We are on the precipice of a critical, multivariate, highly uncertain future driven by uncertainty in the path - development and deployment - of AI.

And one of the most exciting things about this moment is that there are so many viable, seemingly reasonable but deeply contradictory points of view about how it resolves.

• Inference races to zero or stays expensive forever

• Models keep getting better or they plateau

• A multi-model world or consolidation to a few frontier labs

• AI will take all the jobs or create all new jobs

There are any number of such scissor statements about AI, society, business in the air right now.

No matter what you do, you are making a bet on what the future/the steady state of the technology looks like. Any company you build has to be a reflection of that bet on the future expressed through every choice you make: your product, business model, and team/culture/org chart. It’s all one choice.

So “what do you think a valuable company looks like in the future” has to be the starting point for what kind of company you build toward and how you build it from 0.

That’s especially true for team and culture, which are upstream not just of what you can build but also how you can build it. They’re the necessary precursors to the product you build and the motion you execute.

There’s not true or absolutely path dependency. Talented people can obviously do all kinds of different things. But there is a relevant optimal path for the pieces as you’ve assembled them. Which pieces you assemble, and to what end, is hugely important and highly determining.

Last week I wrote about Phoebe and “post agent” companies with Justin, who has an extremely strong and explicit POV:

The post-agent business believes and embraces that software and cognitive/computer-based labor is going to zero. And so the post-agent business counter-positions its product, go-to-market, and strategy around that which is still scarce and valuable: context, attention, trust and brand to produce sustainably differentiated outcomes via network effects […]

Engineering and systems thinking is the means by which we build the machine that runs Phoebe. Engineering becomes the highest point of leverage and every function is engineering-assisted (our head of CX is our most active Devin user) […]

I also fundamentally believe in betting on deeply AI-native, junior talent with no prior assumptions on how businesses/products are built and no constraints on the problems they want to work on.

The point here isn’t whether his bet is correct (though I generally think may well be); it’s that a clearly articulated point of view necessarily drives every other choice for the business.

This matters more for startups than for anyone else because startups are deeply and persistently disadvantaged in almost everything; they have less/no money, access, team, or time. It’s hard for them to just say “we’re going to hire the best people or do things in the best way.” Instead, they need to find and exploit advantages, which means knowing where they are advantaged. It’s necessary to have a sense of who the best people are for you, rather than merely trying to get the best people in some generic sense. Here, local > global optima.

Almost everything in startups amounts to leverage and thrust-to-weight ratios. To hire spiky people requires having a sense of what trade-offs you are willing to accept.

There are two obvious ways to spot an inauthentic commitment to vision or a long term bet.

1. Cargo culting. Not necessarily internally incoherent, but assembled based on aesthetics and rote imitation. All signs and symbols. Think: “I want to look like Cursor, so I’m going to make choices that look like the choices Cursor makes” without understanding the thinking/bets behind them. This is the most common and the easiest to spot because it’s all external.

2. Incoherence. Your team doesn’t match the GTM, or your end state doesn’t match your roadmap. This is harder to spot, because it requires interrogating each/all.

These are both company killers. You have to make a bet in order to move quickly and decisively on all other choices.

Even claiming that you don’t have a point of view is still a perspective. It carries with it that the most important thing is to maintain optionality.

That’s a totally viable approach. Believing the future is highly uncertain, and wanting to preserve your options, has implications for how you build, just like going all in on one particular POV does. It is neither riskless nor riskier than anything else, except insofar as you don’t reflect it back to yourself through all the downstream choices you make.

But optionality is never free, you pay a premium to defer choices. You have to make a choice, and every choice is expensive, including the choice to preserve optionality and defer having an explicit point of view.

If anything, startups are advantaged here: they can afford to make choices that are irrational for other people, which lends itself to spiky choices and min-maxing. Optionality is one such advantage. Startups can hold it precisely because they have no infrastructure and no sunk costs to justify... at least for a time.

So it’s not style or “taste” but YOUR style that counts.

Choose wisely.

Programming note: I’m going on my honeymoon next week. This is my first non-laptop vacations in ≈10 years.😅 The newsletter will return in mid-July.

Editor’s note: This is a co-authored post with Justin Woodbridge, the founder and CEO of Phoebe. Phoebe is building a network of agents for healthcare labor. I mostly just added jokes. We’ll be back to our regularly schedule programming next week. Go Knicks.

White Noise

Everything feels astroturfed. Last week we were token-maxxing. Now we are token-minning. We are rolling up, we are buying out. We are taking private (funding secured), and everyone is monitoring the situation. Tech is in a DeLillo novel, with the hum of manufactured, short-lived narratives getting louder every week.

Underneath it all, everyone is trying to figure out where does this all end up? What is the end game?

Each big-bang moment gives new primitives, scrambling the core axioms around which the big winners are built. PCs gave us Apple and Microsoft, Web gave us Google and Amazon, Cloud gave us Salesforce, mobile gave us Uber. Each of these technologies came with new and unique assumptions for how to build a company for the moment; business model, GTM, and org structure are artifacts of each respective paradigm shift.

So what does the agent-shaped company look like? What is the new primitive, and what assumptions does it reset?

Vertical Market Genealogy

Historically, there have been two strategies for vertical tech companies: pure SaaS (sell the tool) and full-stack (build the tool, sell the work). This current cycle has seen the emergence and popularization of two AI-inflected archetypes built on that prior art:

Sell the tool as a “system of action”

• AI and agents power tremendous experiences and value for customers. There is a great race underway to be first to serve them by localizing horizontal workflows for their needs.

• These companies look like 2010s vertical software companies in their business models, GTMs, and teams/cultures. They bolt AI features onto stuck-in-the-mud SaaS thinking and are “AI native” only insofar as they were founded after the release of ChatGPT.

Do the work as a “neofirm” or “AI rollup”

• These companies build agents to transform traditional operating businesses and own all the upside.

• Doing the work as a neofirm or rollup (taking advantage of new intelligence) relies on commodities and does not produce the scarce resource: attention, context, and reach.

As models get cheaper and competition increases, what differentiates my agent from your agent, my firm from your firm? Without differentiation and decommodification, it’s all execution to eek out advantages; that’s not the end-game, nor the strategy to be the last-and-final mover.

The trillion dollar question: what is wave 3?

Infinite, free, on-demand cognitive labor as the new primitive

At the limit, agents represent a new building block: labor that is digital, directable, and deployable on-demand, and most importantly, free (eventually, through open source and local inference).

So the end game is not building more capabilities and applications for computers. Traditional software companies are built around assembling specific capabilities for computers, and profiting off the time and expense that it takes to do so (teaching computers what you wanted them to do was hard). Increasingly intelligent models compress all the capabilities out of the box.

The mid game trap is to get addicted to the human:agent arbitrage (sell agent labor at human prices). The end game to bet on cheaper, faster, local and open source models that make intelligence free.

The post-agent company will win this marshmallow test.

The defining post-agent company isn’t going to be horizontal, that’s for the labs. The most exciting question is not what kinds of apps and software tools can you build - it’s the new business strategies and models they enable.

Instead of “teaching computers,” the new opportunities revolve around “using computers;” - directing and deploying 24/7 intelligence to do more work (raw volume), cheaper work (cost), better work (quality), and the long-tail of tasks that never made economic sense to do before.

And the most promising places to do so are huge, fragmented, physical verticals like healthcare, logistics, and construction where fractious human labor is the critical input (not computer tasks).

Neofirms are the wrong strategy in these markets. You can’t simultaneously networkmax and controlmax. One must give way to the other as the p0, totemic belief (breadth vs depth). And control in real-world, coordinate-constrained businesses is a mirage; you are sending people you can’t observe, into locations you don’t own, to do work you can’t automate.

