The world is changing incredibly fast. Public software company valuations are falling through the floor. Startup financing is either untethered from reality or suddenly much harder to come by. At the same time, the cost of building software is approaching zero.

Layer on top of that the most powerful technology humans have ever invented improving at a rate that is hard to fathom. Uncertainty is everywhere.

When the present feels chaotic, it’s often useful to look backwards.

During the Industrial Revolution, machines like the Spinning Jenny and the Power Loom wiped out entire categories of work. Hand weavers, cottage textile workers, and skilled artisans who produced cloth manually suddenly found themselves competing against machines that could produce fabric dramatically faster and cheaper. Workers famously protested, most notably the Luddites, fearing permanent unemployment. The language they used sounds strikingly similar to what we hear today about AI creating a “permanent underclass.”

But the net effect of the Industrial Revolution was the opposite of what many feared. Textile production exploded. Clothing became dramatically cheaper. And total employment in textile manufacturing actually increased, just in different roles.

We saw the same pattern repeat itself in later waves of innovation. The automobile displaced the horse economy. ATMs were supposed to eliminate bank tellers. Personal computers replaced typing pools. Each wave destroyed certain jobs while creating entirely new industries on the other side.

Today we find ourselves in another period of real uncertainty. The macroeconomic picture is shaky and the technology curve is steep. We are already seeing early signs of disruption in the labor market, particularly for early-career professionals.

The promise of AI is extraordinary. The fears are well-founded.

When blog posts written by people nobody has heard of go viral and move public markets, it’s a good signal that people are afraid. Fear tends to dominate the narrative during moments like this.

What many of these conversations miss is that a large number of so-called “software companies,” particularly those selling into compliance-heavy or slow-moving industries, are not really software businesses in the traditional sense. They are compliance businesses, distribution businesses, and trust businesses. The software itself is often the easiest part of what they have built. That reality gives many of these companies more durability than the market currently assumes.

At the same time, the barriers to building new software have collapsed. This wave of technology will allow new entrants to expand into areas that have historically been dominated by legacy incumbents. Terrible, decades-old systems of record may finally begin to break in some industries. Inside those systems, much more work will become automated. Interfaces will simplify, often reverting to chat or voice, and long, multi-step workflows will collapse into a single request.

The economics of software companies may change as well. More companies will become successful with far fewer employees. The era of software businesses hemorrhaging money indefinitely appears to be ending. Increasingly, these companies will be judged the way most businesses always have been: on their ability to generate real profits.

Alongside that shift, we will almost certainly see entirely new categories of work emerge. New job types will appear, new industries will form, and more individuals will build meaningful businesses on their own. At the same time, some roles we once assumed would always exist will disappear.

The biggest risk I see in the near term is a training gap. Entry-level work has historically been how people learn an industry. It is how analysts become operators and how junior employees become leaders. But many of those early tasks are exactly the kind of work that AI is best at automating.

If those roles disappear too quickly, companies may unintentionally stop training the next generation of talent. Fast forward five or ten years and you could see a strange organizational problem emerge. Senior leaders retire. Middle management moves up. And suddenly there is a hole in the middle of companies because nobody ever learned the fundamentals that those entry-level jobs once taught.

That could become a real structural issue for many organizations. It also means a generation of young adults may have a much harder time entering the workforce than those before them.

Wealth inequality may widen in the medium term as well. There is not an obvious way around that outcome while productivity gains are being captured primarily by those who own technology and capital.

And yet, despite all of this, I remain optimistic.

Humans are remarkably creative. Much of the energy we spend worrying about relatively trivial problems is a sign of how comfortable life has become. The average person in the developed world is not thinking about where their next meal will come from.

Given enough time, and often a bit of pressure, people invent entirely new forms of work. They start companies, create industries, and solve problems that were previously invisible. Every major technological shift in history has looked destabilizing while it was unfolding.

But over time, people have always adapted.

They have always figured it out.

My bet is that this time will be no different.

That’s some crazy whiplash. And if you’re paying attention, you’ll see that these are trends that have been years in the making. It makes you wonder what’s coming next.

Most of the AI labs are my customers, along with some of the largest companies deploying AI at scale. I live in San Francisco and can’t go to my local coffee shop or grocery store without seeing another friend who’s building an at-scale venture-backed AI company. And I spend way too much time on Twitter.

So what happens next? Is your company going to survive this? Are the layoffs just starting? How fast is this actually moving?

Bottom line up front: I believe full diffusion of AGI is about one decade away, and things will change at an accelerating rate alongside those 10 years, with the world on the other side of 2030 looking as different as it did pre/post 9/11, COVID, and World War 2.

To the AI people on Twitter, they’ll think 10 years is way too long. My average American probably thinks 10 years is too optimistic. Let’s dive in.

Diffusion is not cope

My recent number one source for good AI content has been Dwarkesh Patel’s podcast. He hosts some amazing guests and asks good questions. It’s like Joe Rogan for AI but less muscle mass and more brain power.

On a recent Dwarkesh podcast episode with Dario Amodei, Dario talked about how challenging it is to predict research progress, allocate resources, and balance research vs. deployment. He highlighted how beyond R&D, “diffusion,” the ability for us to put AI into the hands of individuals, remains a real challenge. Dwarkesh countered by saying “diffusion is cope,” that any sufficiently good AGI model would be able to be used reliably by any human to do any task.

I tend to agree with Dario. Even if the models can become good enough within a couple years (and I think they will!), it’s going to take a serious amount of human work to get to ubiquitous distribution. There are three reasons for this.

The models still aren’t good enough yet. The labs’ job isn’t done. It continues to get easier over time, but there are still global-scale logistics problems to solve.

The models just barely started being able to draw silly pictures reliably.Their coding has gotten REALLY good, but not good enough. I do believe that OpenAI, Anthropic, and DeepMind are working on this with good intentions and the risks in mind, but there’s real distance left to cover.

People and organizations resist change. Getting the world to adopt a completely different way of living and working will require a massive amount of education. We’ve made significant progress on the digital transformation of our society, but even that work isn’t complete. There are many systems that you cannot completely self-service digitally without having to escalate to a human. Voice AI will help with that. If you give an agent a credit card, web browser, email, and phone number, you can get a lot done. But it cannot open a bank account. Not yet, at least.

Physical infrastructure can’t keep up. Beyond the organizational challenge, you need to serve all these models (often called inference). You need GPUs. So many GPUs.

GPUs are really only one part of this. You also need data centers, and then you need chips, and then you need power. Power is the hardest one right now, and there are several incredible companies focused on revolutionizing the way we structure, store, and handle power. I’m incredibly bullish on them.

How fast can you build a power plant? What about 1,000 of them? Even if we’re able to get the latest AI, we’ll need to FUEL it with power. AGI might arrive but be impossible to distribute globally because of power and GPU demand to actually fuel it.

The STRUCTURAL barriers of human society and the PHYSICAL barriers of our infrastructure to power AI will create a LIMITING FUNCTION on how soon AGI will fully diffuse into our world.

We may get AGI sooner, but it will be unfairly distributed to those with capital and connections.

What this means for software

OK, so we’re 10 years from full AGI diffusion. What happens in the meantime?

It’s going to be harder and harder for software companies to win by just existing. Disruption is going to be easier than ever and starting a competitor is going to be easier than ever. Market cap is no longer generated by feature differentiation. Ability to execute at scale, brand, and customer trust will matter most.

There’s a popular idea in tech called the bitter lesson, which states that no matter what you do, a future iteration of AI models will also be able to accomplish that same task. People apply this to software all the time, saying “it isn’t worth building that piece of software because the next generation of models will simply generate it in one shot.” There’s truth to that.

Software land has benefited from “having more features” being a viable way to increase market caps and beat competitors. We had so much excess in software that building a company that was completely open source was actually a viable strategy, because it was so challenging to run infrastructure at scale. That era is over.

Predominantly software companies are going to face a lot of pressure to improve their ability to execute and retain market cap in the short term. Organizational restructuring, layoffs, risky product bets.

Don’t believe me? Look at how much of the S&P 500 is constituted of software companies, then look at the cash balances of those companies and you’ll see that they’re ready to weather this storm.

None of this means jobs disappear, though. People think AI is going to kill employment wholesale, and the data says otherwise. Software engineer employment is actually increasing.

AI can make it easier to replicate what software does, AND demand for human work can keep growing because more capable people unlock more demand. Which jobs change and how fast is the real question.

So what companies are “safe”?

I believe there will be only four types of companies in the future:

AI labs are companies turning compute into knowledge, then delivering that knowledge. There will be many types of labs. Some will specialize. But their primary function is creating and generating intelligence.

Infrastructure companies give the labs more capabilities and create the rails for that knowledge to run on. Power, internet, water, chips, and digital infrastructure too, like connection to payment rails or digital services via a browser.

Systems of record are the incumbent places where all of our current knowledge lives, is organized, and is accessed. They’ve become these giant systems that are very hard to move away from because the world is so entrenched on them. I think those continue to exist.

Services means humans doing services for other humans, or people employing agents to do service on your behalf. McKinsey is a services business. So is a hair salon. There’s a wide range, and there will continue to be.

Your business either already is one of these four types, or it will need to become one.

So are we screwed?

No. The most important thing you can do right now is not overreact, but certainly react. Think critically about how to start understanding these new types of technology and systems.

People can draw a very crazy line to a world that’s completely utopian or dystopian, but I think that’s unrealistic. It’s almost like the meme: nothing ever happens. The truth is somewhere in the middle. Things are going to change, and there will be turbulence, but it will only be truly impactful if you aren’t staying in the loop, understanding things, and growing yourself. You definitely cannot continue forward assuming everything’s going to be the same.

I really do liken it to COVID or other major changes in history. If we ignore what’s going to happen, it’s going to hurt. But actually taking an optimistic view, leaning into this future, being a part of it, being involved with it, I think that’s going to be a really huge advantage.

Regardless, the world will change. I guess I’ll need to take another vacation to see what happens

You receive a data room: 60 gigabytes. 45,000 documents. ~600,000 pages.
You’re told to enter 75 data points per unit, reconcile conflicts across documents and financial statements, and finish before the deal clock runs out.

Forty-five minutes later, you’ve completed your first row! There are 999 to go.

At the end of the week, after hundreds of hours across multiple analysts, someone asks a simple question: Where did that number come from?

You don’t know. You search. Eight files look plausible. Two hours later you find the answer and realize you used the wrong value in several places. The downstream models are now wrong. The cycle repeats.

This is not a failure of effort or intelligence. This is simply how multifamily diligence works today. Even at the most sophisticated institutions in the world.

The Real Problem: A Fundamentally Broken Data Environment

Multifamily diligence is uniquely painful not because the industry is unsophisticated, but because the underlying data environment is hostile to analysis.

