Every time I come back to the Bay from a business trip, it feels like going back to the future.

Not because of the demos or the new models or the agents running in the background. That part is expected now. Everyone has something moving faster than it used to. Work that took hours compresses into minutes. The tools are genuinely good and getting better everyday.

But the more work trips I make away, the more I notice something running underneath it all when I return.

The pace is faster. The direction is less clear.

There’s a kind of reactive energy in the air right now. People moving quickly, but mostly in response to what’s coming at them. It’s very inbox-driven, feed-driven, tool-driven. I thought about that scene from Alice in Wonderland, the Queen’s Red Race, where you have to keep running faster and faster just to stay in the same place. That’s what it feels like. Everybody running. Nobody going anywhere new.

A GP friend put it well in a conversation we had today. His team had been pushing on something internally: the difference between being reactive to everything coming in versus deciding what you actually want to do. Not that they were making bad decisions. But they questioned if they were making optimized ones. Too much motion, not enough intention.

The instinct right now is to meet speed with more speed. But I don’t think that’s where the real edge is.

I kept coming back to a Google commercial running during the NBA Playoffs. The one about Luka. They break down his step-back and land on a single line: the brake is better than the gas.

It’s not just how fast he can go. It’s how well he can slow down.

That’s what creates the separation. Defenders are prepared for acceleration. What they get instead is a pause, a shift, a moment of control that opens up space. Luka plays at his speed. And his deceleration is as much a weapon as anything else he does.

That translates more than we admit.

In a moment where everything is speeding up, slowing down deliberately can be an advantage. Deciding what matters before engaging with everything that doesn’t. Being intentional about who you spend time with instead of defaulting to whoever is in front of you. Taking the timeout instead of pushing through the run.

There’s real pressure to keep up appearances right now. To show you’re using the AI tools, seeing the deals, in the flow of everything. And there’s a fear that if you slow down, even for a moment, you’ll miss something.

But if everything is moving faster, there will always be another wave.

You don’t have to catch every one. You have to know which ones matter.

That’s the part the tools don’t solve. They can help you move, process, execute. They can’t decide what game you’re playing or what style you want to play it.

That still comes from judgment and taste. And in a market where everyone is getting faster, the people who know when to take the step-back may end up with the real edge.

Can you make your deceleration an advantage? I am.

with gratitude,
earn

There is a long-running argument in comic book circles about whether Batman can actually compete with Superman. Superman has the powers. Batman has the tools, the money, and the mind. The debate always comes down to one variable: prep time. Given enough of it, the argument goes, Batman can beat anyone. He studies the opponent, builds the right countermeasure, and shows up exactly prepared for the fight he expects.

The lesson people take from that is about tools. I think it’s about something else. Batman’s tools only work because he already knows exactly who he is and what he’s trying to do. Without that, the gadgets are just expensive.

That is the conversation venture is having right now, whether it knows it or not.

Read more

I remember pausing on that. Not because I didn’t understand what he meant, but because I didn’t understand why it mattered. If we believed in our analysis, why did it matter whether we lined up with Citigroup or Goldman? Was the goal to be right, or was it to be close enough to everyone else that we wouldn’t stand out?

Over time, I understood the game. Public markets reward being directionally right, but they punish being wrong and alone. Consensus is a reference point. It helps anchor expectations. It protects you, in some ways, from being too early or too far off. There is a place for it.

But that place was never at the beginning.

What struck me this week is how much that same instinct has crept into the earliest stages of venture. Pre-seed, seed, and my beloved Series A used to be the part of the market where consensus didn’t matter much because there wasn’t enough information for it to matter. Someone met a founder, saw something others didn’t, and made a bet. It was closer to: we think this works, we’re going in, we’ll see what happens. That was the job.

The early stage was about being early and being right. Not perfectly right, but directionally right about a founder, a market, or a shift that hadn’t fully revealed itself yet. You were trying to see around the corner.

Now it feels like that sequence has flipped. Instead of conviction first and validation later, validation is driving conviction. Investors are circling deals because other investors are circling. Founders signaling who else is in as a primary input into the decision. Rounds that start to take shape only after a certain kind of social proof appears.

A strange version of game theory has taken hold. Not the kind where independent actors make decisions, and the market converges over time, but an inverted version where everyone is trying to anticipate everyone else and move together. The goal is not to be right, but to not be wrong alone.

That shift sounds small. I don’t think it is.

If the incentive at the earliest stage becomes “don’t be wrong,” the system starts to select for a different kind of founder. A different kind of company. A different kind of outcome. You get founders who are easier to explain, not necessarily more interesting. Companies that fit into existing narratives instead of stretching them. Rounds that fill up quickly because everyone is reacting to the same signal at the same time.

One example came up recently that stuck with me. A company gets one real term sheet, and suddenly there are more than twenty within two days. Not because twenty investors independently arrived at the same conclusion, but because the presence of that first signal changed how everyone else processed the opportunity.

That’s not conviction. That’s coordination.

And underneath it, there are other agendas running. Not just wanting to win, but wanting to say you saw it. Wanting to be associated with it if it works. Wanting to avoid the feeling of having passed. It pulls the decision away from the more honest question of whether you actually believe in the company.

When coordination shows up too early, it reduces exploration. Venture has always worked, at its best, by funding multiple paths at once. Different teams, different approaches, different assumptions about how a market might evolve. Most of those paths don’t work. A few do, and those few carry the outcomes. When consensus compresses those paths too early, we don’t just change who gets funded. We change what gets discovered.

There is also a structural problem that is hard to ignore. At pre-seed and seed, there is very little data to anchor to. For years, the conventional wisdom was that the founder and the team carried the vast majority of the decision at that stage. You were underwriting a person and a market direction. That is inherently a conviction-driven exercise. If that starts to drift toward 50/50 with what others think, you are introducing a proxy for validation into a moment that is supposed to precede it.

None of this is to say consensus is inherently bad. It has a role. As companies mature, as data accumulates, as financial performance becomes legible, it makes sense for the market to converge on a view. The issue is not that consensus exists. It is that it has moved upstream.

I still think about that moment in equity research. Checking consensus made sense there because the job was, in part, to understand how your view compared to the market’s expectations. But if that had been the starting point, if we had built the model by asking what everyone else thought first, the work would have been different. Safer. And less valuable.