The post-agent business believes and embraces that software and cognitive/computer-based labor is going to zero. And so the post-agent business counter-positions its product, go-to-market, and strategy around that which is still scarce and valuable: context, attention, trust and brand to produce sustainably differentiated outcomes via network effects.

Stretch the New Primitive to Its Logical Conclusion

What happens when you do the work, but take it one step further and give away most-or-all of the rote/labor/replacement-level value?

Instead of building and selling software tools, you are building digital workers to do the work?

What if instead of trying to monetize the work, you commoditize the work?

The defining opportunity is one of newly unlocked business models, strategies, highly-potent, nascent go-to-market motions.

Business models

• The post-agent company should be focused on as efficiently and swiftly as possible vacuuming up what is going to be scarce at the end-game: context, attention, trust and brand to produce sustainably differentiated outcomes via network effects.

• These companies will be asset-light attention aggregators, coordination infrastructure, and trusted clearinghouses / agent-native marketplaces. (whether monetizing through transactions, advertising, attention routing)

GTM

• The post-agent company flips the inevitable commodity-nature agents on their head and uses that as a sharp GTM strategy to penetrate quickly, efficiently, and expand from there.

• Digital workers do the work and give it away to gather context, bootstrap flywheel and build network effects. The single player mode is instantly valuable enough to attract demand and jumpstart the process.

Team

• The post agent company is deeply technical AND obsessively entrepreneurial at all levels and in all roles, transcending the iron triangle (and the “conjoined triangles of success”). To be a truly post-agent organization, a business will embrace agents in every part of its business, expecting every member to use these tools to ship and solve problems for customers.

• When intelligence is an opex line, commercially minded, entrepreneurial individuals and teams will spin up teams/pods/agents/resources on demand unbound by classical thinking about empire building and headcount.

Give away the labor, capture the network that runs through it. When your competition is built around profit from software or services, how can they respond to you doing the work for free?

The post agent company orients itself around continued, rapid improvement in the cost and performance of underlying models and intelligence. It assumes, at every level, that the future does not merely look like the past but rather sees discontinuity and a true paradigm shift, just like what PC, web, cloud, and mobile ushered in.

Phoebe

We’re putting our time and money where our mouth is with Phoebe. Phoebe is building a network of agents for in-home healthcare, the largest AND fastest growing workforce in the country:

1. digital employees for agencies to manage schedules and run their business

2. talent agents for labor to make more money from better shifts

3. care advisors for families to get the care they need

Today, Phoebe the product and Phoebe the company (Recurrence, the maker of Phoebe) are one and the same. But our real product is the engine that designs, propagates, and manages the network of agents.

As a post-agent company, we aren’t just building external workers for our customers. We build internal workers to execute all the functions of the organization. That’s why we’re building the company first and foremost around engineering; 60% of the company has a CS degree and ⅓ of the company are former founders. Everyone ships.

Engineering and systems thinking is the means by which we build the machine that runs Phoebe. Engineering becomes the highest point of leverage and every function is engineering-assisted (our head of CX is our most active Devin user)

Phoebe’s culture in practical terms

Our company runs on a stack of internal agents and tools that revolve around a central database and API (Phoebe Core). Here we fold in the raw exhaust of every system of the company: each email and call (sales, post-sales, hiring), product metrics from Posthog, agent traces, GTM data, etc. Crucially, the same harness that runs our customer-facing product also powers these company agents.

Phoebe core is explicitly designed to be a vibe-coding safe zone and allow tooling to emerge bottom up, from every corner of the company:

• In CX, Cindy and Dash built an increasingly-automatic post-sales CRM, with agents that run on schedules to monitor communication, check account health, usage metrics, and all the other core CX tasks

• In Growth, Suket and Srini run outbound and paid increasingly through a variety of automatic systems that we won’t discuss further here (no alpha leaks).

• In Engineering, our product is becoming increasingly self-healing. Our harness automatically reads and reviews its own work, looking for bugs and poor performance for customers. It’s aware of its own code, knows our customer base, and can search through all of Core (production logs, call transcripts, Posthog data), to fix bugs and gaps.

I also fundamentally believe in betting on deeply AI-native, junior talent with no prior assumptions on how businesses/products are built and no constraints on the problems they want to work on. We have early career engineers working in every org: sales, support, and of course engineering.

Skills mean very little now. Mindset, flexibility, and motivation mean everything. More than that, we look for people with Jeremy-Strong-levels of off putting, candorous love of the game. People that want to win, that are commercial minded, and treat business like the athletic game it is.

Phoebe is hungrily adding AI-native, AGI-pilled talent at every level and for every function to build toward the end game and live the real-world Factorio dream.

Stayed alive, Knicks in five

I was at MSG for game 1 of the Eastern Conference Finals when the Knicks came back from 22 down in the fourth quarter. It was the most miraculous sports moment I’ve ever seen and I didn’t think it would ever be topped... until the right hand of OG.

Wednesday night started out terribly and it seemed like the officiating would once again take the day. But, between some brilliant coaching and leadership, a legendary choke job from the 8 foot loser corn dog, and stunning late game play the Knicks took back the vibes and momentum in the series.

We are increasingly bereft of monocultural events that focus the energy and attention of the world onto a single fulcrum. This Knicks run is the single most exciting, captivating, unifying event in my life as a New Yorker (with the exception of 9/11 but that’s different).

After game 4 I saw people completely take over the streets. I saw bus drivers laying on the horns. I saw children (1000 drunk teens) dancing on cars and bodysurfing. The on court action was incredible and the off court feelings are magical. I love the bandwagoning for the same reasons I want more transplants in New York: it creates more vibrant fanbase to celebrate with. Why would I want to be alone at a party?

We’re seeing something special and I have no doubt it will birth a newly engaged generation of Knicks fans.

Dunking on VCs

Last week there was a big brouhaha on twitter dot com about the bad behavior of VCs over the years. It was not pretty. While the behavior was varied (pulling deals, falling asleep in meetings, light crimes, etc.) the theme struck a nerve and there was clearly a receptive audience to see some venture capitalists taken down a page.

Clearly there’s some simmering rage against VCs...

But none of this is new nor is it random. In 2019 during a different bout of dunking, I wrote about why VCs draw such widespread ire:

In venture capital you can’t really build a brand or reputation on performance. You can’t advertise competence via returns. It just takes too long... Instead, VCs have to cultivate brands for themselves and their firms through omnipresence: they and their takes have to be everywhere always... Venture capitalists need to look smart because it’s impossible to prove they’re smart. Perception, not returns, is reality

This is, obviously, extremely cringey and annoying.

And beyond being generally annoying, VCs are, for most people, avatars of their own failure - even successful people.

After all, the overwhelming majority of interactions between the overwhelming majority of VCs and the overwhelming majority of founders ends in rejection. For most founders, raising capital feels like a humiliation ritual.

That breeds resentment and makes people eager to take VCs as a class down to size. Venture capitalists are the living avatar and personal embodiment of your own feeling of rejection.

Other people are prolific rejectors too. But HR managers don’t gloat constantly on the internet. Because VCs are principally in the narrative business (not the returns business, at least not on any useful or measurable timeline) they do all that rejecting while crowing and bragging on the internet. As a class, it’s deeply unsympathetic and un-endearing behavior.

Worse, VCs have become emblematic of a world arbitrarily stacked against you (as you struggle against being consigned to the permanent underclass). People feel powerless in the shadow of this faceless market that blesses some and damns others, and almost overwhelmingly does not bless them. Seeing that market and that arbitrary kingmaking power personified in a small coterie of people, and then seeing those people be demonstrated to be stupid, dishonest, lazy, disrespectful, etc., is validating and vindicating.

The catch, of course, is that it’s the hopium of proving VCs wrong when, for the modal or even the 75th percentile founder, they’re not wrong.