1. The data is unstructured, inconsistent, and fragmented

Every party, buyers, sellers, lenders, brokers, operates from a different partial view of reality. Information lives across document repositories, PDFs, scans, property systems, spreadsheets, and email threads. Slightly different versions of the “same” document exist everywhere. Nothing reconciles cleanly.

2. The documents were never designed for computation

Leases, amendments, vendor contracts, estoppels, and financials were written for legal and compliance purposes, not analysis. They are long, exception-heavy, inconsistently formatted, and often scanned. For decades, humans were the only viable parser.

3. Accuracy matters more than the tools allow

Small errors compound into massive downstream risk. Yet most diligence relies on sampling, because full coverage is too slow and too expensive for humans alone. Spot checks become norms.

4. The process depends on tribal knowledge

Every firm runs diligence differently. Templates, formulas, heuristics, and expertise live in people’s heads. When someone leaves, goes on vacation, or gets overloaded, quality drops. There is no system of record. No memory. No automatic validation.

5. Time pressure forces bad tradeoffs

Deals move fast. Deadlines are fixed. Teams are stretched thin. Accuracy is supposed to be non-negotiable, yet teams routinely choose between “good enough” and “we don’t have time.”

This is why diligence remains slow, fragile, and error-prone, even for institutional investors managing billions. The process persists not because it’s good, but because there was no alternative.

Why This Can Finally Be Automated

The reason multifamily diligence hasn’t been automated before is simple: the technology wasn’t ready. Today, it is.

1. Extraction quality finally crossed the trust threshold

Modern multimodal models can reliably extract structured data from the messiest of files with high accuracy. Full-coverage extraction is now more accurate than human sampling, and discrepancies surface immediately.

2. Document processing is no longer a bottleneck

OCR, parsing, version comparison, and reconciliation now run in minutes, not days, across massive data rooms. Parallel processing makes exhaustive audits viable for the first time.

3. Reasoning is the real unlock

We can now:

• Cross-check values between documents

• Resolve conflicts logically

• Identify missing or inconsistent information

• Validate rent rolls against leases

• Interpret exceptions and conditional clauses

This is true skilled labor automation.

4. The ROI is immediate and undeniable

Teams are overwhelmed. Outsourcing leads to untrustworthy, inconsistent results. Labor is scarce. Errors are expensive. Automation now delivers what executives actually care about:

• 95%+ time reduction

• Full-coverage

• Traceability for compliance and investors

• Faster closings

AI isn’t a novelty here. It’s the only scalable answer.

Enter Rely

Rely ingests entire data rooms and runs full-coverage audits across leases, rent rolls, contracts, financials, and more. Every output is traced directly back to its source, creating a single system to understand, validate, and manage the entire transaction lifecycle.

We are not building another underwriting tool. We are replacing a massive, manual human process that underpins how commercial real estate actually operates. Here is what our customers say.

Where We Are Today

Rely is live, actively used by some of the most sophisticated operators in the industry, and backed by partners who understand both the scale of the problem and the importance of getting this right. Operators like Cardinal Group Companies and Elm Grove, experienced executives like Howard Smith and Chase Gilbert as well as investors like Better Tomorrow Ventures.

The interest we’ve seen so far has been overwhelming and it’s coming from teams who feel this pain every day.

What We’re Looking For

We’re continuing to learn from the best operators, investors, and diligence teams in the market. If you’ve experienced this pain or have strong views on how it should work, we want to hear from you.

This is the beginning of a fundamental shift in how commercial real estate transactions get done.

Hawk Hill Ventures has been active too! We made 20 angel investments across infrastructure, dev tools, vertical SaaS and even a PE-leaning veterinary business and looked at hundreds more.

Below are our predictions for 2026.

George’s 2026 Predictions

2026’s Breakout Companies Automate Entire Jobs, Not Tasks

For years, software promised leverage but delivered dependency: more tools, more steps, more dashboards. In 2026, the winners flip that relationship. The most important startups will be the ones that collapse full job descriptions into automated systems, taking work that previously required teams of analysts, coordinators, auditors, or operators and executing it end-to-end.

Prediction: We will see a higher percentage of 2026 winners that achieve this outcome. Fewer humans will do the same work at higher quality, which finally changes the economic shape of whole industries.

A High-Profile AI-Related Bankruptcy

The last two years were marked by an unshakeable belief that AI rising tides lift all boats. In 2026, one of the boats will sink and loudly.

Some companies built their valuations on access to models, not durable distribution, data, or workflow ownership. Others built headcount-heavy pseudo-services businesses packaged as AI. And a handful raised money at valuations the underlying business simply will never support once capital markets tighten.

Prediction: A major AI darling, high burn, thin real-world margins collapses in a very public way. The shift will force investors to start looking at AI businesses for sustainability which will be healthy for the category.

SBF Gets Out of Jail

I just feel like this is going to happen.

Paul’s 2026 Predictions

AI models will be able to work for a week straight

If you’ve been paying attention, AI capabilities have kept improving as models get more sophisticated. One way to measure this is task duration. The METR team benchmarks how long models can work on complex tasks, measured in hours. This has grown exponentially over the past several years. I think model capabilities will continue to improve in 2026, and by year end, the trend in the graph above will continue. Models will complete tasks spanning multiple days, up to a full week.

2026 will have the most tech exits of the past 10 years

Valuations are elevated and there are clear signs of market froth, but the party isn’t over. Real businesses are being built, and the need for capital will outstrip what private market investors can provide. For the largest players building the most important technology in the world, increased regulatory scrutiny may demand they go public to improve transparency and corporate governance. This seems inevitable. The backlog of giant companies that held back on IPOs due to regulatory concerns or no need for capital will fade as the biggest companies keep winning. And smaller companies will get acquired faster as the pressure to move quickly and bring in technology has never been higher.

Everything will be ok

If you work in AI, follow the industry closely, or have read enough sci-fi, these improved capabilities can feel scary. You might worry your job will be automated. “What’s my value if not my ability to make spreadsheets?”

I have a different view.

Working with our customers at Browserbase, I keep seeing the same thing: the bottleneck to AI adoption isn’t model quality anymore. It’s the ability for humans to teach others about complex problems and evangelize how we solve them with technology. It’s a human problem holding back AI adoption, not a capabilities problem.

The singularity isn’t coming tomorrow to take all our jobs. It’s going to be gradual. And I’m confident that in 2026, everything’s going to be ok. Quality of life will improve for many people. Those willing to accept and understand this new technology will have an edge. The only people who won’t be ok are the ones who reject this technological revolution. Burying your head in the sand won’t avoid the inevitable. The way we do everything is about to change, hopefully for the better.

Onwards to 2026

We really appreciate everyone who follows our tiny newsletter. Selfishly, we mostly write for ourselves. Putting ideas on paper helps us think and push our companies forward. Thank you for continuing to read, question, comment, and share our thoughts. We always appreciate your engagement!

We hope you have a great holiday season, and we’re looking forward to continuing to be in your inbox every month in 2026.

That was a tough realization. You have an incredible number of existential thoughts: “I can’t lose everyone’s money. We can figure something out, right? I am not willing to be a failure. Am I a failure? Only if I give up right?...” The dialogue is wild because that same day you can feel like a genius for any other reason.

We weren’t killing our product because we were giving up, we were killing it because our initial hypothesis was fundamentally flawed and we had enough data to know why. It was a nice to have, very hard to quantify a return on and nearly impossible to sell.

Most importantly, we wanted to solve the most pressing problem possible for our customers and this wasn’t it…we were actually shown something much more compelling in our discovery process. That process is the beauty of starting a business.

You come out of the gates with a view on the world. A view on your market, the product that must exist and who you are going to serve. But then, you start to spend time with your customer. And only after you’ve spent thousands of hours with them can you truly understand what they really care about. Empathy doesn’t come from moonlighting, it comes from living and breathing a problem space and an industry. THAT is how we found our north star at Rely.

I’ve been approached by a number of other founders who all need the same advice on pivoting, staying in the game and just building resilience into their DNA. I am by no means an expert because this is indeed my first rodeo. Below are just some lessons I have picked up along the way.

Every no is one step closer to your yes. It is hard to describe just how much you will get rejected as a founder. Investors, customers, early employees…often you’re lucky if you get one single yes per day at the beginning. The no’s aren’t a problem but how you respond to them really matters. Use it as fuel, as education and as guidance. From there you can start to adjust your heading.

Be shameless. I reached out to a high school ex-girlfriend for an introduction to a large real estate acquirer. It didn’t work but I tried! Ultimately, if you don’t ask, you don’t receive so just keep putting yourself in the position to experience luck.

The people who succeed just stay in the game the longest. Behave like it. This is a marathon and it may not even be your only start up. Don’t be discouraged by the little things and more importantly, don’t burn bridges and act in a short-sighted fashion. Structure offers and deals in a way you’d want to receive them, work with good people and good people only, build a tribe, treat people well…if you make decisions by trying to get someone to join you on your next mission, you’ll treat everyone better. This isn’t a zero sum game. So, to simplify this, be a good person.

Taking other people’s money is a great forcing function…its way easier to stomach losing your own money. Oh and send them updates. They care! I’ve found this pressure to be healthy. It motivates me to want to always excite them. Investor updates are a great way to continue to foster the community around you and leverage your investor base when you need them. It creates accountability as well. These people bought part of your company so they’re going to want to help. Said a different way, you let them buy some of your equity, why wouldn’t you at least try to get something out of them. Lastly, share the bad news too. Just be honest.

Your attitude is contagious…and if you’re not fooling yourself, you’re not fooling anybody. If you feel defeated, you are defeated and everyone else can tell. You aren’t going to be able to convince anyone else to support you, join you, buy your product if you yourself don’t believe in it anymore. That mentality will crush you until you figure out how to fix it.

Talk to literally everyone. You have no idea where your “a-ha” moment will come from so in discovery, cast a very wide net, be genuinely curious and ask each person you meet who you should talk to next.

If or when the time comes to put your first idea down, do it with pride. This moment is a hard realization but its so freeing. You probably knew it weeks or months before you decided it was time. So be free, move on and move on with confidence knowing you’re a bit closer to the right path than you were before.

I’m sure there are a bunch of other lessons I missed but ultimately, you get in this game to build something of value. That thing of value has to be informed by the people you serve and if your first thing doesn’t work, then just figure out what does. It isn’t embarrassing to change direction, it is embarrassing to receive feedback and ignore it.

Hopefully this is helpful to people going through this journey as well. TLDR: be a good person, listen to others, stay in the game.

George Callaway Matelich was born on October 22nd, 2025. I can’t wait to read his memos one of these days. Congratulations George, Anne and their families!

Two years ago, I wrote An Internet Browser for AI. I didn’t know it at the time, but that memo would change my life. It became the foundation for Browserbase and set me on a path that still feels both inevitable and unpredictable.