Early stage was never about agreeing early. It was about disagreeing early and being right later.

There used to be a kind of purity to that. It was hard, uncertain, and often wrong, but it was honest. You were making a call based on what you saw, not what everyone else confirmed.

Early-stage venture is starting to feel like it is checking consensus before it has even built the model.

And if that becomes the norm, it is worth asking what kind of outcomes we are quietly optimizing for.

~earn

There is a scene in the show Paradise where the president walks through a bunker built to survive the end of the world. It is overbuilt. Redundant. Thought through at every level. And he stops, looks around at all the preparation, and says something that has been following me around ever since.

Empires do not fall because they were not prepared. They fall because they believed nothing could touch them.

He runs through the list. The Romans. The Aztecs. The Bulls. Six championships in eight years. Two three-peats. A team so dominant it felt structural, like it was written into the rules of the sport. And then the window closed, the way windows always do, and they have not been back to the Finals since 1998.

That monologue resonated with me this week because it does not just describe history. It describes something I keep hearing in rooms across this industry right now.

****

Over the last few weeks I have been in a lot of conversations with founders, seed investors, LPs, and GPs. Different cities, different strategies, different fund sizes. But the same undertow in almost every conversation. There is a growing fear around kingmaker checks. What happens to your portfolio company when its main competitor raises from a multi-stage firm with billions behind it? Is that a death sentence? How do you advise your founders when it happens? And what if your founders are deciding between a mega fund and a Series A fund, do you tell them the brand matters? Do you tell them it does not? And do you believe whatever you say?

Underneath all of it is an assumption that has been baked into how this industry operates for a long time. That capital, at sufficient scale and from the right source, can still crown a king.

I think that assumption is breaking. Not slowly. Structurally.

The reason is mechanical, not philosophical.

Read more

I spent most of this week in San Francisco. Coffees, dinners, walks, side conversations in lobbies and on the street between meetings. Founders and investors moving very fast, a lot of people with a lot going on. And the more time I spent in those rooms, the more I kept noticing the same thing underneath the motion.

Most people could not tell me what they were moving toward or where they thought the world was going. They could tell me what they were running from. The bad vintage. The missed window. The LPs who passed. The competitor who raised. The peer investor or founder who closed funding. The market that was supposed to open up and did not. They were moving with urgency, genuinely, not performing it. But the urgency was borrowed. It belonged to someone else’s timeline, someone else’s scoreboard, someone else’s definition of late.

I have felt that pull myself this year, the end of Q1 arriving faster than expected, and realizing that what I was calling preparation was sometimes just perfectionism in nicer clothes. I thought adulthood had taken most of the perfectionism out of me. It had not. It just taught it to dress better.

Which brings me to something that I have been following for the last few days.

A clip circulated from the Founders podcast featuring Marc Andreessen making the case that great men throughout history had little to no introspection.

Read more

I have been in enough rooms this year to notice a pattern. You walk in, the credentials are impeccable, the conviction language is fluent, and by the third sentence you realize you have heard this exact pitch before. Not a similar one. This one. Same thesis framing, same market sizing approach, same founder archetype they claim to prioritize, same language about adding value beyond the check. Different city, and let’s be honest it is the same two cities, different logo on the deck, different person across the table.

It reminded me of something I said to a friend recently when we were comparing notes on the market. Venture has become franchised. You have got a different Chick-fil-A on every corner, but it is still Chick-fil-A. Same menu. Same training. Same talking points. And the franchise model works beautifully when the market rewards consistency. When interest rates are near zero, multiples are expanding, and the rising tide floats every fund above its entry price, you do not actually need a distinct vision. You need a playbook. And for about thirteen years, that was enough.

That era is over. And the people who built their entire operating identity inside of it are now running around reacting to every new signal like a herd of wildebeest the moment one of them smells something. One nose goes up and suddenly everybody is pivoting to physical AI, and do we even all mean the same thing by that. Everybody has an agent infrastructure thesis. Everybody is at the same conference talking about the same three companies. The herd does not know it is a herd. Each individual animal thinks it is being decisive.

What I actually want to talk about is what the job requires now that the playbook era is over. Because I think a lot of people in venture, across roles, are still trying to find the new version of the old playbook. And I do not think that is what this moment is asking for.

The Specialist Is Not Enough Anymore

For a long time, we exalted the specialist. And it made sense. Information was hard to find. If you had seen twenty SaaS deals at a specific price point in a specific vertical, that knowledge was genuinely scarce. You earned something. LPs paid for access to it. Founders sought it out.

Now, anyone with the right tools and a clear question can synthesize that landscape in an afternoon. The specialist still matters, but the bar for what makes specialization valuable has moved. Knowing the space is table stakes. The edge lives in what you do with that knowledge in the actual relationship, in the room, under pressure, when the information is ambiguous, and the founder is stressed, and the board is split.

What I keep coming back to is a different kind of person entirely. Not the analyst or the specialist but the journalist. The person who is genuinely curious across domains, who can travel to unfamiliar territory and come back with something useful, who makes connections that a specialist would miss because a specialist is looking for confirmation within their frame. I grew up code-switching. Moving between different social worlds, different expectations, different registers. For a long time, that felt like a liability, like I was not deep enough in any one thing. I am now watching that skill become one of the more valuable things a person can carry into this market. Because the founders worth finding are found in all kinds of places. The patterns worth seeing cut across sectors. And the LPs worth building real relationships with are not particularly impressed by sector fluency anymore. They are trying to figure out whether they trust your judgment.

Track Records and the Era They Came From

A lot of the numbers that get cited in fund marketing right now were built in a zero-rate environment where leverage was cheap, multiples were expanding, and the biggest risk in most deals was not moving fast enough. That is a different game than what we are playing now. I am not saying those numbers are fraudulent. I am saying they were produced under conditions that no longer exist, and evaluating a manager purely on that basis is like judging a quarterback’s arm strength based on throws he made in a dome when you need him to perform outdoors in January.

There is a version of this playing out in the engineering world too that I think is more relevant to venture than people realize. A lot of experienced engineers right now feel like they have one last big payday in front of them before AI compresses their leverage permanently. They did not see the copilot and LLM solutions coming. So they are making moves, chasing the biggest contract available, optimizing for the near term, not necessarily building toward something. I see the same energy in parts of the GP community. People who built their identity and their track record inside a specific era are making decisions that are more about protecting their position than building something new. The number looks right on paper. The orientation is off.