VCs are ultimately extraordinarily privileged. We don’t work that hard. We don’t add that much. We don’t risk anything. We’re generally all very lucky to get to spend our time with brilliant, passionate people working on the single thing in the whole world about which they are the most brilliant and passionate. So on balance, I don’t think it’s bad for people to air their grievances and ask for more from their capital partners.

On the other hand, the readiness to join the dunk-a-thon is a sign of the times, and of how seemingly arbitrary the market has become, even compared to a few years ago when I first wrote this.

Scott Heiferman (a real NYC tech OG) rolled a mini app to find Knicks watch parties around the city.

I’ll see you at the Garden on Monday.

Anthropic filed its S-1 and SpaceX goes public next week. There’s ≈$3T of market cap hitting in the next few months with at least another trilly on the way (OAI). Google is selling ≈$80B of stock and it’s not much of a story.

We are witnessing the total takeover of already-concentrated financial markets by a single story; debt and equities, publics and privates, newcos and established names. This will be an entirely unprecedented wealth creation and liquidity event. The exuberance about AI the technology will have become fully overtaken by exuberance for AI the financial asset.

These are (or at least have been) obviously great stocks. Whether or not they are great businesses is the most important question in financial markets right now. I don’t know ¯\_(ツ)_/¯

The Bull Case

World-changing technology. They wind up with infinite surface area, and everyone pays them for everything. Intelligence is the most valuable market in the world (all white-collar work), and they become both an accelerant for and a tax on global GDP. LLMs either scale in other domains like they’ve done in coding or it turns out that every task is actually a coding problem after all.

And this is not even accounting for true AGI and a post-scarcity / post-economic world.

The Bear Case

Foundation models are obviously incredible products but turn out to be terrible businesses.

Training is a cost of revenue, not an R&D cost. Stop training, everyone churns and moves on to your competitor or a distilled Chinese version. The labs are perpetually investing unbelievable sums in building these fast depreciating assets (capex and training).

Every model is bigger and more expensive to run and also gets distilled faster, so these companies have to make huge up-front investments that depreciate extremely quickly. They are competing against one another and with the in-house labs of huge companies that throw off huge amounts of cash. They’ll never be able to charge enough because of competitive pressures. They will just run out of money and places to get it.

So what?

On balance, it’s almost certainly better to bet on the bull case, despite how much upside has already been sucked out / pulled forward at these prices.

If there’s any chance the bull case is right, and you have either a low enough cost of capital or are otherwise short the proliferation of intelligence, you have to get exposure here.

And everyone who works a white-collar job is structurally short. We’re all waking up every day betting that our intellect and creativity will matter. Nobody has any incentive not to be in it and every incentive to be in it, even beyond all the high-falutin’ stuff. (You just look like a fucking moron if you don’t have exposure).

So we’re definitely in a reflexive bubble. If you don’t believe the incentives/structural reflexivity just look around and use your eyes.

Is the bubble worth it?

Will it be like railroads, which ultimately produced a ton of valuable infrastructure (I think yes), or like crypto, which was basically a wash (though of course we wouldn’t have AI without all the chips and stuff from crypto)?

Here, the intelligence and capabilities are a positive externality of the bubble/exuberant build outs. And the more distillation happens, the more the frontier gets eaten by open source, the better for everyone. It does seem clear by now that as long as we don’t completely bungle the second order impacts (on society, politics, and business) that we are all going to be the winners. I mean this literally. Productivity is the main thing. It doesn’t require AGI or LLM maximalism to believe that we’ll look back on today’s standard of living (our material conditions) and standard of toil (the drudgery we accept for ourselves) the same way we look back at a pre-electrified world: impoverished, immiserated, and base.

It’s possible the labs ultimately get wiped out. I don’t know, and I’m not betting on it. But the bubble being worth it for the world and worth it for the equity are not the same question. If information wants to be free, then intelligence wants to be unbounded.

None of this answers the actual question of the business vs the stock and the bonanza won’t settle it one way or the other. A business is not a stock, and a stock is not a product; the price won’t be a verdict.

Read More

• Don’t wait for the AI shock

• Investing in the second-order effects of AI

• Sam Altman’s Endgame

  ---

It’s time to lock in (sound on).

The agony and ecstasy of Knicks fandom is profound. I fell in love with basketball around 6th grade, c. 2005, the first innings of nearly 25-year, uninterrupted run of mediocrity and pain for the franchise.

While some of the bing-bong antics today are over the top, it all comes from a place of profound love and desperation. New York is a basketball city and this team/organization/coaching staff has finally given a team worthy of our love.

Knicks in four.

Slow Takeoff

Last week in New York we hosted Takeoff our now bi-annual mini-conference for young partners, principals, and emerging managers - and our counterparts on the limited partner side.

We are betting on this cohort to be the next generation of important investors in this asset class (and I’m lucky to call many of them my friends and peers). Some of them are running funds already, some are still building careers inside larger firms, all of them are doing the work without the security of knowing they’re right. We get to follow and support them, co-invest with them, and steal their best ideas. The least we can do is buy them a drink and put them in front of each other (and LPs) once a year.

The event itself was off the record so I won’t repeat/reveal much but we had a phenomenal group of speakers:

• Emerging Managers: Taylor Greene (TwelveBelow), Finn Murphy (Nebular), Niko Bonatsos (Verdict)

• Our Thinky Boys: Aaron Harris (Magid), Jeremy Giffon (Octave), Mike Dempsey (Compound)

• Fireside chat with Hunter Walk (Homebrew / Screendoor)

• LPs: Erik Strobel (Granite) and Matt Auxier (University of Chicago)

There were a few threads and themes that really jumped out.

Heart in the game Skin in the game used to do the work of binding people to outcomes; it doesn’t anymore. Everyone trades in and out, gets paid regardless, can leave for a portco or a frontier lab. But who would make a decision that costs them in the moment because they actually give a shit and are planting a long term stake in the ground?

Firms are products You don’t get to bolt together separate decisions about fee structure, fund size, partnership composition, check size, and LP base and call it a firm. They have to cohere or you get the bootleg cassette effect: a third-generation partner playing back a model built for and around people they’ve never met, audibly degraded each copy.

What comes after the navy The alternatives world ate venture. The platforms behave less like risk-taking partnerships and more like large institutional asset managers, because that’s what most of them have become. The pirates became the navy. The interesting question, which we did not resolve, is what comes after. Almost everyone in the room is making a bet, with their career and their LPs’ money, on some version of that question.

Tremendous thanks to our supporters who help make this possible: Cooley (Slow’s law firm), JP Morgan (a very deep partner to the firm), and Sydecar (our friends and thought partners).

Ask Yoni

This week I shipped something fun: a Yoni’s 99D MCP to know what I think/have written about stuff (this newsletter). I made this for myself as a more efficient/portable way to surface relevant context for writing and investing projects after getting annoying by constantly trying to manage project memory/context. I do tons of copy and pasting links which is annoying.

I figured someone, somewhere in the world my want to read my newsletter (or have their agent read my newsletter) within chat. One step closer to Yoni GPT.

You can try the MCP server here and the github is here.

COGE

Mamdani announced the Comission On Government Efficiency, which is an obviously dumb name for what remains an obviously good idea. My immediate skepticism about DOGE (since vindicated - it was a complete a failure) came from my (correct) read that it was a bad faith effort by unserious leadership. But we shouldn’t let that poison us against the thought that government can/should/must work better.

It’s also obvious that tech/software/AI should be a big part of improving public service and delivery. Democrats, progressives, and technocratic abundists (check, check, check for me) should be rooting for ideas like this.

Cost overruns, delays, bad service, etc. all undermine faith in government and spur the death spiral of state capacity. The only way to reverse the trend is to reverse the trend. Personally, I want to be involved and supportive in any way I can.