When I wrote it, I remember thinking how neatly my background came together for this moment. I started in infrastructure as an engineer at Twilio, learning how large-scale systems actually run. At Stream Club, I was a founder working directly with headless browsers, building products on top of them and understanding their quirks. Then at Mux, I led teams, worked on developer tools, and spent time on developer marketing, learning how to connect infrastructure products with their users. Those three experiences (engineering, founding, and communicating) fit together in a way that made this company feel like the obvious next step in hindsight.

When I finished the memo, I sent it to a few people whose opinions I trusted most. They’d seen earlier drafts of my thinking before, but this time the reaction was different. Everyone agreed that the idea was strong and original. Some people even wanted to invest. That was when I realized the memo wasn’t just a piece of writing, it could actually become a company.

Predictions and Risks

Looking back, most of the predictions in that first memo proved right.

I wrote that AI would need to use the internet the same way people do, and that browsers would become one of the core tools for it. That’s true today. Web agents exist in many different forms and already power a wide range of applications.

I said scraping and browser automation would remain essential even as new access methods appeared. That’s still the case. Even with the rise of MCP, there is no escaping the need to interact directly with websites.

I argued that existing tools were suboptimal, and I still believe that. Playwright has improved a lot and deserves credit, but Stagehand continues to be faster and more token efficient.

I also believed that open source, community, brand, and a bottoms-up strategy were the right way to build in this market. Two years later, that’s worked. Stagehand’s adoption has grown faster than I expected, and our brand has become one of Browserbase’s biggest strengths.

The broader category has expanded dramatically. Playwright’s installs have multiplied since I wrote the memo, and browser automation has become a standard developer primitive.

Some of the risks I listed turned out to be smaller than expected. Models improved at an unbelievable rate. Inference costs dropped steadily. Legal and ethical fears that could have slowed the space haven’t materialized. If anything, the tone of the conversation around AI has become more constructive and forward-looking.

The core thesis, that browsers would remain essential, has only strengthened. The internet is not being rebuilt for AI anytime soon. The long tail of websites still needs to be read, clicked, and filled out, and that means there will always be a need for a browser layer built for automation.

MCP and the Long Tail of the Web

When I wrote the original memo, I didn’t predict that MCP would exist, let alone that it would take off. I believed AI would need to use websites and tools, and that websites would make up part of that toolset. What I didn’t expect was that there would be a standardized protocol for connecting AI systems to tools at all. That part surprised me.

MCP has turned out to be a clever idea. It gives developers and agents a shared language for connecting to external capabilities. That’s a big step forward for interoperability. But it doesn’t replace the web. If APIs didn’t eliminate the need for browsers, MCP won’t either. It’s mainly a wrapper around APIs and serves a complementary purpose.

Most of the internet will never have an MCP server. The same organizations that haven’t built APIs are unlikely to build these either. That long tail of messy and outdated websites is where AI still needs to operate. That’s the part of the internet that Browserbase is built for.

Customers and Market Reality

Our customer base has evolved alongside the category’s maturity.

In the early days, most of our users were bleeding-edge startups. They were small, fast, and experimenting with AI in ways that others hadn’t yet imagined. They helped us learn faster than any large enterprise could.

Those early builders pushed the limits of the models and created the patterns that defined how people use browser automation today. Features like utilizing accessibility trees and the syntax that made Stagehand approachable came directly from watching how those customers worked.

That created a product-adoption flywheel that still drives us. Bleeding-edge customers had new problems, which pushed our roadmap forward. Solving those problems made our product better than anything else on the market. That, in turn, attracted more customers with new and harder problems. The cycle kept repeating, and the product kept improving.

Things began to change after OpenAI launched Operator in early 2025. That was the moment people outside our corner of the world realized that AI could actually use the browser. Suddenly, the concept clicked. From that point, we started moving upmarket—mid-market and enterprise customers, including some very recognizable names, began adopting Browserbase.

I still love working with the early builders, but now there’s a new kind of energy coming from the larger companies figuring out how to deploy this technology at scale.

The Next 24 Months

The next big shift will come when browser agents stop being treated as toys that only do simple things like automating LinkedIn. When they start performing complex tasks reliably, people will realize how capable they are.

I think browser agents will eventually feel as powerful as humanoid robots. Today, people look at them the way they look at Roombas: limited, clunky, useful in narrow ways. Over the next couple of years, that perception will change. These systems will start to feel extremely capable.

We’ll know the inflection has arrived when people stop talking about “browser agents” as something new and start treating them as a normal part of software development. It will feel obvious, the same way AI coding assistants feel now.

It’s hard to predict exactly when that happens. Maybe 2026, maybe sooner. But I’m certain it’s on the way.

Reflections

I wrote the original memo sitting on my couch, watching football, blissfully unaware of how much life was about to change.

I’ve learned a lot in these past two years. Every Monday, I walk into the office wondering what fire I’ll have to put out that week. The thing you learn is that you always find a way through it.

If I could tell my 2023 self one thing before publishing that memo, it would be to trust my instinct. The idea was right, and the timing was even better than I realized.

Two years later, the browser isn’t just a tool for humans. It’s becoming a universal interface for intelligence itself. It connects the digital world in ways no other layer can. And we’re still just getting started.

We love to feel busy, well-informed, and thorough. But in reality, for many kinds of investments, that busywork has little impact on the ultimate outcome. The thing that separates winners from everyone else isn’t their spreadsheet. It’s how they see the world.

The Illusion of originality in the math

I talk to hundreds of institutional real estate buyers. Almost all of them lead with: “We have a proprietary underwriting model.” Then I see it…and it looks identical to everyone else’s.

Doing this work isn’t bad, it is really important. But it gets you to the starting line, not the finish line.

The danger is mistaking the model for the edge. You start to believe that because “multifamily performed poorly in 2023, all markets performed poorly.” In reality, Indianapolis and several other markets saw rents and property values rise in that same period.

The models rhyme. The differentiation is tiny. The edge comes from interpretation, not the cell references.

Real Estate: same spreadsheet, different bids

I’ve seen the same model lead to radically different decisions:

• Buyer A sees commodity apartments in a tired secondary market.

• Buyer B sees a secular shift toward Portland spillover and undervalued Southern Maine towns.

• Buyer C sees an inflation hedge and capital preservation.

• Buyer D sees a platform for operational excellence with tech-enabled NOI expansion.

Same math. Different worldview. Worldview drives the bid — and the return.

Tech parallels

• Airbnb — Math: TAM looked laughably small. Worldview: Trust + design could rewire travel.

• Stripe — Math: Crowded, low-margin. Worldview: Developers as the new distribution channel.

• Tesla — Math: EVs looked niche. Worldview: Cars could be electric and aspirational.

• Facebook — Math: Social networks are fads. Worldview: Real-identity networks compound.

• Amazon — Math: Unprofitable e-commerce. Worldview: Internet scale would crush margin fears.

• Shopify, MongoDB, ServiceTitan, Toast, Snowflake, Nubank — each “too small” or “too risky” on paper; each massive in reality because founders bet on a worldview, not consensus math.

Math is consensus. Worldview is non-consensus.

The spreadsheet makes you feel in control, but it rarely gives you an edge. Everyone can run the same LTV/CAC ratios, cap rates, or DCFs. What separates winners is what they believe when they look at those numbers.

In venture especially, market math is often obviously wrong, but it gives people an excuse to be lazy. Most VCs don’t beat the market for a reason. Doing the work to form a thesis that requires real change in the world is hard.

Closing

In the end, much more of investing, and succeeding in the world, is about gut and grit than we like to admit.

The spreadsheets give us comfort. They make us feel like we’ve turned the future into something repeatable, quantifiable, inevitable. But that’s an illusion.

Every deal, every company, every market inflection comes down to a leap of faith. You either believe in the world as it is, or you see it as it could be. You either stick with the consensus math, or you take the non-consensus worldview that feels just a little insane at the time.

We want to ascribe repeatability to all things. But the truth is: calculated luck, conviction, and the willingness to act when the numbers don’t look perfect are the only ways to actually create something new.

These companies are worth studying closely because they don’t just win in their markets. They create the markets. They endure ankle-biting competitors and come out stronger. They become the mental model for how that category should work.

What does it take to build one? The pattern is surprisingly consistent. Category defining companies require three things:

• Choosing a nascent or risky market

• Relentless execution on product and go-to-market

• Good timing, often helped by luck

Each by itself is not enough. Together they create the rare outcome where a company can become the default.

The Three Ingredients

Market Choice

Category defining companies almost always begin in markets that look too small, too weird, or too risky. At the start, these markets are dismissed by incumbents and ignored by most investors. That is exactly the opportunity. By the time others realize the space matters, the category leader has already been established.

Execution Compounds

The companies that win execute better than anyone else. Their product is 10x better, their go-to-market motion is tuned to the way their category buys, and their culture is focused on speed and iteration. Execution is where the advantage compounds.

Being Early vs Being Ready

Even with the right market and great execution, timing matters. Many of these companies looked too early at the start. What made the difference is that they survived long enough to be ready when the wave hit. Along the way, each got lucky with a break that gave them credibility or scale at a critical moment.

Five Category Definers

Twilio: Turning Telecom into Software

When Jeff Lawson founded Twilio in 2008, the idea that a startup could disrupt telecom seemed absurd. The industry was dominated by carriers and hardware vendors. Developers had almost no ability to add communication features into software.

Twilio changed that. Jeff’s insight was that communications could be abstracted into APIs. Instead of cutting deals with carriers or wiring up hardware, a developer could add a few lines of code and instantly send texts or make calls. It was a risky bet on a market that technically didn’t exist yet: communications-as-a-service.

Twilio focused on developers, not IT buyers. The APIs were simple, the pricing was usage based (which was unique), and the documentation was excellent. “Ask Your Developer” became more than a slogan. It was the strategy. Developers adopted Twilio in droves, and companies like Uber built critical features on top of it.

The smartphone and app boom created massive demand for embedded communications. Twilio was ready, and by the time the world caught up, it was the default. By 2018 Twilio had crossed $500 million in revenue and effectively defined CPaaS as a category.

Stripe: Payments for Developers

Stripe became famous for one thing: seven lines of code to accept a credit card. A developer could copy-paste those lines and start taking payments instantly.

Patrick and John Collison built Stripe in 2010 to fix a broken system. Online payments meant bank paperwork, clunky gateways, and confusing APIs. PayPal worked, but it pulled users into its own flow and was never designed for developers.

Stripe changed that with an API that worked out of the box. The pricing was transparent, and the focus on developer experience was relentless.

The growth motion mirrored Twilio: bottoms up and product led. Startups could adopt Stripe instantly, without sales calls or integration consultants. As those startups grew, Stripe scaled with them. Over time it expanded into subscriptions, marketplaces, and fraud prevention, building a broad payments platform.