At the same time, capital is consolidating in ways that create real pressure nobody wants to say out loud in a pitch meeting. The biggest allocators are making fewer, larger bets on a shorter list of names. Not because the emerging manager market got worse, but because when your career risk is concentrated in a single CIO tenure, the path of least resistance is writing a check to a name that will not raise eyebrows. One LP told a fund manager friend of mine that it was genuinely hard to differentiate between managers right now. I understand why she said it. I also think it is the single most important problem to solve if you are building a new firm, because the answer cannot just be better branding. It has to be something people actually feel when they are in the room with you.

The Job Looks More Like Running a Team Now

The venture game used to feel more like poker. Here are the metrics, here are the comps, here are the odds, make a decision. It was analytical. It was defensible. It was replicable. You could write a memo about why you did it.

Now it feels like running the Indiana Pacers. You have a first-round pick, and you are underwriting toward an All-Star regardless of where in the draft you are selecting. But the analysis looks completely different depending on whether you are looking at a player who has four years of college tape or a teenage phenom who came out of an international league we barely track with no documented history and a lot of potential. You are not applying the same rubric. You are asking a different set of questions for each person, and the constant underneath all of it is: can this person become exceptional at the level that matters, and do I have the judgment to see it before everyone else does?

That requires something the checklist era never really tested. We can no longer lean on a hard ARR threshold as the primary signal. TAM has never felt less useful as a stand-alone metric. What I look for now alongside the fundamentals is harder to put in a spreadsheet. Learning velocity. Proximity to customers. Ability to attract talent. Resilience under ambiguous conditions. Those are not metrics you calculate. They are things you observe over time, in how someone reacts when things break, in whether they are telling you what they think you want to hear or what they actually believe.

Surviving Is the New Thriving

The firms that are not rattled by this moment are worth studying. You feel it immediately when you are around them. They are not running toward whatever is generating the most hype. They have a clear point of view, a consistent way of engaging founders, and a real thesis about why they are positioned to win in this specific environment. You know it when you feel it. They are not performing stability. They have it. The firms that thrive in chaos built something durable enough to absorb it.

The firms that are struggling tend to share something. They were built during the optimization era. Smooth processes. Clean positioning. Respectable thesis language. Nothing that would make a sophisticated LP uncomfortable. Nothing that would cause a founder to remember them three months later. When things were going up and to the right, those firms could move fast and call it conviction. Now that conditions have changed, the absence of a real point of view is suddenly visible.

Surviving in this stage of venture is thriving. I genuinely believe that. Not as consolation. As a fact. The cockroach quality, the willingness to keep going when the environment is hostile and the feedback is slow and the outcome is genuinely uncertain, that is the thing. A lot of talented people have left or are leaving because the alternatives are good and the carry checks have not come. The ones who stay, who keep doing the work and keep building relationships and keep finding the right founders in unexpected places, they are the ones who will define what this industry becomes.

One Idea Running Through All of It

The specialist, the track record, the GM analogy, the cockroach era are not four separate observations. They are the same observation from four different angles. The playbook era rewarded a specific kind of person. Someone who could execute a repeatable process cleanly, pattern match against known comps, and present a defensible thesis in familiar language. That person thrived for over thirteen years. That person is now mismatched to this moment, not because they lack talent, but because the moment is asking for something different. It is asking for judgment that cannot be systematized, relationships that cannot be automated, and a point of view that was formed through real experience rather than assembled from what everyone else is saying. The people who have that, who have been building it while the market rewarded the playbook version, are about to find that the ground has shifted in their direction.

This week, Santosh and I are recording the first episodes of Carry On. The whole premise is the stuff people do in venture but do not say in public. The X’s and O’s of how this business actually works, what LPs are observing, what founders are signaling they need, and what it costs to build a firm with integrity and a real point of view in an environment that has rewarded conformity.

If there are questions or thoughts you have always wanted someone to answer honestly about how venture works or does not work, send them our way. We want to start with the things people already know but nobody is saying out loud.

With gratitude,
earn

He did not hesitate and gave me three things.
Be curious about everything.
Be nice to everyone.
And schlep.

Schlep was the Sesame Street answer. It was not like the others. And it is the one word on that list you cannot fake.

We are in a moment where information is cheap and getting cheaper. Two people looking at the same market can now produce a decent synthesis in minutes (or their AI model can). That is no longer the edge. The edge is in what you do after the summary runs out. It is in getting on the plane. Going across the country or globe when everyone else is waiting for the deck. Asking one more question. Showing up somewhere you were not expected and paying attention.

That is schlepping. And fewer people are willing to do it because our screens convince us that access to information is the same as doing the work.

The kindness part is easy to underestimate too. Not the absent-minded politeness of glancing up from your phone when someone walks in. Real presence. Learning names. Looking people in the eye when they have nothing obvious to offer you. The allocator reiterated it costs you nothing to be kind and it compounds quietly over a career. I believe him.

Curiosity keeps you open. Kindness keeps you grounded. Schlep keeps you honest.

I have been trying to articulate why I want to build the most analog firm in a digital age. This one conversation this week got me closer than a year of thinking about it. The framework is not new. The willingness to actually live by it is what is getting rare.

With gratitude,
earn

This came out of a prep conversation for Swimming with Allocators. New episodes are out every Wednesday on Apple Podcasts, Spotify, and YouTube.

A few weeks ago, a friend of mine (who is an executive coach to tech CEOs and fund managers) described a conversation he had with his dermatologist. Not a venture capitalist. Not a software engineer. Not someone who doom-scrolls tech Twitter for sport. Instead, it was an accomplished, mid-career, probably mid-forties dermatologist. The kind of professional whose credentials represent exactly what a stable, well-built life in America is supposed to look like.

And this doctor was visibly shaken. He told my friend he might need to start thinking about a new career.

This anecdote confirmed something I had been sensing for a while. The anxiety around AI is no longer contained to founders, early-stage investors, or the chronically online. It has bled into the bloodstream. It is sitting in dermatology waiting rooms, at dinner tables, in text threads between old college friends, trying to figure out what the next ten years of their careers actually look like.