Remember, efficiency isn’t just less money in, it’s outcomes achieved per unit of work.Mamdani announced COGE - NYC’s Commission on Government Efficiency. This is a very dumb name for a very good . I was/am on record that DOGE was/is a complete failure because it was never a serious or good faith attempt but

1. Making government work better and be cheaper is obviously good

2. Software/tech/AI should obviously be part of that

It is absolutely incumbent on democrats, progressives, and technocratic abundists (check, check, check for me) to demonstrate that government can actually deliver, not just spend more money. We have to start in places like New York which have an immense pool of talent and capital and still face dire budget problems.

I hope to support COGE and be involved in any way I can (step one is coming up with a better name).

Being Public Seems Hard

Seems basically impossible to be a public company right now.

You’re simultaneously getting the feedback that you need to be an AI leader AND that you predictable high margins but:

1. You’re spending tons of tokens, and increasingly, more than the tons you had planned on spending (“we blew our whole budget in 3 months”)

2. All the spend is highly experimental and you don’t know if the results will be there (good or bad) or how they will materialize (more growth, more profits, etc.)

Public markets love predictability (the only better than good news is the expectation of good news, and the only worse than bad news is unexpectedly bad news).

That is basically impossible to deliver right now.

From elsewhere

• Is this a new 5th job or just a new meaning to slop cannon?

• I am extremely into neighborhood restaurants getting “day jobs” as popups/coworking spaces/coffee shops. Margot/Winner has become a staple for me. I’m pretty curious about the economics since part of the appeal is low density/low turnover but hey, if it works!

We’re in the installation period of the AI supercycle where hype, speculative capital, and technological progress are rampant. The deployment period, where it becomes usefully integrated into our productive base, is just beginning and will last generation(s).

Deployment is the most important thing happening in the world right now. Acceleration and abundance vs stagnation rests in the balance.

Accordingly, there is intense, widespread interest in making bets directly on deployment. Betting not just on the beneficiaries (new AI-enabled products and services) but also on the delivery (the IT services working on AI transformation projects for legacy institutions).

System integrators and other consultants/AI services firms are/will be the connective tissues between installation and deployment.

There are three or four main flavors of AI services right now:

1. Fast moving FDEs/bodyshops: (elite) generalist teams parachuting into enterprise transformation work horizontally and on a per-engagement basis

2. The captives/joint ventures: the deployment arms and JVs being spun up by the labs and hyperscalers

3. The product companies: specialists building an internal product that delivers a specific outcome for a specific kind of customer, repeatedly. We have one company here (operating as a system integrator) and are looking for more. And of course there are a few companies selling these kinds of tools externally to other service providers (arguably the fourth).

The bodyshop approach is a rational response to the current market environment and the belief that things are moving too fast to do anything else. There’s not time to build a product; you just have to do stuff. It’s also a much better deal for a certain kind of talent. The best FDEs get to be the hero and directly monetize their skills instead of running customer success for a research team that sees deployment as overhead.

The JVs/captives are a fundamentally different proposition but also rational for their backers/creators. Their deployment JVs don’t need to create value companies directly; instead they generate EV by expanding the surface area and lock-in for their sponsors. Investing $100M to improve the odds for a $10B position is obviously rational. At $1T it becomes necessary.

Of course there is some risk of a “Hotel California” situation: once they’ve built your AI infrastructure on their stack, with their FDEs, around their models, you can’t check out. Their people aren’t working for your outcome, they’re working for their sponsor’s lock-in. The tradeoff is that they have max recognition and access, so a lot of buyers will take the deal anyway.

The most interesting version of the story is a product company delivering a service, rather than a service company delivering a product.

The internal product enables differentiated service delivery. The product might look like some combination of an agent/harness, context layer, workflow automation, etc. but in any case you’re creating a product that you consume internally in order to deliver a service externally (to your customers) better, faster, cheaper. The product companies bend the cost curve down rather than charging more because of temporal buying pressures and unique access to talent.

One of the most obvious tells is whether you’ll do anything for everyone or if you have a narrowly defined ICP. Are you trying to productize the process that delivers a specific kind of outcome in a specific context, or are you offering to do custom work and build anything your customers want? The product companies will have clear religious beliefs about what they can’t touch or won’t do.

We’re looking to back AI Accenture, not “Accenture for AI.” That is, there’s an immense opportunity for services companies to fundamentally rebuild how they deliver, not merely what they deliver.

Read more:

• The Complexity Crisis

• Implementation is the task of our time

• Modeling software after SaaS

Cliff Club

Introducing Cliff Club: a community exclusively for early employees at venture-backed companies and we’re starting with NYC first!

We’re gonna bring in great early operators and subject matter experts to talk about questions like:

• WTF is QSBS?

• How should I think about early exercising?

• How should my role evolve as the company scales?

• How should I feel about getting layered?

• Should I specialize into a role?

Really excited to work on this with my friends Charley and Leeor.

Take the dive.

Very busy week with a lot cooking include a teaser for a new project. More to come soon.

Venture Grounds

I had a lot of fun on the Rho podcast talking about the second order effects of AI (what happens AFTER we token max) and the right way to back seed founders (give them money and don’t fuck it up).

If the development and, increasingly, deployment of AI is the most important thing in the world, then the most interesting questions are what happens after we’ve done it (to the extent that it’ll ever be “done”).

I’m extremely bullish on and interested in the second order effects of AI:

AI amplifies volume until existing filtering/routing mechanisms collapse and need to be rebuilt.

• Deployment: more apps, faster releases, more integrations, etc. SIs are critical chokepoints and will become very powerful this cycle.

  • Slow company: Stealth, Stealth, Stealth

• Hiring: Applications are free to send. Volume explodes. Recruiters can’t distinguish candidates. Resume screening stops working. The whole funnel breaks.

  • Slow company: Phoebe, MeritFirst, Tofu

• Outbound sales: Sophisticated campaigns available to everyone, flooding channels. The well gets poisoned and new GTMs replace it.

  • Slow company: Memelord, Stealth

• Code collaboration: Git was built for scarce, human-written code. Whatever replaces it will be built for abundance.

  • Slow company: Atomic

• Trust and security: Everything on the internet is fake. Content, identities, credentials. Systems that assumed human-scale production and human-verifiable authenticity stop working.

  • Slow company: Outtake, Sublime Security, Stealth

But broadly speaking, there are a huge number of spaces where attention (abundant) and production (scarce) will flip to devastating effect.

And simultaneously there are greenfield opportunities that would have previously been impossible or unimportant.

Cliff Club: Take The Dive

We love getting together to hang with operators and builders early/earlier in their journeys, whether or not they wind up building companies. I find that much/most of our best network comes organically by hanging out and jamming with people months or years ahead of a transactable opportunity. But we only earn the right to your time if we can be of service.

Introducing Cliff Club: a community exclusively for early employees at venture-backed companies, starting in NYC.

We’re gonna bring in great early operators and subject matter experts to talk about questions like:

• WTF is QSBS?

• How should I think about early exercising?

• How should my role evolve as the company scales?

• How should I feel about getting layered?

• Should I specialize into a role?

Really excited to work on this with my friends Charley and Leeor.

Take the dive.

Drink the Haterade

• For all my Fallon haters: Fallon acts as the high priest of a terrified optimism, his rictus grin serving as a shield against the encroaching silence of the real. Here, in the sanitized, over-lit heart of the American culture industry, there is an inescapable horror. But it isn't a monster lurking in the shadows; it is the manic, unblinking insistence that actually, there are no shadows at all.

• For my Stephen Miller haters: Stephen Miller is surely the most powerful unelected person in America. The 40-year-old’s official roles — he currently has the dual titles of deputy chief of staff for policy and homeland security adviser — vastly understate the influence he has had in shaping Trump’s most hardline anti-immigration policies and rhetoric for the past decade. Whenever the president says something particularly inflammatory or offensive about immigrants in a speech, there’s a good chance Stephen, his director of speechwriting during the first term, told him to say it.