Just as SaaS and e-commerce were breaking out, Stripe had the rails in place. Every new company wanted it. Stripe hit the milestone years faster than peers, crossing $500 million in revenue soon after 2016. Today it is the defining brand in developer-first payments.

Vercel: The Frontend Cloud

Deploying a modern web app used to mean wrangling servers, CDNs, and pipelines. Guillermo Rauch saw that frontend developers were underserved and decided to fix it.

He founded Vercel (fka ZEIT) in 2015 and chose an overlooked market. The strategy was a one-two punch: build the open-source framework Next.js and pair it with a cloud platform tailored for it. Every pull request got a preview URL, deployments were instant, and global edge distribution just worked.

Next.js solved real problems with React, like server-side rendering and routing. Its adoption spread quickly, creating a natural funnel to Vercel’s platform. Freemium entry and enterprise upsells gave the company a scalable model.

When React surged and SEO pressures pushed developers back toward server-side rendering, Next.js was waiting. By 2023 Vercel had scaled past $100M ARR and set the standard for the frontend cloud.

Datadog: Observability for the Cloud Era

When Olivier Pomel and Alexis Lê-Quôc founded Datadog in 2010, most people still thought of the cloud as a toy. Enterprises ran their workloads on-prem and monitoring was handled by legacy tools like Nagios, Splunk, or New Relic. Investors told them it was too early.

They were early, but they were right. Datadog’s insight was that DevOps teams would need a single place to see infrastructure, application traces, and logs as they moved to cloud. A simple metrics collector, hundreds of integrations, and a SaaS delivery model gave customers immediate value. From there Datadog kept expanding into APM, logs, and security.

Its GTM model combined free trials and bottom-up adoption with a sales team to land enterprise deals. Pricing scaled with usage, which aligned perfectly with the growth of cloud workloads.

Cloud adoption hit an inflection point in the early 2010s, and Datadog was one of the only companies ready with a mature product. A lucky break was surviving the lean years with just enough capital until the wave hit. By 2020 Datadog had passed $500 million in revenue and was the clear category leader in observability.

Okta: Identity in the Cloud

In 2009 Todd McKinnon quit Salesforce to build cloud identity. Even his wife thought he was crazy. At the time, enterprises used Microsoft Active Directory, Oracle, or IBM for identity, all tied to on-prem systems. A cloud alternative sounded impossible.

Okta executed by focusing entirely on identity-as-a-service. They spoke with dozens of CIOs to validate the pain, then built a platform that was reliable, secure, and easy to integrate. The Okta Integration Network gave IT teams pre-built connectors to thousands of apps.

Unlike Twilio or Stripe, Okta’s GTM was top down. Identity decisions are made by CIOs and CISOs, so Okta invested early in enterprise sales and support. It wasn’t fast. Okta took years to find product-market fit. But once it did, the trust it had built with enterprises paid off.

SaaS adoption was exploding, and companies suddenly had dozens of apps to manage. Okta became the neutral identity layer they needed. A lucky break came with early reference customers like Netflix. By 2019 Okta had passed $500 million in revenue and cemented itself as the category-defining company in cloud identity.

How They Won in Different Ways

When you zoom out across Twilio, Stripe, Vercel, Datadog, and Okta, you see sharper contrasts in how they each became category defining. The companies followed the same broad pattern (risky market, great execution, right timing) but the way they played the game was different.

Developer-Led vs Enterprise-Led

Twilio and Stripe are the purest examples of developer-led growth. Both built APIs that made complex infrastructure simple, and both went straight to the individual developer. Their products were self-serve, their pricing was pay-as-you-go, and their growth came from bottoms-up adoption. Twilio got embedded into apps like Uber without a central IT decision. Stripe became the default choice for startups spinning up online payments. In both cases, developers had the power to choose, and Twilio and Stripe gave them superpowers.

Datadog started bottom-up as well, but its market forced a hybrid motion. Monitoring often starts with a single team but quickly becomes an enterprise-wide need. Datadog had to layer on a sales force to close large deals. Its brilliance was in keeping the developer experience simple enough for bottoms-up adoption while investing in enterprise sales at the right time.

Vercel is another variant. Like Twilio and Stripe, it focused on developers, but its strategy hinged on controlling an open-source framework. By creating Next.js and then offering the best platform to run it, Vercel built a funnel that was both grassroots and product-led. That gave it a different kind of leverage. Developers adopted Next.js for free, and a significant percentage naturally flowed to Vercel.

Okta sits apart. Identity is not something developers pick in a hackathon project. It is a CIO-level decision about security and trust. Okta had to be enterprise-first from day one. That meant a heavier sales motion, longer sales cycles, and years of building credibility. It looked slower, but once Okta broke through, the stickiness of identity created a massive moat.

What They Displaced

Each of these companies defined their category by displacing entrenched incumbents. But the shape of the incumbents was different.

• Twilio went up against the telecom carriers and legacy contact-center vendors. These were slow-moving giants with no cultural affinity for developers. Twilio reset telecom by turning it into software.

• Stripe displaced PayPal, Authorize.net, and the banks. The incumbents were technically capable but architecturally outdated. Stripe outpaced PayPal by giving developers APIs that actually worked.

• Vercel replaced the DIY approach of piecing together AWS, CDNs, and CI pipelines. It also competed directly with Netlify. Vercel won by integrating an open-source framework with a cloud platform, which no one else had done.

• Datadog unified categories that had been split across multiple tools. It pulled usage from Nagios, Splunk, New Relic, and a dozen other incumbents. Datadog unified what others kept siloed.

• Okta replaced on-prem identity systems like Active Directory Federation Services and Oracle IAM. These systems were not built for cloud apps. Okta redefined identity by being neutral and cloud-first.

The contrast here is important. In some markets, the incumbents were slow and complacent. In others, they were simply mismatched to the new environment. What all five companies had in common is that they saw an opening to create a new standard that incumbents were too slow or too stubborn to meet.

Go-to-Market Fit

A critical piece of execution was aligning go-to-market with the way the market actually buys. This is where the companies diverged.

• Twilio and Stripe thrived on self-service and usage-based pricing because developers could buy without approval.

• Datadog needed a hybrid because observability starts small but becomes a C-level purchase.

• Vercel leaned into open source as the entry point and used freemium to capture the long tail while landing enterprise contracts.

• Okta built an enterprise sales machine because there was no other way. CIOs had to sign off.

This shows that there is no one-size-fits-all GTM model. What matters is the fit. Each company picked the model that matched its category and executed it relentlessly.

What They All Got Right

Despite the contrasts, there are a few themes that run across all five. These themes are what make them category defining.

Correctly Early
Each company looked too early when it started. Twilio was building cloud telecom before most developers trusted APIs for core features. Stripe was fixing payments in a world where PayPal seemed good enough. Vercel was betting on frontend infrastructure before most companies saw it as a bottleneck. Datadog launched when cloud adoption was still questioned. Okta pitched cloud identity before CIOs were ready. They had time to build, refine, and be ready when the market caught up.

10x Product Advantage
Category defining companies are driven by product quality and low barriers to entry. Twilio made telecom programmable. Stripe made payments copy-paste simple. Vercel made deployment instant. Datadog made monitoring unified and real time. Okta made identity secure and cloud-first. The product wasn’t marginally better. It set the new standard.

Developer Empowerment and Trust
Four of the five companies empowered developers directly. They gave individual engineers leverage to make decisions and ship faster. The fifth, Okta, won by creating trust at the enterprise level. Different audiences, same principle: put power in the hands of the user that matters most.

Consumption-Based Growth
Twilio, Stripe, Datadog, and Vercel all built pricing models that scaled with usage. Customers started small and paid more as they grew. This created natural net retention and made expansion built-in. Even Okta, with a seat-based model, grew as companies added more apps and more users. Growth was baked into the model.

Surviving the Desert
Each company had a desert period where the market wasn’t ready, investors weren’t interested, and growth was slow. Twilio pre-Uber, Stripe before e-commerce took off, Vercel before SEO pushed server-side rendering, Datadog before AWS hit escape velocity, Okta before CIOs trusted cloud. What made them category defining is that they survived, and were still standing when the wave arrived.

A Playbook

Category defining companies share a playbook.

• Pick the market others dismiss. Twilio picked telecom APIs, Stripe picked developer payments, Vercel picked frontend hosting, Datadog picked cloud monitoring, and Okta picked cloud identity. Each market looked small, unproven, or impossible. Each grew into a massive category.

• Execute relentlessly. They built products that were radically better, not marginally better. They matched their go-to-market motion to their market. They obsessed over their core user. Execution was not just about speed but also discipline and focus.

• Align pricing to customer growth. Usage-based or consumption pricing aligned their growth with the customer’s success. Expansion was natural. Retention was high. The business models scaled with the market.

• Be early, then endure. All five companies looked too early when they started. But being early gave them a head start. The hard part was surviving the years when no one cared. Each found a way to cross the chasm, and was ready when the market arrived.

• Take the lucky breaks. Uber made Twilio famous. Shopify boosted Stripe. React boosted Vercel. Cloud inflection boosted Datadog. Netflix validated Okta. These breaks weren’t controllable, but they were only useful because the companies had already done the hard work and doubled down when the opportunity struck.

The Next Category Definers

Defining a category is not about being first. It is about being the company that shows the world how the category should work. Twilio, Stripe, Vercel, Datadog, and Okta each did that. They chose risky markets, executed at the highest level, and rode the wave of timing and luck when it came.

The next category defining company will look just as contrarian at the start, until they become the default.

I’ve never been more challenged, more rewarded, more certain, and less sure, of what tomorrow brings. I’ve also never been happier.

Over the past few years, I got married, moved across the country, learned the joys and pains of old home ownership, built a new community, became a dog dad, quit my job, started a company, and now, most importantly, am expecting my first child in November.

Conventional wisdom says not to do all of this at once. But as someone smarter once said: “I’m in a swimming pool, I’m not going to put on a rain jacket.” Life is busy, wild, and about to get even busier, but it’s also more fulfilling than I ever imagined.

Oddly enough, being busier than ever, I’ve never felt freer. Now that I’m building Rely, I can see and appreciate so many parts of my life that I used to take for granted. I used to wonder if I’d be happier chasing my own dream. Now that I am, the answer is unequivocally yes.

How did I stay in “comfort” for so long when it made me so uncomfortable? I recently read something in my friend David Gerber’s newsletter that stuck with me, a quote from Rick Foerster: “The trap is when I do something that I’m good at and rewarded for. Because that situation locks me in: comfortable, successful, miserable.”

That hit hard. The “Trap of Competence” is real. While you’re in it, it feels impossible to escape. But once you’re out, it’s actually very hard to imagine ever going back.