And you can understand why. Over the past couple of years, a very specific narrative has been on repeat. AI will replace knowledge work. One person will build and operate a billion-dollar company. And then data continues to support the emerging story, with companies like Block announcing it is cutting 40% of its workforce, citing AI. Then, a research firm called Citrini publishes a thought exercise written as a dispatch from June 2028, framed as a macro memo from the future describing a full-blown Global Intelligence Crisis. Unemployment at 10.2%, with the S&P down 38%, and a vicious economic downturn feedback loop with no natural brake. It was explicitly labeled a scenario, not a prediction. But the markets sold off anyway, and people screenshot the headlines and started treating a mental exercise as a prophecy.

We live in a hyperbolic culture right now. Everything is breaking news. Everything is existential. And we have been trained, particularly over the last several years, to react faster than we reason. The Citrini piece is genuinely worth reading because it is a serious attempt to model a left-tail risk and think through the second and third-order consequences of AI disruption hitting the real economy. But confusing a stress test with a forecast is a category error. And a lot of people are making it.

So I did what I do, what any trained equity research analyst would do. I read broadly. I took notes. I talked to people. And what I found was not a clean answer, because there is not one. What I found was a much more useful picture than either the doomers or the optimists want to admit.

Greg Ip, the chief economics commentator at the Wall Street Journal, pushed back directly on the apocalyptic case. His argument is grounded in history. Technology has always displaced some jobs. But it also lowers costs, raises productivity, creates demand, and generates new roles. The offsets are consistent enough that, across the sweep of American history, technological change has not raised unemployment for the economy as a whole. He points out that software developer jobs are actually up five percent year over year as of early 2026. Ip also mentioned that the number of accountants and financial analysts rose when Lotus 1-2-3 and Excel arrived, even though everyone predicted spreadsheet software would eliminate those roles. Another example mentioned in the article was that since Google Translate launched in 2006, the number of human translators employed in the United States has risen by 73 percent. Ip’s argument is not that AI cannot disrupt; it clearly can and will. He argues that a sector disruption becoming an economy-wide catastrophe requires a breakdown in how the market economy functions, and there is no evidence of that happening right now.

The Economist offered a different kind of calibration. Their piece, titled “The AI Productivity Boom Is Not Here Yet,” points out something that gets lost in both the fear and the hype. The tools are improving fast. The adoption is real and rising. But the macro data has not moved in any dramatic way. Roughly 41 percent of American workers used generative AI at work by late 2025, up from 31 percent a year earlier. But only about 13 percent use it every day. AI accounts for less than six percent of total work hours. And when you strip out the enormous investment surge in data center infrastructure, underlying productivity gains are close to zero. The article draws on a historical parallel worth sitting with: when factories replaced steam engines with electric motors, productivity barely moved at first. The real gains came decades later, once the factories were physically redesigned around the new reality. We are still in the early reorganization phase of AI. The technology is here. The rewiring has barely begun.

What I take from reading these three pieces together is: things do not change as fast as we fear in a year, but they change faster than we expect in five or ten.

The Citrini scenario is not a prediction. It is a warning about what an unmanaged transition could look like if we get complacent. The Ip piece is not a reason to relax. It is a reason to stop catastrophizing and start thinking clearly. The Economist piece is not reassuring. It is a reminder that the real impact of a new technology is felt when organizations and workflows are redesigned around it, and that phase is coming.

Here is what I think gets lost in most of this discourse. The macro debate is interesting. But it is not actually the question most people I know are wrestling with. The question they are wrestling with is much more personal. It sounds like: what do I do with all of this? Am I going to be okay? Do I have any agency here, or is this just happening to me?

And I see two failure modes among people I respect. The first is the person who dismisses AI entirely. They are not paying attention, not experimenting, not updating their approach, and are going to get quietly outcompeted by someone who is. The second is the person who has convinced themselves that none of their skills matter anymore, that the tools will outpace them no matter what, and who has essentially decided in advance that they have already lost. That person is also going to struggle, but for a different reason. They have surrendered agency before the game is even over. Actually, we are still in the first quarter.

Both of those postures are wrong. And they share a common flaw: they are passive.

I keep thinking about language around fear. I am being careful here because I do not want to overstate it, but some theologians argue that what is often translated as fear of God is, in the original Hebrew and Greek, closer to reverence. Not panic. Not paralysis. Not the feeling that something is about to crush you. A recognition of something larger than yourself, and a call to orient wisely in response to it. That is a much better posture for AI than either dismissal or dread. Respect it. Learn it. Use it. Do not be naive about what it can do. But do not surrender to it either.

The thing I notice when I actually use these tools is that they do not replace what is most native to how I work. I am a connector. I connect people, ideas, capital, and patterns. AI does not disrupt that. It helps me do more of it, faster, and sometimes more clearly. It lets me spend less time on the blank page and more time shaping what is actually mine to say. The question is not whether AI can do more cognitive tasks. It clearly can. The question is whether you use that as a reason to outsource your thinking, or as a reason to deepen the parts of your work that are genuinely yours.

What I want for the people in my network, particularly my cohort in that 35 to 45 range who are mid-career, who built their expertise in a world that rewarded knowing things, is for them to understand that the old premium on simply possessing information is narrowing. The new premium is on what you do with it. On taste. On judgment. On trust. On the ability to read a room, persuade a human, understand context, and make meaning out of chaos. Those things are not going away. But they need to be activated, not dormant.

No five-step plan. Just one thing.

The piece I found most useful, practically, was not the macro analysis at all. It was a letter that Jim VandeHei, the CEO of Axios, wrote to his wife and kids. His core message is simple: you are living history. Do not ignore it. Use AI with curiosity, discernment, and clear eyes. He told his kids that any knowledge work that does not require true expertise or a critical human connection is at real risk. He said treat the tools like the smartest person you have ever met, but remember they are imperfect, probably like the smartest person you know today. And he ended with a call to action that really resonated with me: There is no hero riding to the rescue. You be the hero.

I want to say something like that to my generation. Not to his kids, but to the adults I came up with who are building companies, raising families, managing teams, and trying to figure out what the next two decades of their careers look like. The letter he wrote was for people starting out. But the message applies to all of us.

You did not ask for this moment. It is here anyway. The data is genuinely uncertain about how fast this will move and how wide the disruption will be. Reasonable economists disagree. History offers useful but imperfect analogies. The full productivity impact has not yet shown up in the macro numbers. There is real time to orient. And yes, black swan events could happen more and more.