• For my Iran war haters: Defeat in the present confrontation with Iran will be of an entirely different character. It can neither be repaired nor ignored. There will be no return to the status quo ante, no ultimate American triumph that will undo or overcome the harm done. The Strait of Hormuz will not be “open,” as it once was. With control of the strait, Iran emerges as the key player in the region and one of the key players in the world. The roles of China and Russia, as Iran’s allies, are strengthened; the role of the United States, substantially diminished. Far from demonstrating American prowess, as supporters of the war have repeatedly claimed, the conflict has revealed an America that is unreliable and incapable of finishing what it started. That is going to set off a chain reaction around the world as friends and foes adjust to America’s failure.

For Lovers

• America is experiencing a productivity miracle and we basically don’t know why. It’s not because of AI (at least not yet) but it is really good news.

Last weekend we hosted Slop Con: an all day hackathon for slop cannons to work on personal projects and compete for (cash) prizes. We got 60+ people squirreled away all day on a gorgeous NYC saturday - extremely bullish. Big thanks to everyone who came out.

The fact that it was, by and large, some of the stupidest projects you could possibly imagine (oneshotting Mog/Chop, e.g.) reinforces the point. These are all people who came out because they wanted to have fun, try new tools, and chop it up. Crucially, they cut across technical and non-technical folks. We had AI engineers and equity research analysts. There were founders and interns.

In the weeks since I published the Four Jobs, we’re starting to see the term/idea crop up everywhere:

Even Max at Pangram, the anti-slop company, proudly refers to himself as a slop cannon.

Because there’s a difference between using slop to produce and producing slop.

The slop cannon uses janky tools and rough output as inputs to better work. The bad version mistakes the byproduct for the practice. I’ll flatter myself and say I do a good job of this in how I use (and don’t use) AI to write; bicycles for the mind vs motorcycles for the mid.

Through a willingness to go produce, go fast, and look dumb you too can reach slop nirvana.

There’s a chance that this is ultimately a distinction without a difference. It might not really matter how or why the best people produce when the modal production is both 1) bad and lazy 2) high volume.It’s very possible that, irrespective of that distinction, at least as far as it comes to communications, sales, marketing, etc., we still wind up in an age where all the channels are choked and stuffed.

And no matter what, we still reach the age of inbound where attention (abundant) and production (scarce) flip.

But what about “taste?”

I reject the idea that we need to choose between velocity and judgement, slop vs taste.

1. The whole organization needs intuition and judgement (more meaningful and measurable than “taste”). And the organization as a whole needs to move fast, when the moment calls for it. Gotta walk and chew gum, lads.

2. Taste in the common tech parlance is a meaningless, ineffable meme. It’s a cultural signifier rather than a specific attribute or skill. See also: “tasteslop.”

Slop cannons need to be product- and commercially-oriented to effectively harness their gifts and the blessings of our new AI age. But they need intuition and judgement to point themselves in the right direction, otherwise they’ll run really fast off a cliff.

So remember kids, slop ‘em up.

Slop Cannons Wanted

We have some great companies building fast and looking to add slop cannons.

• An ex-One Medical, ex-Bridgewater founder backed by a16z, Slow Ventures, and 100+ leading angels is building a stealth health tech company in New York - physical AI applied to medicine, vertically integrated from software to clinics. Looking for engineers who think in systems, physicians who ship product, and a Chief of Staff who lives in Claude Code. Email stealth@rungravity.com

• Michael Rado and Sam Kotlove are building Smoke: an agent workspace that runs sales, marketing, and ops for small teams. Set up a dozen agents in minutes, no code, no connectors, enterprise tools included. The core loop is prospecting, outbound, and content marketing. Nothing goes out without your say. Hiring GTM Engineers, Product Engineers, AI Wonks, and Product Designers.

• Justin Woodbridge is building Phoebe: a self-driving network of agents aggregating and directing worker attention in home care. Replacing coordinators inside agencies with agents that never sleep, never churn, and never forget. Looking for a Product Ops lead who wants to be the human in the loop for hundreds of agents running real workflows in production. This is a young team where everyone ships. (NYC)

• Goodlane is a repeat-founder team building an AI-native parent brokerage for freight agents in Brooklyn. Independent agents own shipper relationships and run their own mini-businesses; the parent brokerage automates everything else. Looking for an Agent Services Lead to own the relationship with the agents themselves: recruiting them, onboarding them, and making them successful. Part GM, part founding GTM, part product. If you want to run a business inside a business, this is for you. (Brooklyn)

In other news: AI Deployment

OpenAI and Anthropic have now both launched big consulting joint ventures to do enterprise AI transformation and create a new channel to sell tokens.

My bet is that most companies aren’t really ready for full scale transformation so big deployments will keep failing at high rates. The underlying data infrastructure isn’t there yet in many orgs, to say nothing of the culture.

Remember, I am long system integrators and services-led businesses where you own the outcome.

Read more:

• The Complexity Crisis

• Deployment is the task of our time

I wish I were more interesting, but I generally take the modal view that AI will wind up being a job creator, not a job destroyer, with two caveats:

1. In the short term, it’s almost impossible to imagine a scenario in which there’s not a ton of labor market displacement and economic churn.

2. In the long term, I could be wrong.

Either of those scenarios is good justification for some kind of insurance.

Everyone basically expects some kind of shock to occur, even if only temporarily. Whether that’s job loss, energy prices, a wave of cyber incidents, who can say. What we know is that if Something Big happens, there is almost certainly going to be a period of displacement and torque. We don’t know what it’ll look like but it feels inevitable that it will happen. So it seems obvious to take out / issue some insurance.

The default answer in this conversation tends to be UBI. But UBI is dumb and bad. It’s both too expensive AND doesn’t go far enough because it’s not productive. It creates a permanent underclass of government subsistence vs developing human capital. If people leave the workforce during the tumult, their path back into the workforce is hazy at best and their tendency to become antisocial is high.

Sam writes a lot about AI as a meaning crisis, not a job crisis. Even if we solve the income problem, we’re still left with video games, porn, and nihilism as a 24-hour job.

Better to develop human and physical capital through a public works project that encourages social cohesion and human flourishing.

We should be prepared for massive investments in physical infrastructure, human capital, and human flourishing:

• There’s a huge amount of shovel-ready projects in the country, and no amount of AI job displacement will solve the fact that all of our bridges are falling down.

• Human capital investments (health and education/childcare) increase the productive capacity of the economy over the long term

• And human flourishing (arts and culture) is a necessary condition for social cohesion. If we’re a rich society that’s going to become an ultra-rich society, we should engender the conditions for arts and culture. Otherwise, what was it all for?

Treating AI insurance as a jobs program vs pure redistribution is also a great hedge. If I’m wrong and there’s no risk of crisis, we’re still investing productively in human and physical capital AND it’d be much easier to end a jobs program than to undo an entitlement program like UBI.

Note: There’s also a very real possibility of other kinds of shocks that merit insurance and investment. Think biosecurity and cybersecurity. We are sticking our heads in the sand on this.

Not having insurance for something we think might happen and might be really bad if it does, on ideological grounds is just dumb.

If we don’t have smart stuff ready to go with broad buy-in we’ll do dumb stuff or we’ll do nothing at all and history will happen TO America.

Stray thoughts from my feed:

• Moral leadership requires moral authority and courage. You don’t get to whine about cultural/moral decay while actively promoting it. It’s fine to make money but maybe just shut up and dribble.

• The real American dynamism is not buying low grade cashmere from Quince. We can all do better than the same grey, made in China crew neck.x

As the difficulty to produce/create software goes down and the lack of true moats in most software companies becomes more apparent, the unique value of network effects as a truly defensible position will reprice.