I think about this constantly when it comes to building culture at Rely. Personally, I couldn’t feel more disconnected from the “ohana” vibe at Salesforce or the culture of complacency I experienced during my first job at Workday. That style works for some but it never worked for me. On the other end of the spectrum, I have no interest in creating the “sleep in the office, grind don’t stop” culture that chews people up and is romanticized by YC. I want a team that can sprint when needed, feels fully bought in, but also lives full, happy lives outside of work.

So what kind of culture do I want to build? What’s the version that feels right to me? Who do I want around the table every day? And maybe most importantly: how do we create a culture so strong that even if I weren’t the founder, I’d still want to be here? That’s the question that guides me.

My cofounder and I talk a lot about urgency. We both feel it deeply but our job is to inspire it in others, not enforce it with fear. We want urgency to come from possibility: “We get to climb this mountain. We get to build something great.”

If urgency is one element, agency is another. We want to hire high agency individuals but encourage them to step into more and more as they prove capable. This is the key ingredient of a culture I personally would want to be a part of. Amazon uses the term “bar raiser” in hiring and Frank Slootman talks a lot about hiring for slope rather than intercept. We want to work with high agency people who raise our bar and are just getting started.

And above all, we want to work with good people. People who assume best intent. Who trust each other. Who act like owners and obsess over what’s best for the customer and the business. If that’s your default lens, you avoid most of the silly workplace politics.

As a founder, this is the part I think about more than anything. My day-to-day role changes constantly, some days I’m hiring, others I’m selling or fundraising, but one thing is always true.

My job is culture. Every day. Because culture is what shapes who we hire, how we work, how we grow, and how we win.

Reach out to george@tryrely.ai if you want to join our team! We would love to meet you.

Much has been written about finding early market validation, building founding teams, and raising initial capital. But what happens after you've done all that?

Founders should have one mission: build the product, team, and brand that makes you hard to kill.

Product

The first product a startup brings to market likely isn't defensible. The fact that it could be built in a short amount of time inherently means someone can copy it. And as soon as you show signs of momentum, someone will.

Now, that's a generalization. The Seven Powers do apply to great products. However, in the business of SaaS, it's unlikely the flywheel of monopoly has started spinning ahead of your Series A.

Product velocity doesn't just help you close more customers. It makes you harder to kill. The more aggressive you are with product expansion, including moving beyond individual features to entirely new product lines, the more surface area you're able to cover.

This is not easy. You have to build good products. But if done right, in the best case scenario you can sell more features, and in the worst case scenario you have a hedge.

People

An early startup team can feel very codependent. If your only salesperson is on vacation, sales velocity slows. If your lead engineer is out sick, product doesn't ship.

As soon as possible, it's critical to add coverage across all functions of the business. It allows people the space to get sick, handle life stuff, and learn from each other.

This is why going from 1 person to 2 on a team is more impactful than going from 2 to 3.

I commonly see startups get into a slump after not growing beyond 10 people, as that team size is unsustainable to keep up with the momentum of a rapidly growing business.

Over time, great teams operate like boulders rolling down a hill. As they continue to grow in size and execution, there's very little that can slow them down.

Brand

Brands are like oak trees. They take forever to grow, require significant resources, but once established they're sturdy. Of course, a hurricane can still take them down. But brand is one of the most important moats. Brand creates distribution, earns trust, and closes deals that are neck and neck. It also fuels the necessary hiring pipeline to keep great people coming.

Building great brands requires taste and time. The establishment cost is high, but the compounding growth is invaluable. It's much harder for others to copy.

The ultimate power of brand is becoming the default. Being the default means you've already won the consideration phase of every deal. Competitors have to explain why they're better. You just have to show up.

Becoming Antifragile

Business durability comes from building resilience to survive inevitable failures. Every startup will face existential threats. The question is whether you've built something that can take the hit and keep moving.

Momentum is your most valuable asset, but it's also the most fragile. Products, people, and brand are what transform that momentum from a temporary advantage into a permanent one. They're what keep you moving when everything else tries to make you stop.

The startups that survive aren't necessarily the ones with the best ideas or the most funding. They're the ones that became too stubborn to die.

I’ve seen a lot of hype around using AI for everything: solving problems effortlessly with instant access to the world’s knowledge and expertise at our fingertips. This is truly amazing, but it also opens the door for human nature and laziness to take over. Relying entirely on LLMs for understanding diminishes our ability to create new knowledge. Without a deep, fundamental grasp of what we’re actually doing, we risk losing the critical thinking needed to innovate and move forward. Put simply, those who leverage AI to accelerate their learning will surge ahead, while those who don’t will quickly fall behind.

A brief history of Lumosity

Lumosity was founded in 2005 by Michael Scanlon, Kunal Sarkar, and David Drescher, who sought to make brain training accessible and engaging for everyone. The idea emerged from a shared belief that the brain, like any other muscle, can be strengthened through targeted exercises. They built the platform to help users improve cognitive abilities like memory, attention, and problem-solving. Over time, Lumosity grew rapidly, leveraging neuroscience research to create a product that blended fun and functionality. The company’s success lies in its ability to gamify brain training, making cognitive enhancement both effective and engaging for millions…I guarantee platforms like Lumosity, DuoLingo and anything else that helps us learn will get substantially more valuable in the years to come.

Brain atrophy

As we increasingly rely on LLMs to handle complex thinking and decision-making, brain atrophy could become an unintended consequence. By offloading cognitive tasks to AI, we risk reducing our own mental engagement and the opportunity to exercise our brains. Over time, this lack of intellectual stimulation may lead to a decline in our cognitive abilities, weakening our capacity for critical thinking, problem-solving, and creativity. Just as physical atrophy occurs when muscles aren’t used, brain atrophy may emerge from not actively engaging in thought and analysis. If we don’t intentionally challenge our minds, we could see a generation of individuals who, while technologically proficient, struggle with deeper cognitive processes as they become more reliant on AI.

The result…

Basically, a possible outcome here is the movie Idiocracy, where the average IQ of the human race dropped dramatically and everyone became a moron.

It’s clear that the future holds incredible promise. LLMs are rapidly transforming how we work, learn, and solve problems. But with this comes a critical responsibility to ensure we don’t lose sight of the human element, our ability to think deeply, challenge ourselves, and push the boundaries of innovation. The future is not just about using AI to make things easier; it’s about using it as a tool to accelerate our learning and creativity while still maintaining the intellectual rigor that defines our progress. We have already made our lives so easy that humans get wound up about truly meaningless things…everyone needs a purpose! I would hate to see that get worse…

What excites me is using AI to accelerate learning, not bypass it. We can build a future where our minds are sharper and more adaptable. The future is bright, and the possibilities are limitless if we continue to challenge ourselves, think critically, and use AI to amplify, not replace, our own cognitive growth.

The hard truth about developer tools is that there's no silver bullet to building something great. Success requires identifying a genuine problem underserved by the existing market and innovating in a way that delivers not just a better solution, but one that's both defensible and commercially viable.

Many developer tools founders stumble at this intersection. As developers themselves, they build something impressive, feel compelled to open source it, and share it with the world. But transforming that creation into a sustainable business presents a significant challenge. Success demands both favorable market conditions (stagnant incumbents or rare technological shifts) and technology you can meaningfully control.

Developer tools companies face a classic dilemma: those avoiding open source typically lose to competitors who embrace it, while those who open source struggle with commercialization because they fail to protect intellectual property or capture necessary value. It's a precarious balance, you can't be entirely open source nor completely closed source. The key is threading the needle between building an excellent product with free components while capturing value durably.

This delicate balance eludes many developers. Moreover, what initially brings success rarely scales to build a billion-dollar company, a disappointing reality. I've witnessed numerous talented developer tools founders create outstanding products that never successfully reach market or commercialize because they either give away too much to the community or pivot to enterprise and essentially become service businesses.

The developer tools sector produces relatively few billion-dollar companies compared to the broader B2B market, largely due to differences in value perception and the culture around developer tools, particularly the ability to self-host. At Browserbase, we've approached this differently. We've open-sourced Stagehand, our browser automation framework, believing it shouldn't be tightly coupled with Browserbase. While we maintain an open-source browser infrastructure, its true value isn't in being open source, the real challenges lie in maintenance, setup, and addressing the countless edge cases that make production deployment difficult, issues our team tackles daily.

For developer tools, there's simply no universal formula for building a great company. Success depends on specific circumstances, much like Earth's existence resulted from precise conditions enabling life. Building a commercially viable, successful, and beloved developer tools company requires the perfect convergence of market conditions, technological innovation, talented people, and strategic decisions, a truly rare achievement.

Raising a pre-seed round is as much art as it is strategy. But once I reframed the process as an opportunity to meet incredible people, everything changed. I started having fun—and when you enjoy it, everything gets easier.

After meeting with over 100 investors in 10 days, these are the key lessons I wish I’d known before starting this journey:

1. Preparation and Confidence Are Everything

• Practice, practice, practice. Before diving into a wave of meetings, refine your pitch. Get feedback from trusted friends and advisors (people who have actually done this), iterate, and get to the point where you sound confident, polished, and genuinely excited. Your energy has to scream, “I AM THE PERSON FOR THIS.”

• DO NOT READ YOUR SLIDES. The deck is a tool, not a script. Know your material so well that your passion and clarity carry the conversation.

• Your first few meetings will be throwaways. It just takes a few reps to get in a groove. You need to be in live settings to respond to real feedback and refine the pitch as you hear what does or doesn’t resonate.

2. The Power of Some Introductions

• Warm intros are everything—and founder intros are gold. A recommendation from a portfolio company founder carries exponentially more weight than anything else.

• Be mindful of VC-to-VC intros. If a VC isn’t going to invest in your company but will pass you to someone else, why would the next person invest? There are cases where the investment isn’t a fit and the intro may be very valuable but just be mindful of what this may signal.

3. Build Momentum and Heat

• Timing matters. Cram your pitch meetings into a short timeframe to build momentum. The perception of “heat” is real, and it can influence how investors respond. This means your schedule will be very packed. I’ve had 16 meetings in a single day and over 40 in a week.

• Schedule thoughtfully. Place your least important meetings at the beginning and your most important meetings later. Your pitch will get better and if interest builds, you’ll have a better chance with the pickiest investors.

• Everything is a funnel. Expect to take an unfathomable number of calls and get a ton of no’s. Some of the best founders I know got over 50 no’s before receiving their first yes…which means prepare yourself mentally for a whole lot of rejection. It is part of the process!

• People talk. VCs hear about deals from one another. The fact that you’re talking to people will make its way around.

• You can leverage one offer to get other offers…and improve the terms.

4. Positioning is Important

• Everyone is lucky to be getting the opportunity to talk to you. You aren’t there to beg for money, you’re there to find the right partner to build your business. Remember that and act like it.

• Collect real commitments early. “Soft circling” funds are not the same as having the money wired to your account.

• Tell a big story.

• “It only takes one.” One great investor, one term sheet, one believer—that’s all it takes to move your round forward.