But the right response to uncertainty is not paralysis. It is engagement.

Figure out your superpower. It’s there. I promise. What is the thing that, when you are doing it, you feel most like yourself and most useful to other people? What is the topic someone could call you about on a random Tuesday with no prep, and you would have something really helpful to say? That is what you protect. That is what you build around. That is what AI should amplify, not hollow out.

Experiment with the tools. Not performatively, but genuinely. Find out what they are good at and where they fall short. Use them to go faster at the things that already come naturally to you. Do not use them as a substitute for thinking. Use them as a pressure test for your ideas.

And maybe most importantly: do not mistake reading about AI for engaging with AI. The discourse is not the same as the practice. One of them will make you more anxious. The other will make you more capable.

The Citrini piece that started the recent anxiety attacks on the market and group text threads even ended with a bit of realist optimism similar to VanDeHei's advice, reminding us all that “We are reading this in March 2026, not June 2028…The canary is still alive.” The feedback loops described in the Citrini piece have not started. The reorganization The Economist describes has barely begun. There is still time to shape what this transition looks like, not just for yourself, but for the teams you lead, the companies you build, and the family, friends, and people coming up behind you.

So be engaged.

With gratitude AND agency,
Earnest

Articles referenced in this essay:

• Citrini Research: The 2028 Global Intelligence Crisis

• Wall Street Journal (Greg Ip): Tech Has Never Caused a Job Apocalypse. Don’t Bet on It Now.

• The Economist: The AI Productivity Boom Is Not Here Yet

• Axios (Jim VandeHei): Blunt AI Talk in Letter to His Kids

* * *

Twelve days ago, my paternal grandfather, Earnest Sweat Jr., passed away. Yesterday, an “Earnest Sweat” Google alert pinged my inbox with his obituary. I clicked it, started reading, and something about the math stopped me cold.

Read more

About halfway through the conversation, one of the co-founders said something that I have not been able to shake. The hardest part of design has never been execution. It is deciding what it should be. AI, right now, is largely pattern recognition. It can generate, remix, and approximate your preferences. But the question of what a thing needs to communicate, and what it needs to feel like to someone encountering you for the first time, that is a judgment call. And judgment calls require someone who has actually lived something.

Because of how my mind works, that concept sent my mind in three directions at once.

Three Reference Points

The first was a documentary called Exit Through the Gift Shop. It was supposed to be about Banksy. Instead, it became a portrait of Mr. Brainwash, a man who rode proximity to real artists and an instinct for replication into something that looked, at a distance, like a career. Whether he was a fraud, a genius, or just a mirror held up to an art world that had already lost the plot is a question the film never answers, which I think is the point. He forces you to ask whether the system itself is complicit in elevating replication. And in doing that, even he becomes useful.

The second was Andy Warhol. The first time I really encountered his work was in my (required) Art Humanities course at Columbia, and I remember sitting there thinking, this is not art in the way I have been taught to recognize it. He was not grinding pigments or obsessing over perspective. He was selecting, repeating, and amplifying. At the time, I thought maybe he was just vampiring culture, attaching himself to whatever was already popular. But decades later, the soup cans still stand. The Marilyns still stare back at you. He was not just copying. He was diagnosing something about celebrity and consumerism that most of society had not yet learned to see.

The third was Nora Ephron. There is a documentary about her life called Everything Is Copy, which was her thesis about how to live and work. Her argument was simple: life is material. Conversation, heartbreak, awkward dinners, and even betrayal. Nothing is wasted. Everything can be transformed into something that makes another person feel less alone.

Which leads to the question that connects all three. If AI can remix everything documented on the internet, and Nora Ephron remixed everything that happened to her, what is actually the difference? That is when I pulled Baldwin off the shelf.

Baldwin Sets the Bar

That is the line between slop and soul.

Baldwin is not talking about style or technical craft. He is talking about extraction. Art is not the rearrangement of available inputs. It is the relentless forcing of meaning from lived experience, taking the disorder of a life actually lived and compressing it into something ordered and intentional, something that carries weight precisely because you can feel the pressure behind it.

AI can remix what has been documented. It can recombine what has been digitized. It can produce something that looks like design or reads like prose. But it cannot suffer. It cannot metabolize. It cannot extract the last drop, sweet or bitter, from a life it has actually lived. That is not a limitation that gets patched in the next model update. It is a categorical difference.

The Archive Was Never Neutral

There was a second thread in my conversation with Civilization that also resonated with me. The researcher and strategist at the firm was explaining a design library one of the co-founders has been building, focused on the history of modernist design with particular attention to resistance and protest movements, documenting the visual culture of communities that were never centered in the dominant archive. Her observation was pointed: pattern recognition can only work with what has been documented, and what has been documented is not evenly distributed.

Everything AI generates is made of what was put into it. The likelihood of any specific design choice is weighted by how often something similar appeared in the training data. Which means that generating from the dominant archive tends to reproduce dominant patterns, dominant aesthetics, and dominant assumptions about what looks credible or innovative or trustworthy. The bias is not always visible, but it is always there.

This matters beyond design. It matters for anyone trying to signal something genuine in a marketplace about to be flooded with content drawn from the same underlying archive. When the floor of acceptable output rises for everyone simultaneously, the thing that differentiates you is no longer access to tools. It is provenance. The undocumented story. The perspective that was never in the training data because it came from somewhere the algorithm did not know to look.

What the Premium Shifts To

AI will make the average acceptable, and that is already happening. A competent pitch deck, a functional brand, a passable strategy memo, all of these are now within reach for almost anyone with a subscription and a good prompt. The floor has risen and it will keep rising.

But when the floor rises, the premium shifts. Right now, I think it is moving toward three things that cannot be generated: provenance, conviction, and coherence. Provenance asks what it actually cost you to arrive at an idea, what you have seen and survived that gives your perspective weight. Conviction is the difference between someone who decided what something should be and someone who generated options and picked the one with the best metrics. Coherence is the quality of something that feels integrated, like it came from a sensibility that has been tested against reality rather than assembled from available parts.

In venture, I see an enormous amount of output right now. Pattern matching, hot takes, this-looks-like-the-next-X analysis. We are very good at recombination. The firms and founders that endure are not just recombining. They are extracting from cycles they have survived, from customers they have actually sat with, from markets they misunderstood and had to relearn the hard way. That extraction is the work. And it is work that cannot be outsourced.