Chris Dixon gave us “come for the tool, stay for the network.” James Currier gave us “market networks.” I want to add a third: agent networks. Self-driving software that starts useful to one user, then backflips into multiplayer coordination as those users’ agents start transacting with each other. Build the tool, acquire the network.

AI will actively create more and new opportunities to build network effects, even if the playbook and form factor diverge wildly from web 2.0 marketplaces.

Platform shifts and primitives

Historically, marketplace and NFX businesses were built on a set of primitives: payments, identity, messaging, and risk/compliance. With each new platform shift (internet, then mobile) these businesses have added/relied on new primitives as the base metal powering the network, beneath the interface and UI.

The first marketplace businesses (think eBay) were built on a hugely important and newly possible set of primitives enabled by the internet as a platform. Payments and messaging made those web 1.0 businesses work. Finding, talking to, and paying strangers was a hugely valuable new capability made possible by the internet and powered a new trillion dollar economy of online commerce.

The next platform shift (mobile) spawned a wider array of 2- and 3-sided businesses (Uber, Airbnb, Doordash, targeted ads, etc.) that added identity and risk/compliance, allowing NFX businesses to extend deeper into the messy real world of people and services. These (mostly) locally constrained services companies were brutally expensive to build because of the requirements for density and the cold start problem BUT they could address bigger markets.

For a brief time, it seemed like “market networks” would be the next evolution here: companies like AngelList and HoneyBook would start as software companies to attract one side of the market and then power connectivity to create NFX and lock-in with the other. But there was no meaningful platform shift powering these businesses and no new primitives; the single player utility wasn’t high enough or new enough at the start. While they didn’t fail (those companies still exist after all), there are no IMPORTANT companies to emerge from this hypothesis.

Now with the AI platform shift, context will be the fifth primitive for marketplaces and agent-networks are the emergent NFX paradigm. Even if it’s self-driving and has no conventional interface it functions on the same basic metal.

With each new platform shift, NFX businesses tend to monetize at a higher rate. As you add new primitives, you creates new monetization pathways, which ultimately drives take rates and viable category sizes. Context as a primitive powers the addition of a SaaS-like business model, which obviously opens up new sizes of categories.

Agent networks

Agents finally make the single-player tool valuable enough to stand on its own. You can build in sequence and grow revenue while in pure single-player mode because the utility of agents is so high. With obvious, out of the box utility you get cheaper acquisition and avoid the cold start (kind of one thing), and the shape of the network changes.

The user’s agent is a way to acquire context and serve them in single-player mode. My agent does stuff for me, part of which might be talking to you, which sets up the company to sell an agent to you. Then our agents talk together / to each other (you know, business).

There are a handful of companies cropping up and starting to run some version of this idea already.

• Phoebe does this in home care. Scheduler agent for agencies, talent agent for caregivers, family agent for clients. Three sides, one network. More in the announcement post.

• Ando does this in restaurant staffing. Forecasting and scheduling for operators, Ando Passport for workers (a verified identity that compounds across shifts and is portable across employers).

• Hero and Nas.com do this for sellers. Take a photo, the agent identifies the item, prices it, writes the listing, posts it. Acquire sellers now, run the marketplace later.

• TruckSmarter does this in freight. Owner-operator truckers pay for an agent that aggregates loads, handles scheduling and paperwork. The talent agent is the wedge into a network nobody has been able to consolidate. Notably TruckSmarter is pivoted into this; it’s not a new company.

Across these verticals (and many more) the pattern is/will be the same: Pay a single-player agent to absorb grunt work on the side of the market that is hardest to acquire. The network follows.

Self-driving networks have to drive themselves, after all.

Where to build

The pattern selects for different TAM and acquisition cost profiles than web 2.0 marketplaces. What to look for:

• Coordination and memory is the grunt work on the acquirable side. Middle-of-the-curve tasks. Not physical, not skilled, not requiring personal agency to get started.

• Buying propensity (not P2P). It’s too hard get normal people to pay for a single-player agent (normal people don’t buy lots of software for themselves). B2B or B2C where one side can absorb a SaaS-like agent, or P2P close to a transaction where one side has prosumer-level coordination load (Hero, Nas.com).

• Sneaky big TAMs: context makes the TAM’s bigger but you still can’t do it anywhere that’s micro, or you at least have to have a point of view on long-term category size.

The two places it doesn’t work:

• Networkable but not agent-acquirable. You still can’t do Uber from scratch. There’s no market creation opportunity if you need existing people to benefit from an agent, and there’s no good case for single-player Uber as an AI agent that a driver would have paid for. Riders obviously aren’t gonna pay for a ride booking agent.

• Agent-acquirable but not networkable. Vertical AI that works like SaaS. Real single-player utility but no multiplayer layer to lean on. If external coordination can’t/won’t be on the roadmap, it obviously can’t work in multiplayer mode.

Conclusion

I’m not convinced lots of standalone, non-NFX driven agents will have a lot of long term value, at least not unless they can move past SaaS-like experiences, business models, and GTMs. Switching is easy and you have to sprint to stay in the same place because of model improvements.

Agent networks are different. The NFX drives utility you can’t recreate in single-player mode even if you perfectly exfiltrate data and infra. The data and infra make the system better, of course. So it’s networks, then data, then infra, in order of importance. The infra is the least important because it’s the most replicable; anyone can build good software on their own but you can’t just produce the data, and even if you could it would be useless without a network.

The graph is the moat and context is the new primitive driving a business model shift for marketplaces.

Read More

• Life on Autopilot: on on what agents can, can't, and shouldn’t do

• Don’t call it a moat: on why software moats mostly aren't

• The phone company always wins: on the durability and appeal of NFX-business

Slop Con

On Saturday, May 2nd, we’re hosting NYC Slopcon, a hackathon for product engineers, vibe coders, commercially minded engineers, technically minded sales people, high velocity operators, etc..

We are convinced that the world is reorganizing around four archetypes and ways of working: slop cannons (non derogatory), SREs, adults, and hot people.

The best AI native companies are increasingly recruiting commercially minded engineers regardless of the role. They explicitly want people who are comfortable using tools AND thinking about product AND thinking about customers. The salespeople are shipping (at least internal tools and automations for themselves) and the engineers are relentlessly focused on customer value.

The highest performing companies will have ‘product engineers’ and slop cannons in every role (product/eng, sales, ops, talent, finance, CX, marketing, etc); it is a multi-hyphenate skill set crucial to accelerate each area of the business.

We’ll have prizes and credits from OpenAI, Vercel, Pangram, and Memelord. Join us. Slop em up.

There’s a meme war going on right now between “it’s all about SBC” and “it’s all about terminal value.” The obvious answer is that it’s both. But the important is *why* it’s both, and what that tells you about where the SoftWars™ are headed.

The market has decided that >95% of software is an annuity (fixed life) with $0 terminal value. If you’re running an income stream, you need to care about income and constrain SBC (a cost). If you’re accelerating/have TV (building a perpetuity/compounder), you don’t. That’s the basic dichotomy of the market right now.

If you are one of the happy few companies that the market believes not only has INCOME/earning potential but the potential to be a perpetuity (has terminal value), then it doesn’t matter because nothing matters. 25% of revenue as SBC is merely the price to pay for acceleration. The problem is that your low growth, moderate adjusted-EBITDA SaaS name is not one of them. Cutting SBC will make that business more attractive to a value-oriented financial acquirer but doesn’t give you a future.

Running the badly run middling software companies a little (or even a lot) better won’t change the narrative albatross weighing them down, except perhaps to the extent that running them better means just going all in on AI and acceleration.

From Matt Slotnick:

So how does software fight back?

You build.