• Who to talk to: Ultimately the warmth of the intro determines where you start. The more senior the connection, the better, within reason. You just don't want to talk to the partner who is mostly checked out. It is also helpful to know who has backed deals like yours in the past. Worth identifying pre-introduction but also not always within your control.

• Know your ideal round construction before starting to pitch. People will ask what you want and from whom. Always say a range of money and a range of dilution. That gives VCs two levers to play with and doesn’t lock you into anything.

5. Know the Reality of Fundraising

• Some meetings are over before they start. Within the first five minutes, you’ll know if someone is genuinely excited or just taking the call. Learn to recognize these signs and save your energy for the ones that matter. Some VCs meet you with immediate skepticism, some VCs meet you with curiosity. The more curious they are, the better chance you have. There are also people who have no experience or interest in the end market and that often shines through. Those people are very hard to turn.

• If they don’t follow up in a couple days, move on.

• Fundraising is draining. Emotionally, physically, and mentally, it’s a marathon and a sprint. When you’re in the process, don’t expect to get much else done.

• For pre-seed multi-stage funds come with risks. If they don’t follow on later, it can create negative signaling. Sometimes smaller investors and smaller pre-seed rounds preserve more optionality. Also, if you want your VCs to spend time with you, you’re better served finding a smaller group with fewer investments.

• Sometimes, you never stood a chance. Some people are literally only meeting you to gather information and share it with their portfolio companies or form a market thesis.

• Geography matters. Being in Maine adds a degree of difficulty, but east coast VCs care less. Talk to everyone but spend time locally as well. While this matters less today than ever, it does still matter.

• Elitism is still very present. Someone literally said “if you raise money from a fund I haven’t heard of, I’ll just wonder why Sequoia passed.” Just expect interactions like this from time to time. As I said before, you want a partner and you’re interviewing them too. If you get these vibes, move on.

• Some people will try to waste your time. Only go through a long, drawn out process if you really want to and like the firm. Otherwise, just respectfully pass.

• Everyone thinks their strategy is the right one. Talk to everyone, be exclusive, have a deck, write a memo, only have angels, don’t raise at all…

• An offer is not real until it is in your hands. A verbal offer can always be rescinded. I’ve heard the words “I’m in” only to receive a call 5 days later saying “I’m not in.”

6. Early-Stage Realities

• Many funds claim to do pre-seed, but the reality and their expectations don’t align. Research before engaging and when someone says “we do pre-seed” but also “we like to see $100k-$500k in ARR” then just skip it. Know the fund structure, the number of deals they do each year, per partner. Ultimately, they might even like what you’re doing but the deal could simply be incompatible (check size is too small, market they don’t understand, etc.).

• Early stage investments are founder bets. Ultimately, pre-seed rounds are entirely based on the founding team. It is important to sell yourself and your co-founders as much or more than what you’re building. Angels and small funds can be powerful allies this early. They’re often more flexible and willing to bet on the founder.

• Balance focus with building a network that works for you. Some VCs believe meeting with more than 10–15 funds signals a lack of focus; others think you should meet everyone.

• Be wary of “advisors” who will trade time for equity. Advisors who invest demonstrate actual confidence in your ability to execute. Advisors who just want equity in exchange for their time are often leeches, especially if they approach you.

• Don’t price yourself for perfection. If you raise too much money and too high a valuation too early, it will be very hard to raise money again and the expectations of you will be very high.

• Excel is still the best fundraising CRM…

7. Create Energy Before You’re Ready

• Start building relationships early—even before you think you’re ready. It builds “potential energy” in the pipeline and makes the formal raise less intimidating.

• As you move forward, always be raising, but strategically. Know when to shift from informal conversations to full-on fundraising mode and when you do, return to bullet number one!

8. A Note of Gratitude

• Sometimes, the difference between success and failure is having the right people around you to light a match. Surround yourself with people who believe in you, push you, and keep you grounded. Lean on them when the process gets overwhelming. And do the same for others when it's your turn.

Fundraising is hard, messy, and unpredictable—but it’s also really fun and an amazing learning experience that forces you to grow as a founder. Take it one step at a time, and never lose sight of the bigger picture: building something worth fighting for. Remember: closing the round means you’ve just been invited to start the race…now you’ve got to actually do the work!

It’s a miracle that new developer tools companies come into existence in this unforgiving market. But when evaluating the companies that endure, you’ll find they overcome three main hurdles:

1. They start on the right foot: The original strategy of successful companies like Stripe, Twilio, Clerk, Supabase, and Vercel has been continuously accurate. There wasn’t an early pivot. They had a straight line from idea → execution. Just like how software needs to be architected well from the beginning, developer tools companies need solid strategic foundations.

2. They grind for their first 100 believers: Each customer in this stage is extremely precious, and is taking a massive bet on a very much unproven product. These first 100 customers will have a bad time using your product, but it’s how you respond to them that sets the tone for the duration of your company. At 100 paying customers, you’ve likely found the majority of the sharp edges of your product, validated the core offering, and established yourself in the market.

3. They innovate on distribution: Going from 100 → 1000 customers requires not only excellence on product, but also true innovation on getting the product into customers hands. Developer tools companies must find, educate, and win over developers, who are notoriously critical buyers.

The focus of this memo will be on hurdle #3.

Developer Trust

The most important factor to building a category defining developer tools company is to win the trust of developers. Especially in the early days. I’ve been a CTO, technical lead, and buyer at companies large and small. The first thing I would do when making a build in house vs buy decision was simple: I’d ask another developer what they recommend.

Developers often feel like they have no control. Their CEO demands a feature be built, and doesn’t care about the tech debt or tooling issues that are really weighing them down.

That’s why when a developer feels understood, they have unusually positive reactions.

Word of mouth and positive developer reputation is one of the hardest things to build. It takes a long time, and can easily crumble. You can’t growth hack your way out of a negative reputation.

That’s why the most important driver of growth is positive NPS. At the core of everything we do, we need to ensure that we’re continuing to earn the respect and trust of our customers.

But how do we do that? It’s hard to capture every way within a single document, so you should start with this core value: Will a customer speak positively of this interaction with Browserbase?

A central part of this is customer communications. We need to be prompt, polite, and genuine in our communications to customers. The customer isn’t always right, but as long as we’re authentic when speaking to customers, they cannot fault us.

This happens in many small interactions via our support channels, but also bigger conversations like sales calls and product documentation. If we’re able to master the ability to empathize with customers and win their affection, we’ll benefit from the word-of-mouth growth that’s core to everything we do.

Product First Growth

How would you define our product? What if I told you it was much more than our APIs and SDKs.

Our product encompasses every touch point a developer has with Browserbase. From our website to our documentation, our onboarding flow to every little email they receive from us. It’s critical to think holistically about our product as a driver of developer trust.

While it may seem unrelated, a customer will judge the reliability of our infrastructure based on the care and consideration we put into our documentation. Would you hire the gardener who has weeds growing in their front yard? Investment in our product experience outside of our core offering is a multiplier on growth.

Time To Magic

The most important milestone in a prospective customer’s journey is the “a-ha” moment. However they discovered us, they’re still trying to put the puzzle pieces together on why they need what we’re offering. If a prospect doesn’t understand why they need us or what we do, we have an uphill road to climb.

The most scalable way to educate customers on what we do is to “show not tell”. And luckily for us, our “show” is pretty compelling! With one click of a button in our Playground, we can show them a web browser being automated on their behalf.

Even if we’re not a good fit for a prospect, if we’re able to successfully educate them on what we do, and impress them with our product experience, we may win a potential advocate.

The Art of Objections

I wish it was sufficient to simply get our prospect to the “a-ha” moment. But any savvy developer will have a plethora of reasons why our solution won’t work, won’t scale, will be too expensive, etc.

Now, we can’t overwhelm them all at once with all of the counterpoints to their objectives. That feels sales-y, and developers hate being sold to. Instead, we need to offer breadcrumbs to the next step in gaining a deeper understanding of our product. There are a few channels for this:

• Our dashboard experience

• Our documentation

• Our quick start guides

• Our customer emails

All of these are canvases for furthering the customer journey. If we’re able to think critically about each of the objections a prospect might have at each stage in their journey, we can surface the right next feature or next step they need to take to further their understanding of our product.

The Sales Shortcut

Developers don’t like sales calls. They want to tinker, try things, and build. They want to come to their own conclusions.

But sometimes, they don’t have enough time to explore on their own. They want someone to explain to them how to use this thing. Other times, they’re a “special snowflake”, who believes they’ll need a bespoke implementation or feature.

The challenge: scaling this to thousands of prospects is expensive, and we can’t do personalized demos for every individual developer.

Our “Contact Sales” button is the gatekeeper for our time. We don’t want to waste our time, and we most certainly don’t want to waste our customer’s time!

That’s why we need to “qualify” customers before we invest time in them. Oftentimes, they might be better served spending more time in the product.

I really like how Posthog does it. There’s a demo video on the page, and information form that allows us to gate keep the calls. This form is an opportunity to remind customers of the self service documentation that already exists.

Each additional prospect we take on results in a lower likelihood to closing the others.

If product-led growth is a machine gun, sales-led growth is a sniper rifle.

Growing Weeds & Hunting for Whales

It’s no secret that 80% of our revenue will come from 20% of our largest customers. This is the nature of infrastructure platforms and usage-based businesses. If someone isn’t using your product at true scale, ACVs won’t grow.

But how do you find these large customers? There’s two main channels:

1. Growing Weeds: We are able to win customers early in their life and retain them as they become a household name.

2. Hunting Whales: We get in front of the qualified prospects that already have meaningful adoption and win their business.

Both of these channels are incredibly important for infrastructure businesses.

The former is more natural for a product-led-growth motion, as it’s incredibly hard to predict which early startup will be a break out success (just ask our investors). Being able to win the bottom of the market gives us an opportunity to capture the next Airbnb at inception.

The latter follows a traditional sales led motion. It’s highly unlikely that giant contracts will fall out of the sky and meaningfully adopt our product. Instead, we have to go to them. To repeatably win larger prospects, we need to be creating opportunities and inserting ourselves into their roadmap.

Running a dual motion (PLG + Sales Led) is generally ill-advised. Most companies that try this often fail to do well at either. However, I believe that a PLG motion, combined with targeted, high effort outbound can be very complimentary.

Here’s what targeted, high effort outbound looks like:

• Crystal clear understanding of our ICP, value prop, and differentiation

• A target logo list that is realistically serviceable by our current product

• Warm introductions to champions and execs via investor + partner network

• Persistent, candid follow up on conversations to move things along.

• Regular, in-person engagement opportunities with our buyer (dinners, events) to build rapport.

• Readiness to engage in a multi-quarter sales cycle (never rushing a deal)

What we aren’t doing is building a team of SDRs and cranking through every logo on the F500. Why? Well, at our stage, our GTM motion needs to be as lean and efficient as possible. But secondly, without strong reference-able customers, each deal will be extremely challenging to win.