So maybe the future of design, and of venture, and of any work that requires judgment, is not a battle between humans and machines. Maybe it is a sorting mechanism. The average will be easier to produce. The noise will get louder. But the exceptional will become priceless, because the exceptional will carry evidence of a life behind it. It will feel like someone actually paid attention, actually suffered the question, and actually forced from their experience the last drop it could possibly give.

That is not something you can scrape from the internet or prompt.

That is something you have to survive.

With gratitude,
earn

A founder said that to me last week, and it startled me a bit. This wasn’t someone speaking from the outside looking in. He has built AI agent infrastructure twice. Once inside a hyperscaler, and a second time at a large consumer AI company. He is smart, deeply technical, and has lived inside large AI departments with seemingly unfettered access to resources most operators can only dream about. And after all of that, he came to what he described as an inevitable conclusion.

He had seen bot detection everywhere in the cloud, with APIs tightening and firewalls hardening at every turn. His consensus was blunt: enterprises are simply unwilling to punch holes in their systems so that some general-purpose agent can roam freely through their data. They have spent too much protecting it to give it away now.

The dream that has been sold to consumers, builders, and investors was seductive. Autonomous agents floating through the cloud, clicking buttons, pulling data, doing your work while you sleep. One prompt and you could lean back like George Jetson while your digital assistant handled the rest. The reality is something far more humbling. Most of those agents get blocked at the door. I have experienced it myself, trying to use large language models to do research at scale, to automate my daily workflows, to scrape, synthesize, and run multi-step processes. You hit invisible walls faster than you expect. Captchas. Rate limits. Logins. Access restrictions that were never announced but are always enforced.

The models are getting smarter. But the gates are getting stronger. And that tension matters far more than we are willing to admit publicly.

Every major technological shift produces new direct and indirect problems and opportunities. It never produces the predicted utopia or the predicted dystopia. It produces something messier, more complex, and ultimately more interesting. The closing internet is one of those moments.

From Netflix to Walled Gardens

Remember when everything was on Netflix? One login, infinite content, frictionless distribution. You could share it with a few friends and feel like you had unlocked the entire history of television. Then Disney pulled Marvel. HBO pulled its catalog. Peacock appeared. Paramount Plus. Apple TV. Everyone built their own platform because they eventually realized that distribution without ownership is a trap. The streaming dream fractured into walled gardens, and now people are paying more for less, navigating five different apps to find one show.

The internet is going through something remarkably similar. Early Web 3.0 4.0 (sorry I’ve lost count) felt genuinely open. SEO arbitrage worked. Scraping worked. APIs were generous, and growth hacks scaled beautifully. Distribution was cheap, data was accessible, and if you were willing to hustle, you could build an audience or a business on top of someone else’s platform without asking permission.

That era is over.
Data is now proprietary and central to a platform’s valuation.
Platforms are defensive. Distribution is pay-to-play.
Access is permissioned.
And AI does not eliminate this dynamic. It accelerates it.

If you are building in 2026, you are not building on an open plain. You are building inside gated communities. The sooner founders accept that reality, the sooner they can design for it instead of fighting it. And that shift in posture leads to one foundational insight that I keep coming back to: in a closing internet, trust becomes infrastructure.

So what do you watch out for? What can a founder do?

Read more

What I loved was how unromantic and disciplined Erik was about the tradeoff. This isn’t a branding exercise. It’s a daily operating system: stay professional as an investor, stay honest about the political momentum, and don’t confuse “good intentions” with durable outcomes. It reminded me of every great corporate VC seat I’ve seen. Strategic value and financial value aren’t separate lanes. They’re the same road. The difference is whether you can hold both truths at once without getting sloppy.

Somewhere in the conversation, my brain went back to why I attended Columbia in the first place. Not just New York, though that was a real perk given my love of East Coast hip hop, but the core curriculum. Being forced given the opportunity to read the canon and wrestle with Western history was my early attempt to understand why the world was the way it was before trying to predict what it would become. I doubled down on that curiosity and chose economics and political science because I couldn’t stop noticing how superpower nations and multinational companies were starting to behave like city-states, with their own incentives, alliances, and gravity. Listening to Erik talk about a country of 6 million people with a strategy in a shifting order made that whole mental model feel less academic and more alive.

And that’s the field note: even in a world getting more insular and polarizing, the opportunity is still in building bridges. When globalization strains, it doesn’t end. It decentralizes. It becomes person-to-person, firm-to-firm, state-to-state. The winners won’t just be the biggest or the loudest. They’ll be the ones who can form real partnerships, build trust across borders, and make the relationship itself a strategic asset. Find the bridge-builders. Invest in the bridge-builders. Be the bridge-builder.

Swimming with Allocators is back soon, and you can hear all of Erik’s great insights.
Catch up on past episodes here: https://www.youtube.com/@AllocatorsPod/videos

Her point was simple and heavy at the same time: the way a lot of us are operating right now is probably not healthy, and we will likely see strange downstream effects from it. Not just burnout. Not just anxiety. Real physical stuff. The kind of outcomes that make you stop using the word “lifestyle” and start using the word “risk.”

What hit me is how quickly “connecting” with others has drifted into “performing.” Adults do it all day.

Zoom calls that feel like webinars.
Group chats that replace real conversation.
Quick voice notes that keep things moving but never really let anything settle.

Even the way we FaceTime now can feel like a broadcast. I’ve watched it with my son. He’s seven, and when he calls his grandparents or aunts, half the time he’s not even catching up. He’s in webinar mode. He wants to show them tricks from his media class. He screen shares. He’s presenting.

It’s funny.
And it’s also a mirror.

Because we’re training a generation (and ourselves) to equate connection with demonstration.

It’s less of that slow back-and-forth where you learn someone’s mood. Where you sit in a little silence. Where you ask the second question after the first answer. I’m not blaming him. It’s probably generational. Going through Covid as a young child. Having access to devices early. Being raised by millennials.

But I’m hopeful, especially when I look at the success of a consumer product company a friend invested in called Tin Can. The pitch, as I understand it, is almost laughably old school, where kids ages roughly 5 to 12 want a regular phone experience again. Not a smartphone. Not another feed. Something closer to what we had with corded phones. Just calling people. Just talking. The fact that something like that is “ripping” says more than a thousand op-eds. It suggests the demand for unoptimized connections is not romantic nostalgia. It is a market signal. We miss being with people, and we miss being reachable in a way that does not require us to perform.