But not in the measured, incremental, multi-year roadmap manner that they’re used to. The frontier shifts too fast for a multi year roadmap.

Take a portion of free cash flow and stand up a small, empowered labs group, with a real, honest to god mandate supported by the CEO. A mandate to build new solutions for your install base, unencumbered by existing offerings, pricing, messaging, or value props. You have to build, and you have to build fast. Because AI is moving fast, and prevent defense never works.

Do everything you can to instill this entrepreneurial DNA within your company. Hire and promote people who truly get it. And part with those who don’t Compensation structures probably have to change. Talent assessment very likely has to change.

Strong endorse.

Now to make this about ME. This matters as much or more at seed than anywhere else.

Basically the only questions that you can really even ask are:

1. Do I love this guy/gal

2. Is there a story that matters

3. Is there something convex/interesting to spend money on today

4. Is there terminal value

This often gets muddled by asking “what’s your moat” which doesn’t make sense and has no good answer. You of course can’t have a moat early. If you could, you probably couldn’t start. (Some notable exceptions include things like regulatory capture through political connections, but those are rare).

What you can (must) is a point of view on the long term margins, earning power, defensibility. A thesis about where the moat comes from and what the business looks like over time. If nothing else it’s a more interesting/revealing conversation to have early.

And if you can’t/don’t have a strong POV on terminal values in 10 years you shouldn’t be investing in a category with 10+ year hold periods. This goes for founders and investors alike.

I increasingly find myself saying “I don’t doubt you can do X reasonably well and earn a bunch of revenue but let’s both accept there’s no TV in that so what’s it for?” Obviously things are hard and you can only make those claims in a narrow set of companies but the point stands. And I it’s not an academic question; it winds up really impacting what you build and prioritize today (capital allocation).

Counterintuitively, predicting the long term is actually easier/higher confidence than predicting the medium term. If Trump had lost in 2024 the world today would be very different but the world in 2040 might be largely the same either way because the forces/stories propelling the world are broad and deep. It’s variance around durable trend lines.

We are not traders. We’re investors who, as a bug or feature, have no liquidity. All we have is terminal value - hopefully our companies will too.

We Co-led Prefix’s Seed Round

A couple years ago I wrote about why repair and maintenance is a super elegant way to get into industrial workflows efficiently and build a many-armed tentacle monster:

• Many shots on goal to make a sale; something is always broken unlike equipment sales which is infrequent.

• Motivated buyers: repairs are urgent. When something breaks it needs to get fixed right now because uptime drives revenue.

• High value: unlike single use supplies, machinery and heavy equipment is high value and fixing it is high skill - not just a question of price.

That was a preview of Prefix (and Heave).

Prefix is a marketplace for restaurant and retail facilities maintenance. Independent technicians on one side, big multi-site operators on the other. The company makes repair and maintenance better, faster, and cheaper for some of the most demanding brands in the country, including Chipotle, Raising Cane’s, and Bojangles, across thousands of locations.

Restaurants are a great place to start because they’re large but highly federated accounts. A national chain has hundreds or thousands of locations, each depending on local technicians for HVAC, refrigeration, plumbing, and electrical. These are big customers that don’t want to deal with rinky dink procurement across every market they operate in.

Prefix lets big brands interact with small, local service providers on more of a like-to-like basis: buy from Prefix, get service fulfilled by the best local players. Best of both worlds. And on the other side, it makes those big accounts winnable for small businesses that would otherwise never get in the door.

By starting with repairs (recurring, urgent, valuable) Prefix can profitably earn the right to start building relationships with customers on both sides. Over time there’s obvious expansion opportunities into equipment purchase, capex financing, single use supplies, software, and self service offerings for small independent brands.

We co-led a $7.5M seed round with Collide Capital to help Prefix keep scaling. The company is expanding aggressively beyond QSR chains into retail and other multi-site verticals, and over time we’re excited to see them go deeper on both sides of the marketplace.

Prefix is a revenue driver and sales orchestration opportunity for the supply side, and a coordination layer that keeps getting more valuable for the demand side. This is how useful technology gets into the real economy and makes long tail entrepreneurship more possible and profitable.

Prefix is hiring AEs and senior salespeople around the US.

Slop Con

On Saturday, May 2nd, we’re hosting NYC Slopcon, a hackathon for product engineers, vibe coders, commercially minded engineers, technically minded sales people, high velocity operators, etc..

We are convinced that the world is reorganizing around four archetypes and ways of working: slop cannons (non derogatory), SREs, adults, and hot people.

The best AI native companies are increasingly recruiting commercially minded engineers regardless of the role. They explicitly want people who are comfortable using tools AND thinking about product AND thinking about customers. The salespeople are shipping (at least internal tools and automations for themselves) and the engineers are relentlessly focused on customer value.

The highest performing companies will have ‘product engineers’ and slop cannons in every role (product/eng, sales, ops, talent, finance, CX, marketing, etc); it is a multi-hyphenate skill set crucial to accelerate each area of the business.

We’ll have prizes and credits from OpenAI, Vercel, Pangram, and Memelord. Join us. Slop em up.

The world needs attorney client privilege for AI

Messages to LLMs are not privileged/are discoverable. We’re starting to see them show up in litigation along with the advice for regular people not to ask Claude about the law.

This is a gift to the legal cartel and it’s bad for everyone else.

Consulting LLMs about legal matters in parallel to lawyers, before lawyers, and instead of lawyers is pro-social behavior. The alternative to an LLM reading your employment contract is not having a fancy lawyer read it; it’s signing it blind. Same goes for leases, demand letters, non-competes, etc.

Worse, the genie can’t go back in the bottle. The landlords, employers, corporations, and scammers will use AI to drop the cost of contracts and litigation threats to zero. Consumers will get screwed.

Someone should build the product that serves the behavior and protects the user.

The shape is probably something like: minimum viable pretense for attorney-client privilege over LLM conversations. You sign up, you sign an engagement letter, there’s a licensed attorney on the other end blessing the channel, and your conversations with the AI happen under that umbrella. I’m sure there’s nuance here, but the broad strokes seem solvable by a good lawyer and a good product person working together.

The business model is either standalone consumer AI and/or PLG for a consumer neofirm.

On one hand this is “safety driver in the Waymo” which is dumb. On the other, it’ll ultimately be good for people, which is good.

I had a conversation with an investor managing a large institution and many billions of dollars this week. This guy is, by any measure, orders of magnitude more successful than I am and pushed hard/called bullshit on some of my usual talking points.

He’s not gonna read this but I’d still like to state my case a little more eloquently than I did IRL, especially in light of “the group chat” essay which is cynical, distasteful, and completely true.

Cheap like expensive is neither inherently good nor bad. A business is just a business, and price at seed is mostly a reflection of the market, not inherent worth. Remember, any reasonable DCF/the true FMV of every early-stage investment is approximately zero.

Believing that something being out of favor makes it good is just as specious and lazy as the opposite belief: that something being expensive makes it good. In either case, you’re outsourcing judgment to others.

Is there any objective measure of (non-) consenus? Obviously yes but also no, not really.

“Non consensus” can either be an input (price) or process (judgement).

In the former, you’re agknowledging/claiming that only way for something to be non-consensus is for it to be cheap. The idea of being a “contrarian” while paying top quartile prices is obviously insane nonsense.

Conversely/in the latter, you’re focused deriving a judgement/value of a business irrespective of what others think (how hot-or-not it is). That will sometimes mean paying up!

With enough companies/a big enough sample, the two ideas should converge at the median and diverge at the average (you might independently conclude Anthropic really is that good and be right!).

In an industry/asset class with power law outcomes, it’s totally reasonable to pay multiples more for the 99th percentile than the 90th percentile of quality opportunities, but caveat emptor. Distinguishing between those without real meaningful data beyond what others think is basically impossible or, at the very least, approximates to a random walk.

