th Birthday, Apple! appeared first on Allen's Thoughts....

Every VC says they back high-quality companies.

That is like saying humans need to sleep, eat, and drink. True, yes. Useful, no.

The more important question is: what does “high quality” actually mean in venture?

For us, it means the potential to become uniquely world class.

A company that can become the clear winner in an important category. A company with real moats. A company that can build something enormous.

This is why we spend so much time trying to understand what is truly unique about a company. Not what is interesting. Not what demos well. Not what sounds differentiated in a pitch deck. Not what the ARR is today. What is actually hard to replicate? What gets stronger over time? What creates a widening gap versus everyone else? And the only way to know is to spend time with these deep tech founders and really understand how the technology, the product, and the company work.

I have written a long blog post on this topic on Two Small Fish’s website. Here is the link.

The post Uniquely World Class appeared first on Allen's Thoughts....

Any software engineer who has built scalable systems, myself included, knows the special kind of hair-pulling pain that comes with logs.

Logs are supposed to tell you what happened. In practice, they often bury the answer under a mountain of machine-generated noise. When a system breaks, slows down, or starts behaving in a way that makes no earthly sense, you go to the logs. Then you search. Then you search again. Then you question your life choices. I have spent enough time with log4j that I will never get back, and I suspect I am not alone. Just ask Brandon, who had to deal with logging every single tap, swipe, and click from Wattpad‘s 100 million monthly users. That could easily add up to billions of log entries a day!

That is why YScope clicked for the TSF team right away.

YScope was cofounded by University of Toronto Professor Ding Yuan, Professor Michael Stumm, Dr. Kirk Rodrigues, Dr. David Lion, Jack Luo, and Beverly Xu. The company is building an open-source platform for log storage, search, and analytics, designed for humans and AI, at the edge and in the cloud, that delivers extreme efficiency. At its core is CLP, a compressed log processor that delivers a real step change in both performance and efficiency. In a category where teams are often forced to choose among speed, storage cost, and usability, YScope is advancing all three at once.

The timing could not be better. We are moving into a world where intelligence is becoming abundant. As the cost of intelligence collapses, more and more systems will be making more and more decisions every second. A robotic finger moves. A self-driving car makes a slight turn. An AI agent calls a tool, retries a task, or hands control back to a human. Each of those actions creates an event. Each event becomes a log record. Multiply that across robots, vehicles, industrial systems, edge devices, and software agents operating continuously across edge and cloud environments, and you do not get a modest increase in logs. You get an explosion.

That explosion creates a massive opportunity for the company that can handle this new reality better than anyone else. Better compression means lower storage and transmission costs. Faster search means faster debugging, faster investigations, and faster feedback loops for both engineers and AI systems. This matters especially at the edge, where bandwidth is limited, latency matters, and shipping everything back to the cloud is often expensive or impractical. When you can compress logs aggressively, keep them searchable, and analyze them efficiently across environments, you are not just improving an existing workflow. You are changing the economics of the entire stack.

More importantly, CLP provides the ideal logging infrastructure for the next frontier of the agentic AI and robotics era. Logs contain massive volumes of noisy and redundant information. CLP turns unstructured logs into structured data, which significantly improves an LLM’s ability to understand what is happening in the system while reducing noise and the data footprint. This structure also enables much faster queries, allowing efficient exploration of large volumes of logs and the quick identification of the signals that matter.

That is what excites us about YScope. This is not just a clever research project with a great demo. It is infrastructure for a world that will produce exponentially more machine-generated telemetry. As intelligence becomes cheaper and more abundant, the infrastructure required to store, search, and analyze the resulting torrent of logs becomes increasingly strategic. CLP has already demonstrated this impact at scale, powering Uber’s production logging platform, operating across more than 1.5 million connected electric vehicles and being used by other Fortune 500 companies.

At Two Small Fish Ventures, we invest in the next frontier of computing and its applications. We focus on deep technologies where foundational breakthroughs create the conditions for new category-defining companies. YScope fits squarely within that thesis. It sits at the intersection of AI infrastructure, edge computing, and the coming proliferation of physical AI and autonomous systems. Seeing the future is our superpower, and we believe YScope is building an important piece of that future.

TSF is proud to lead YScope’s US$3.9 million financing, with UTEST, Snow Angels (the Snowflake alumni investment syndicate), Next Wave NYC, and other successful founders participating. Congratulations to Professor Ding Yuan and the entire YScope team. We are thrilled to be part of the journey!

P.S. Read the announcement from the YScope team here.

OpenClaw has taken the world by storm, but for many users, the practical issues are clear: it is difficult to install, it can create a real security risk if not used properly, and it can become expensive quickly.

Crate from our portfolio company Gensee addresses those issues directly:

1. One-click install in 60 seconds

2. A secure sandbox that only accesses what users explicitly allow, built on deep expertise from Dr. Shengqi Zhu and award-winning operating systems expert Professor Yiying Zhang, whose work on agentic optimization and efficiency makes this possible

3. Completely free to use, enabled by that same focus on performance and efficiency

Crate also offers several important advantages beyond setup, security, and cost. It includes Gensee’s proprietary AI search engine, ranked #1 on Source Bench for finding the highest-quality web sources. It comes pre-installed with common, useful skills vetted by the Gensee team for safety, while still allowing users to install more skills themselves. And it gives users flexible control through features such as multiple instances, pause and resume, snapshots, and rollback.

That is why I am using OpenClaw myself through Gensee Crate exclusively.

For a deeper look, I wrote a longer post on Allen’s Thoughts. You can also try Gensee Crate here.

Originally published at

https://twosmallfish.vc

on March 12, 2026.