The Logo Flywheel

One of the biggest hurdles developer tools need to overcome is market recognition. How can someone trust that we’re not going to break their product? Buying a developer tool is risky, especially when a prospect is committing to an annual contract.

The easiest way for us to derisk our product is to build social proof. It’s a tale as old as time: You’re more likely to go into the crowded bar than the empty one, because clearly something good is happening in there.

Recognizable customer logos earns us credibility. Credibility makes it easier to win deals. And thus, a flywheel begins to spin.

But I can not state this enough: it has to be backed up by a good product. Regretted enterprise churn is some of the worst out there. It spooks investors, employees, and prospects all at once.

If someone sees a logo on our website and decides to call the founder of that company, and they have nothing but positive things to say about us, it does more for our ability to close a deal than anything that comes out of our mouth.

In Summary

Building a category-defining developer tools company starts with winning the trust of developers.

Our product is our biggest lever for distribution, and investment in the product experience is a multiplier on growth.

Luck is what happens when preparation meets opportunity. By combining amplified word-of-mouth, impressive product experiences, and targeted high-effort outbound, we'll create opportunities that our product is prepared for, and build the logo flywheel that makes Browserbase the default choice in our market.

Ever since college, I’ve wrestled with finding meaning in my work. The closest I’ve come has been through side projects like Slice of SF, Nexus Maine and Hawk Hill Ventures. I’ve spent countless hours admiring those I look up to, wishing for their success, while avoiding the steps required to get there. I was enamored with the destination, not the journey. But over time, I began to notice a pattern: my moments of professional excitement and fulfillment always revolved around one thing—agency.

Updater gave me my first taste of agency, and for that, I’ll always be grateful. But the longer I stayed on the traditional path—software engineer to product manager to director of product—the more I realized it wasn’t for me. I wanted to build. I wanted to sell. I wanted to understand customers deeply and market to them effectively. I wanted to learn. Yet, in most of my roles, I kept hearing, “That’s not your responsibility” or “Stay in your lane.”

Nine months ago, I woke up and finally recognized that only I could change my situation.

The Team

First of all, none of this would be possible without the incredible support of my wife, Anne and her belief in me. Absolutely none of it.

People are the single most important ingredient. I am incredibly lucky to have found my co-founder, David LoBosco. Introduced by our mutual friend and his former CEO, David Engle, David LoBosco is a rare talent—a 10x engineer, intuitive designer, and product thinker who always puts the customer first. His skills complement mine, and he challenges me to think bigger and execute better. This journey wouldn’t be possible without him.

I also want to acknowledge Paul Klein, my partner in Hawk Hill Ventures and a founder I deeply admire. Paul’s resilience, grit, and generosity have been a guiding light for me. Watching him tackle the highs and lows of building Browserbase has been inspiring. His support gave me the confidence to take this leap.

To the rest of my family, friends, and mentors—your encouragement has been instrumental, and I’m beyond grateful for the role each of you has played in getting me to this point.

Getting to the Idea

Sometimes you can’t get an idea out of your head and it just starts to eat away at you. You question the very nature of an industry, someone’s way of working, the way time is spent, something you wish existed…

And sometimes, you want so badly to force something into existence by sheer power of will, regardless of the need. You want to create a reality that is different from your current one. You ask yourself “who do I admire? What do those people all have in common? What do I like about my life? What don’t I like about my life?” The list goes on and on and eventually, you take steps in the direction you think is right. However, aftering hearing about why I thought it was time for someone to build better gym software because MindBody sucks, my friend and advisor Drew Geant asked “are you playing to play, or are you playing to win.” That knocked me back a few pegs…

The idea maze is a tricky place to be and when you lack direct experience and a historical background on what you’re considering, it can feel neverending. The process sets you on this rollercoaster ride of highs when you think you’ve figured it all out and lows when you realize you’ve just come up with A) a horrible idea, B) an entirely undifferentiated idea, C) a tarpit idea, D) something impossible for you to build, or E) all of the above! After a while in the maze and thinking about Drew’s observation, it was time to take a step back. It was time to ask myself what problem I actually know about, what unfair advantage do I have, what motivates me.

There is one problem that has actually annoyed me for years on end and hasn’t gone away but the path to solving the problem seemed daunting and improbable…but that is because I kept looking at how it had been done before.

To reference Dharmesh Shah, many of the dots I’d been collecting for so long finally started to connect. The vision started to become clear. The problem now feels urgent, and the solution almost inevitable. It’s not just about iterating on the old but instead creating something entirely new; it’s about transforming the way people work, freeing them from inefficiencies, and building something that feels essential to the way we will work in the future.

What you’re about to read has finally been born from this drive to solve real problems in a meaningful way and modernize the way the physical world is managed.

The Industry

The property management industry generates about $100 billion annually but it feels like a small, tight-knit community. It’s an industry where everyone knows everyone, and even competitors maintain friendly relationships. I love it, I’ve never experienced anything like it and I am motivated to help improve it.

In this world, there are a few universal truths:

• Products are sold, not bought.

• Nobody likes their property management system.

• Nobody will rip out their property management system.

• Site-level staff turnover is absurdly high.

• Software is rarely bought by the people who actually use it.

Property management software emerged in the 1980s with pioneers like Yardi and RealPage. These companies built the first on-prem systems, and while they’ve since moved to the cloud, their solutions remain far from modern. Their dominance has created a fragmented ecosystem of point solutions, resulting in silos, login fatigue, user dissatisfaction, inaccuracies...

After decades of unbundling, the industry is now craving rebundling. Several challenger property management systems have recently emerged, raising significant funding and sparking hope for a better future.

Introducing, Rely…sort of…

Rely was born out of a desire to combat the pervasive effects of turnover in property management. We’re building software for the people who actually use it—an experience designed to empower end-users, simplify workflows, and modernize the way the physical world is managed.

More to come on this soon though. Stay tuned.

Conclusion

When I graduated from middle school, I got to pick a quote that I liked or felt I could relate to and what I chose is, in hindsight, wildly embarrassing. I chose something from John Mayer: “everyone’s a stranger, but that’s the danger in going my own way.” While I was probably 13 years old at the time, something about that quote still resonates with the log-sized chip I have on my shoulder.

This decision to start my own company feels both daunting and exhilarating. For years, I’ve looked up to founders and creators who had the courage to step off the beaten path and build something meaningful. Now, it’s my turn.

For me, starting Rely is about more than just solving a problem. This is also about living with intention, chasing the vision of the life I want, and creating something that matters. It’s about choosing the harder, more uncertain path because that’s the one that aligns with who I am and what I believe.

I know this journey will be filled with challenges, setbacks, and unexpected twists. But I also know it will be filled with growth, learning, and the joy of creating something from nothing with amazing people. For the first time in my professional life, I wake up every day with a clear sense of purpose and excitement about the road ahead.

To everyone who’s supported me, challenged me, and encouraged me to take this step—thank you. Your belief in me has been a guiding force, and I’m endlessly grateful.

To anyone reading this who is contemplating their own leap of faith, bet on yourself. The journey may be uncertain, but it’s the only way to truly discover what you’re capable of.

It’s time to chase the dream. This year, I woke up and decided that “I am going to play for the asymmetric life.” – Paul Klein – Graham Weaver

2024 has been a monumental year—not just for us, but for the world at large. From a closely contested presidential election to escalating global conflicts, from the Paris Olympics to SpaceX’s jaw-dropping feat of catching a rocket with its chopsticks, this year has been anything but ordinary. As we reflect on these transformative events, we also look back on a year of growth, learning, and impact at Hawk Hill Ventures. This marks our 12th and final post of 2024, a year we’ll remember as one of bold moves and incredible moments.

Looking back

In 2024, we reviewed over 300 deals, met with over 100 founders and made 13 investments. Browserbase closed pre-seed, seed and series A financing rounds all in 2024! These are really exciting but it isn’t fair to just pat ourselves on the back. How did we do on our predictions from last year?

George’s 2024 predictions

2024 is the year of fintech - Grade: A

Fintech has far outperformed the broader market with market leaders like Affirm up over 50%, Dave up almost 10X, SoFi up over 60% and more, we can confidently call this one correct. VC funding for fintech hasn’t increased a ton but overall the market reception for this category of company has improved tremendously.

VC contraction is just beginning - Grade: B-

2022 to 2023 saw a massive contraction in overall capital raised by VC firms and number of active funds. That trend continued through 2024 but not at nearly as substantial a rate (Source).

The groundhog will see his shadow - Grade: F

Puxatawny Phil did not see his shadow, unfortunately.

Paul’s 2024 predictions

Design-to-code AI tools gain mainstream adoption - Grade: A-

It turns out AI is really good at code generation! In the original prediction, I specifically called out v0 , and since then it has become much more popular for developers building frontend MVPs. I’ll go with an A- for this because it isn’t really design-to-code, but idea-to-code, as people are describing their ideas to Claude and getting entire applications designed and built for them.

Early stage venture has its best year of the decade - Grade: B+

If George gave himself a B-, I’m going with a B+ here! While total capital raised by funds was down y/y, capital deployed was up! We saw some amazing companies formed across many categories in 2024, and I’ve never been more thrilled by the opportunity + true disruption being driven by these ambitious founders.

49ers win the superbowl! - Grade: F

In classic niners fashion, we choked in overtime and could not pull it off. I’m not getting my hopes up again this year.

Looking ahead

Rather than adding another round of predictions here, we will take a moment to build on Paul’s memo, The Art of Consequential Choices.

“Perhaps the key isn’t to seek perfect decisions, but to make thoughtful choices that align with who we are, while accepting the inherent uncertainty of every path we choose. In the end, life is much more about the journey than the destination anyways.”

As we reflect on these words, we’re reminded of the power of embracing calculated risks—balancing thoughtful decision-making with the acceptance that uncertainty is part of every worthwhile endeavor. This mindset has been central to our growth this year and continues to guide us as we look ahead.

2025 is shaping up to be an exciting year for us, with new directions and opportunities that we’re eager to share soon. Stay tuned—we can’t wait to take you along on this journey.

How can one accurately forecast the right choice? Truthfully, I don't believe you can. There's so much variability and luck involved that even a statistician cannot guarantee the highest EV selection will be the right one.

So in high-stakes decision-making, where the commitment is both irreversible and demands significant resources (cash, time, reputation), how does one make the best decision possible?

It comes down to optimizing for what you can control - the guaranteed outcomes from each path. Take starting a company: regardless of success or failure, you'll learn invaluable lessons about business and resilience while building a network you wouldn't have otherwise. Even if the venture goes to zero, you've gained something valuable.