The LP also made a second point that connected the human health angle to the private markets angle in a way that should make every fund manager lean in.

Private markets are network-based. Trust-based. Pattern recognition-based. The comfort of home, and your social bubble, is real. But deals do not get made at home. Early stage, especially. The work is embodied. It is eye contact. It is shared context. It is the after, not just the meeting. And if you believe, like I do, that attention is the scarcest resource now, then the analog path is not a retreat from the future. It is a way to compete for attention without contributing to the fatigue of constant digital interaction. Boots on the ground, done thoughtfully, is suddenly differentiated again.

So when I say “most analog firm in a digital world,” I’m not just talking about notebooks, phone calls, and dinners. I’m talking about building a system that respects human limits and human biology. That is where good judgment lives. I’m talking about conversation as a craft, not content as a substitute. I’m talking about a life where I can do the work without frying my central nervous system. And a firm where founders feel something steady when they are around us.

If we are heading into a decade where the body keeps the score for how we lived through this screen era, then the contrarian play might be simple: do as the Gen Z cool kids say, “touch grass.” Or used to say it. I’m old enough now that the slang changes while I’m still using it. Which is basically that SNL bit in real time.

But the advice holds.

Touch grass.

Get in rooms.

Talk with people, not at them.

Build trust where the algorithm cannot follow you.

with gratitude,
e

From the outside, my day-to-day work probably doesn’t look like work at all.

It looks like conversations. A lot of coffee tea. Zoom calls without decks. Walks that run long. Dinners where the most important part happens after the plates are cleared.

Friends and family sometimes ask, politely, what I’m actually doing all day.

This past week clarified the answer for me.

If this phase of the work doesn’t feel slightly uncomfortable, you’re probably skipping something essential.

The conversations weren’t hard per se. They were hard in the honest sense. The kind that require you to say what you actually believe before it’s fully formed. The kind where boldness costs something.

That’s when the analogy clicked.

This is the part of the process stand-up comedians live in long before you ever see a special.

They work rooms.

Not theaters.
Not arenas.
Rooms.

Small clubs. Amateur nights. Inconsistent crowds. Jokes that die quietly. Jokes that almost land. Friends in the audience offering punch-ups afterward. Try this tag. Shorten that pause. Flip the ending.

That’s not performance.
That’s crafting.

Read more

Lately, I’ve been thinking a lot about talent.

Not hiring.
Not org charts.
Not comp bands.

Talent.

Specifically, whether founders still know how to see it early enough.

And whether they’re brave enough to let it be great, even when that greatness eventually moves on.

Automation is raising the floor everywhere.

Tools are faster.
Execution is cheaper.
Access is wid…

Read more

Last week, a few people sent me a speech from Davos.

I didn’t watch the video.
I read the transcript.

One line stayed with me: Nostalgia is not a strategy.

It sounds obvious. Almost trite.

But the longer I sat with it, the more it felt like a diagnosis.

Not just of geopolitics.
Of venture.
Of investing.
Of leadership.
Of this moment we’re all standing in, whether we want to name it or not.

The speech referenced an essay by Václav Havel, The Power of the Powerless.

Havel tells a story about a greengrocer who hangs a sign in his window every morning:
“Workers of the world, unite.”

The shopkeeper doesn’t believe it.
No one does.

But he puts the sign up anyway.
To avoid trouble.
To signal compliance.
To get through the day.

And because every shopkeeper on every street does the same, the system persists.

Not because it is true.
But because people keep performing it as if it were.

Havel called this living within a lie.

That phrase landed hard.

Because once you see it, you start seeing signs in a lot of windows.

In venture, we have our own versions.

We repeat sourcing and diligence playbooks that worked in 2014 and call it discipline.
We cling to pattern recognition long after the patterns have decayed.
We talk about differentiation while funding sameness.
We say the market will “come back” as if markets have memory or loyalty.

None of this is malicious.

It’s human.

Nostalgia feels like wisdom.
But a lot of the time, it’s just fear with good branding.

The hardest part about this moment is not that things are changing.

It’s that the old stories still half-work.

Brand still opens some doors.
Capital still attracts capital.
Reputation still buys time.

So it’s tempting to keep the sign in the window.

To perform confidence.
To perform certainty.
To perform continuity.

But performance is not the same as truth.

And in investing, living within a lie is expensive.

Being a great investor today requires real adaptability.

Not slogan-level adaptability.
Not “AI-first” adaptability.
But the kind that forces you to re-underwrite your own instincts.

You have to be a real stock picker again.

You have to understand why something is working, not just that it worked before.
You have to get comfortable saying “I don’t know yet” and doing the work anyway.
You have to let go of old signals that no longer predict outcomes.

That’s uncomfortable.

It’s much easier to repeat what once made you feel competent.

But the market does not reward nostalgia.
It rewards intellectual honesty and curiosity.

This is showing up everywhere I look.

In how venture firms are being built.
In how capital is concentrating.
In how trust is being allocated faster and with fewer chances.
In how founders decide who to spend time with.

The old rituals still exist.
But they don’t carry the same weight.

And pretending they do doesn’t make you principled.
It makes you late.

What struck me most about the Davos speech wasn’t its politics.
It was its posture.

A refusal to accept inevitability.
A refusal to confuse realism with resignation.

There was a line about “taking the sign out of the window.”

About naming reality instead of performing comfort.

That feels relevant far beyond geopolitics.

I feel it in venture conversations every week.
I feel it in how founders talk when the cameras are off.
I feel it in how LPs ask questions they didn’t used to ask out loud.

There is turmoil. There is uncertainty.

But I refuse to be hopeless.

Not because hope is naive.
But because hopelessness is passive.

Adaptation is an act of agency.

So is rebuilding.
So is reframing.
So is choosing to create instead of retreat.

That applies to investing.
To firm-building.
To how we show up in our communities.
To how we think about this country and world, even when it’s hard.

We are not obligated to accept decline as destiny.
We are not required to keep performing stories we no longer believe.

Taking the sign down doesn’t mean you have all the answers.

It means you’re willing to tell the truth about the moment you’re in.

And from there, to act.

Slowly.
Imperfectly.
But honestly.

Nostalgia is comforting.

But it is not a strategy.

And I’d rather live in the work of adaptation than in the performance of certainty.

With gratitude,
E

A few months ago, I kept hearing the same question in different accents.

From LPs, sometimes on the record, sometimes off.

From GPs, especially the ones with real scar tissue and a real point of view.

From operators circling the industry, wondering if the door is closing.

The question is simple:

If more capital is being gobbled up by the biggest brands in venture and growth capital, how will new firms ever break through?

It is a fair question. It is also the wrong mental model.

Because if you look at how the great franchises in venture were actually born, you start to see a pattern.

Most enduring firms did not simply “raise a fund.”

They were put in business.

Some LP, or small group of LPs or individuals, stuck their neck out. They wrote the first meaningful check. They recruited the next check. They gave the GP the oxygen to build. Not just invest.

And in this market, where attention is scarce, and patience feels like an endangered species, I think we are about to see that formation mechanism evolve.

Not disappear.

Evolve.

The chasm is real, but it is not new

Venture always had a power law. It always had brand gravity. It always had inside lanes.

What feels different now is the shape of the market.

Upstream, you have an explosion of managers at pre-seed and seed. Some are excellent. Many are loud. The ecosystem is more distributed than ever.

Downstream, you have mega-funds and growth platforms consolidating capital, talent, and mindshare. More dollars. More infrastructure. More ways to win deals.

And in the middle, you have the zone where franchises used to get built.

The Series A and Series B layer that could grow into a fund-to-firm-to-franchise over time.

That middle layer is not dead, but the on-ramp is harder.

Not because there are fewer great investors.

Because the old formation path is getting squeezed by two forces at once:

1. Capital concentration

2. Trust compression

Capital concentration is obvious. Bigger funds have more capital and more momentum.

Trust compression is subtler. Everyone is trying to decide faster. Committees want fewer unknowns. Founders want fewer meetings. LPs want fewer surprises in a market that has plenty.

When those two forces hit at the same time, “raising your first fund” becomes less like a rite of passage and more like an obstacle course.

The numbers explain why this feels different.

By 2024–2025, the top tier of venture firms captured a disproportionate share of new capital. Depending on the dataset, roughly the top 10–15 firms by fund size raised between 35–45% of all venture dollars in a given vintage, while the top 25 firms captured well over half.

At the same time, the number of active venture funds continued to grow, and median fund sizes either flattened or declined. In other words, more managers were competing for a smaller slice of discretionary capital, while brand-name platforms absorbed the majority of commitments.

This is not just capital concentration. It is formation pressure. The market is signaling that scale, trust, and distribution matter earlier than they used to.

Read more

The streaming wars did not start because Hollywood got more creative.

They started because distribution stopped being enough.

Netflix used to win by shipping DVDs. Then by being the easiest place to watch. Then by licensing everybody else’s best work. And then the day came when licensing was no longer a strategy. It was a dependency.

So Netflix did the only rational thing.

It became a studio.

That evolution has been stuck in my head because venture capital is going through a version of the same transition. Founders feel it. LPs feel it. Operators definitely feel it.

It shows up as a question I keep hearing, sometimes said directly and sometimes implied:

Why does it feel like every VC is now a social media influencer, a PR firm, a media company, or a streamer?

That question is not cynical. It is perceptive.

Narrative became table stakes. The way you do it became the moat.

A friend told me a story recently that made this painfully concrete.

He built a podcast miniseries out of spite after losing a hot deal.

The story goes: he sent a term sheet to a founder. The founder passed. Later, the founder admitted something that sounded harsh but was probably honest.

“I didn’t really see your expertise online.”

My friend has real operating experience. Real credibility. Real pull in his corner of the world.

But in a market where everybody is busy and everyone is sorting quickly, “real” is not always visible.

That is the new problem.

Narrative is no longer decoration. It is legibility. It is searchability. It is trust, compressed.

And once narrative becomes table stakes, the differentiator shifts.

Not whether you produce content.

Whether your narrative is purposeful enough to become a moat.

But you have to try to find out. The same investor friend once told me that a peer VC once told him that “the optimal level of cringe is not zero.”

The private markets are now a buyer’s market for attention.

Read more

I’ve fallen back in love with this stretch of the calendar.

Late December. Early January. Janus time.

I’ve written about it before, the Roman god with two faces, one looking back, one forward. But this year it’s hitting differently. Maybe it’s age. Maybe it’s responsibilities. Perhaps being wired as a helper for so long makes it easy to forget the experience itself, in favor of the outcomes.

At my best, people have described me as having a kind of boyish joy. Not naive. Not unserious. Just genuinely energized by the work and the people and the possibility of things turning around.

Somewhere along the way, that joy can get crowded out by scorekeeping.
Did I do enough last year?
Did I miss something?
How fast do I need to make up ground?

Culture does not help. Dopamine everywhere. Instant feedback. Everything is framed as urgent, critical, existential, and unprecedented.

And yet.

There’s something beautiful about this season. My birthday is January 9th. Capricorns are the best. I won’t apologize for that. But what I really love is that everyone becomes a Capricorn this time of year.

New year, new discipline.
New plan.
New gym membership.
Strong convictions.
Tickets already bought.
Diet started yesterday.

Read more

A friend asked me two questions this weekend that I’ve been sitting with ever since.

The first was simple. The kind of question people ask when they’re not playing around.

Are you really committed to building a venture firm?

The second question had teeth.

With so much changing in the venture world, do you genuinely believe you can deliver top-decile returns?

In my younger years, questions like that would have triggered the chip-on-the-shoulder version of me. The part that wants to prove something before it wants to explain something. I know that guy well. He has helped me in some moments.

But he’s not the best positioned to build for the long haul.

This year, I’ve been trying to approach questions like this differently: not as an attack to defend against, but as curiosity to honor. That shift has done more for my conviction, my relationships, and my clarity than almost anything else I practiced in 2025.

Because here’s what I’ve learned, slowly and repeatedly:

The quality of your outcomes is usually downstream of the quality of your questions.

A good question is like a great journal entry

A great question does what a great journal entry does.

It forces you to say what you actually believe.
It exposes gaps and assumptions.
It makes the fuzzy parts visible.

Read more