The key question becomes: Is my guaranteed upside worth more than my guaranteed downside? Many people focus solely on best and worst-case scenarios. But extreme outcomes are both hardest to predict and statistically least likely.

When contemplating leaving his lucrative investment banking career to sell books online, Jeff Bezos asked himself: "At age 80, will I regret not having tried this?" In his mind, the idea, his life’s circumstances, and the timing of the market made the decision once in a lifetime. It's worth considering whether an opportunity like this will come again. If not, you must make peace with that reality when making your decision.

If you do step through a one-way door and things don't go as planned, remember the principles that guided your decision. In poker, you can make the correct choice with limited information and still lose. If you let losses change your decision-making criteria, you'll never develop sound judgment. We must focus on playing the hand we know we have, accepting that much lies beyond our control.

One of life's great paradoxes is how monumental every decision feels in the moment. Our pattern-seeking minds can trace clear lines from past choices to our present reality. But perhaps we overestimate how much any single decision shapes our ultimate destination. After all, our fundamental nature - our values, instincts, and patterns of behavior - tends to guide us toward similar ends despite different routes. Perhaps the key isn’t to seek perfect decisions, but to make thoughtful choices that align with who we are, while accepting the inherent uncertainty of every path we choose. In the end, life is much more about the journey than the destination anyways.

The Cost of Inefficiency

My back-of-the-envelope consulting math suggests that there are roughly 55,000 documentation specialists in the U.S. and 210,000 corporate educators. Worldwide, those numbers are more like 800,000 across both roles. The average salaries range between $68,000 and $75,000 per year, which means companies in the U.S. spend $18.4 billion annually and globally, $54.8 billion per year on these functions alone. This feels unnecessary when we think about the technology available to us.

For many companies, the management of knowledge and internal training still requires significant manual intervention, creating inefficiencies and adding avoidable costs. This is especially true when knowledge workers spend up to 30% of their day searching for or recreating documents, which reduces productivity and employee satisfaction (Sources: Whale, Armstrong Archives). 

Why Should Knowledge Bases Be Verticalized?

The key to transforming internal documentation lies in verticalization—creating domain-specific models trained on the core systems of record for each industry. A one-size-fits-all solution for knowledge management or employee training doesn’t address the nuances of various industries like property management, healthcare, legal services, or manufacturing. Each sector has its own tools, regulatory requirements, and processes that need tailored, embedded knowledge solutions.

Take healthcare, for example. Medical professionals need instant access to up-to-date, accurate documentation about treatment protocols, patient records, or regulatory changes. A generic knowledge management system won't suffice because it doesn't account for the intricate workflows and legal obligations within the healthcare field. However, a verticalized platform trained on healthcare-specific data—integrating seamlessly with electronic medical records (EMR) and regulatory systems—would revolutionize the way healthcare providers access and update critical knowledge in real-time.

A Conversational Future, Meeting You Where You Are

Beyond just being verticalized, the future of knowledge bases must be conversational, a system that employees can talk to, asking questions in natural language and receiving accurate, real-time responses. This system needs to integrate with existing collaboration tools to avoid context switching: Teams, Slack, etc. This conversational AI would continuously learn from user interactions, keep knowledge bases fresh and updated without manual input. This cuts down on the overhead of documentation specialists and corporate trainers, reducing the need for human intervention, and allowing organizations to focus on value-adding activities instead.

The Bottom Line

As businesses continue to move toward automation and AI, it’s clear that there’s an opportunity to drastically reduce the costs associated with internal documentation and employee education. The impacts are far greater, though, in high turnover industries. Initial and ongoing training time is reduced, efficiency improves and this ever evolving system can proactively identify opportunities for process improvement. One day, this system could be a global search product, a unified data platform and more.

By leveraging verticalized, conversational, and self-updating knowledge bases, companies can not only cut costs but also drive efficiency and improve the employee experience across the board.

The future of work isn’t just about automating processes—it’s also about empowering employees with instant, intelligent access to the knowledge they need to do their jobs well.

A better product might not succeed for several reasons: maybe the market doesn’t feel the pain point you’re solving, or perhaps you don’t understand the complex dynamics at play in the industry. Great technology alone isn’t enough. If you don’t get the strategy, timing, and go-to-market approach right, even the most innovative solutions can fall flat.

Consider the challenge of selling to government agencies. It doesn’t matter how advanced your product is—if you don’t know how to navigate procurement or the sales processes that incumbents understand deeply, you’re going to struggle.

At the core, building a great product is only half the battle. Startups need to master market fit and strategy if they want to truly disrupt industries.

Example 1: Blockchain, A Solution in Search of a Problem

The blockchain industry has offered a technology with the potential to disrupt everything from financial systems to supply chains. However, aside from cryptocurrencies, there hasn’t been mass adoption of blockchain-based solutions. Why? In many cases, blockchain technology is a solution searching for a problem. It promises decentralization, but the average consumer hasn’t been compelled to adopt it because it doesn’t address a tangible pain point.

Example 2: Quibi

Quibi, a short-form mobile video platform, launched with a $1.75 billion investment and a high-profile leadership team. Despite having high-quality content and innovative technology, it failed within six months. Why? The product didn’t solve a problem that consumers actually cared about. At the same time, Quibi’s marketing strategy misunderstood the nuances of mobile consumption, and its subscription-based model was not well received. This is a textbook case of a well-built product that missed its market. A global pandemic didn’t help…

Selling to Government: It’s Not About the Product…

Governments are one of the hardest customers to win, regardless of how great your product may be. The complexities of procurement, bureaucracy, and entrenched relationships make breaking into this market incredibly difficult.

Case Study: Palantir Technologies

Palantir entered the government sector with cutting-edge data analytics technology but struggled early on. Government procurement processes and long-standing contracts with incumbents proved to be significant barriers. Eventually, Palantir adjusted its strategy by hiring experts who understood the intricacies of government sales and aligning its product to solve specific issues for intelligence agencies, which enabled its eventual success.

Incumbents: The Power of Market Entrenchment

Incumbents can be unseated, but it's harder in some industries than others. Established players benefit from long-standing customer relationships, brand recognition, and capital that allows them to outlast newer competitors.

Example: Microsoft Office vs. Google Docs

For years, Microsoft Office dominated the productivity software market. Google Docs entered with a 100% free and collaborative platform—arguably better suited to the cloud era than Microsoft’s desktop applications. Yet, it took years for Google Docs to gain significant traction in enterprise settings, largely because Microsoft had deep integration into corporate IT systems and a loyal user base. While Google eventually found its niche, Microsoft’s market entrenchment shows how incumbents can hold off challengers for long periods, even when those challengers offer better solutions.

The Innovator’s Dilemma: Why Incumbents Fail

The Innovator’s Dilemma, as coined by Clayton Christensen, explains how successful companies can lose market leadership by focusing too much on current customer needs, ignoring the potential of disruptive innovations. Startups can exploit these gaps by offering new solutions to underserved segments or by leveraging emerging technologies. The quintessential example is Blockbuster and Netflix.

An Unseated Incumbent: Kodak

One of the most striking examples of an incumbent getting completely unseated is the fall of Kodak. For decades, Kodak dominated the photography industry with its film-based cameras and was synonymous with capturing life’s moments. But when digital photography started gaining traction, Kodak didn’t move quickly enough. Ironically, Kodak had developed the first digital camera in the 1970s but failed to embrace the technology, fearing it would cannibalize its lucrative film business. Meanwhile, companies like Canon and Sony capitalized on the shift, focusing on digital innovation and ultimately taking Kodak's market share. By the time Kodak pivoted to digital, it was too late. The company filed for bankruptcy in 2012, unable to keep up with the technological revolution it had initially pioneered but failed to embrace. Kodak’s downfall is a classic example of an incumbent failing to adapt to disruptive technology, leaving room for competitors to step in and take over

Strategies for Startup Success

1. Identify Real Problems: Focus on tangible pain points that customers are willing to pay to solve.

2. Understand the Market: Take the time to thoroughly understand market dynamics, customer behavior, and competitive landscapes.

3. Tailor Go-to-Market Strategies: The best products still need the right sales and marketing approach to succeed.

4. Be Prepared to Educate: Especially when introducing novel technology, educating the market is often necessary.

5. Adapt and Iterate: Flexibility is key. Startups that can adapt quickly to feedback will stand a better chance of succeeding.

Conclusion

Having a better product doesn’t guarantee business success. Superior technology must be paired with a deep understanding of customer needs, effective market strategies, and the ability to exploit gaps where incumbents are vulnerable. At Hawkhill Ventures, we believe in backing founders who have the vision to create groundbreaking products, the market savvy to find an edge and the ability to execute against that backdrop.

A massive congratulations to Paul and Julia who got married this past weekend. An incredible celebration for an incredible couple. Very grateful to have been included and wishing them an amazing next chapter. Here is a blurry photo of them dancing up a storm.

I don’t blame them. Pivoting is hell, and it’s better to avoid wasting time on an idea that’s flawed from the beginning.

However, our world is full of live players and markets are efficient, which means it’s exceedingly rare for a perfect idea to exist, unexploited. If you’re going to build a business, you’ll have to take some risks. You shouldn’t be looking for the perfect idea, but the perfect opportunity.

A perfect opportunity is a riskier version of the perfect idea. The market is promising, but not yet validated. There are other players, but nobody is winning. And most importantly, the solution is defensible in some way.

Markets make or break companies. In Brett’s words, “some markets are far more fertile than others. An ideal market is one that has enough buyers with an urgent problem to drive significant initial revenue, but ideally not enough that it's a saturated market.” Identifying fertile markets requires vision and unique insight. You have to see how the world is going to change, and you have to be early. This often is a contrarian take! If VCs are doubting your market size, you’re on the right track.

A defensible solution is the second most important. When building a company, defensibility justifies investment. Easily replicable ideas lead to weak margins. There needs to be some staying power. I often refer to the '7 Powers' framework of defensibility. Ensuring that your solution has some sort of economy of scale, network effect, or high switching costs will give your company a chance to establish an enduring business.

Finally, there needs to be some sort of existing competition. But why are founders so afraid of competition? I blame Peter Thiel and his book Zero to One. This startup gospel has been repeatedly misinterpreted by founders everywhere. He preaches that creating a new market is better than competing in an existing one. Often, people believe that the mere existence of competition makes a market less compelling, when in fact, it’s a sign of opportunity. Existing competitors offer market validation, especially when they aren’t executing to the fullest of their ability. Some of my biggest customers have come from the public logos of our biggest competitors.

Investors often say that great founders have “vision.” But what does that really mean? To me, it’s the ability to look at an almost-right idea, and frame it as the perfect opportunity for innovation. This vision isn't just seeing things as they are, but as they could be, turning the overlooked into the highly sought after. Successful early-stage founders need to prove that they can bridge the gap from perfect opportunity to perfect idea. So, stop searching for perfection, and start seeking opportunities where others see none.