If I have one core insight and belief at seed/pre-seed, it’s this: there’s a huge amount of false precision and unreasonable confidence in assessing the best, the worst, and the median opportunities.

What “everybody knows,” what “the group chat decides,” and what’s actually true are not a perfect circle even if it feels really good to believe that they are.

From Kingmaking and the limits of external conviction:

King making is permission-based; it only works on the basis of elite buy-in/consensus […]

At Slow, we believe deeply in founder-led/internal paths to conviction. Whether that happens heads-down in a notebook (calling your shot) or heads-up in the market (calling your customer) doesn’t matter. We can fund you to conviction/insight or meet you once you have it. But it has to come from you, not us.

“The group chat” is externally derived conviction and is highly permissioned and better suited to those who need permission to build and buy.

A New York financier reviews New York Review of Finance

This morning I got the second edition of NYRF delivered in print to my door. It’s great and I like the un-accredited writing in the style of The Economist. It creates the impression that you’re peaking behind the curtain of what folks are ~~really~~ saying but can’t say...

A lot of reads a bit too skeptical and incurious about tech (a preemptive elegy for AI is a bit on the nose) but that is probably the perfect center of the venn diagram between NY financiers living in Brooklyn and harboring literary/cultural ambitions and leftists-cum-Odd-Lots-fans who are reading and writing this (guilty as charged). Excited to see the next quarterly edition.

The art direction is great down to the partners they’ve chosen to work with as advertisers.

Notes

• This is a really fantastic slide from Redpoint’s market update deck/post.

• YC quietly removed delve from its directory without issuing a public statement. I really don’t get who that’s for/that seems like the most mealy-mouthed, worst version of a comms strategy. Stand by your guy or don’t, especially if you want to be the self-proclaimed moral voice of tech, as Garry does. ICYMI Delve did a big fraud on a bunch of other startups (allegedly) and might be at least partially responsible for the Mercor hack... bad!

• GLP-1s are a bigger business than AI. Continue to watch this space. Were it not for AI, they would be widely viewed as the most important thing in the world right now.

• One of my favorite posters posits that maybe only 3 jobs: innovators, marketers, and context carriers. Personally I prefer my four jobs but he makes good points 😉

A note on competition/conflicts:

Here’s an email I wrote to a founder on how we think about investing in competitive companies (which, in fairness, we basically never do). Some people got Big Mad at me online which felt fairly performative - just another salvo in the fight to prove (to whom idk) that you’re The Most Founder Friendly...

To me the takeaway is actually fairly positive sum: companies don’t really succeed and fail on the basis of some competitor, at least not until they’re much further along. We don’t need to kill the other crabs to escape from the bucket.

Slow Security

Super stoked for this: we’re hosting founders, operators, security leaders for a ≈100 person cyber security mini conference in NY next month.

We’ll have Anthropic’s head cyber and NatSec policy for a fireside chat and panels with some excellent security operators and investors.

Sign up to join us. Space is limited and we want to prioritize builders and buyers.

There are only four jobs.

A few weeks ago I wrote this aside about how companies/jobs/functions are reorienting toward a new organizing principle. Then this week it went tragically/comically viral for no good reason (my linkedin is broken now). So it’s now worth restating the case.

There’s a very real possibility that the only jobs in tech companies are going to be:

1. product eng/vibe coder/PM/slop cannon: self explanatory. This is the high velocity, high tool use generalist. they are obviously not restricted to product and eng roles. Anyone can be commercial and product minded.

2. SREs/infra/security/systems: we’re going to be producing so much STUFF across every org that there’s going to need to be really really good people stitching it together, making it stable, secure, and robust.

3. Adults: sometimes you need a grown up in the room to just say “hey, come on.” They are effectively a much needed governor on an otherwise accelerating organization. You will find them across roles but there are obvious places like legal and finance where you’re NGMI without an adult in charge.

4. Hot people: You will find hot people in roles ranging from sales, to people, to CX. There will always be an important place for those who present an easy UX to the world and are pleasant to be around. Remember, there are many ways to be hot.

First, these are working styles and archetypes, not job descriptions. The latent traits have always cut across job titles and orgs. To ask “what about design” or “where does my job fit” is the wrong question/fails to comprehend what’s going on. The right question is how do you approach your job irrespective of the title.

Second, it’s probably not possible to be all of these. It may not even really be possible to be more than one or two. To be spiky in all things is to be smooth. If you’re saying you’re all four you’re probably a) wrong and/or b) misunderstanding my point.

Third, this new quartet very well may replace the classic trifecta of product/design/eng. That iron triangle was organized around categories of output. When anyone can produce code, designs, and specs, organizing around what you produce is meaningless. Instead the highest velocity teams will organize around how they produce.

The best AI native companies are increasingly recruiting commercially minded engineers regardless of the role. They explicitly want people who are comfortable using tools AND thinking about product AND thinking about customers. The salespeople are shipping (at least internal tools and automations for themselves) and the engineers are relentlessly focused on customer value.

The highest performing companies will have ‘product engineers’ and slop cannons in every role (product/eng, sales, ops, talent, finance, CX, marketing, etc); it is a multi-hyphenate skill set crucial to accelerate each area of the business.

Conversely, there are “SREs” everywhere for those with the eyes to see them. The basic function within their org is the same: to provide scaffolding that lets other people move fast and cleans up behind them. Whether that happens in marketing (QA and copy editing) or engineering is irrelevant, the value to their org, and by extension the company, is the same.

In a world of rapid acceleration and internal decentralization (everyone can autonomously execute everywhere simultaneously), you need adults who can rely on judgement and earned intuition. Sometimes that means saying no with authority to prevent catastrophic errors and acceleration off a cliff. Remember, velocity is a vector (it requires direction) and momentum requires mass (the work needs to matter). Adults can help keep you on the golden path. This is self evident.

You cannot win without hot people. Hot people are the interface layer. Externally they make the thing legible and attractive. Internally they make the org cohere. They’re the reason people want to show up to work and the reason people want to buy from you. Remember, there are many ways to be hot and I will not elaborate on this further.

I'm not an employment doomer. The biggest companies will get more efficient (lower headcount) but there will be lots more companies and the boundaries of "tech" get fuzzier every year. And to the extent that there are fewer engineers (objectively wrong at this point) it’ll only be because many of them get subsumed into other “non-technical” roles.

Some Stray Notes

• I’m not the first to point this out but the OpenAI/TBPN deal is clear proof of the immense value of hot people. Jordi and Coogan are equivalent to AI engineers.

• The NYT reported on what purports to be a ≈$1B revenue, one person GLP-1 company. It seems to be a fraud and the numbers are probably either fake or misleading. This is what happens when a slop cannon builds a HC company without an adult in the room.

• This is all orthogonal to agency which is important for everyone in every role in a highly performant, accelerating company.

Slow Security

Super stoked for this: we’re hosting founders, operators, security leaders for a ≈100 person cyber security mini conference in NY next month

We’ll have Anthropic’s head cyber and NatSec policy for a fireside chat along with panels with people building and backing security companies. More to come soon.

Sign up to join us. Space is limited and we want to prioritize builders and buyers.

Read on for internet drama about agency, my plea for more positive sum thinking in healthcare “innovation,” and a plug for our Slow Security mini conference ft. Anthropic.

People were mad online

Last week I tweeted and people got very mad at me for being insufficiently deferential to their very special relationship with claude…

Here’s why I’m right, and here’s what’s at stake.

What an agent is, and where agents are useful

An agent takes action on your behalf where you are the principal. An app does not. We know what apps are. We’ve been using apps for a long time, and there are lots of good AI features that can make apps work a lot better, which is a far cry from saying that the AI will proactively organize your life for you...

There is a reasonable argument that by now ALL AI products are “agentics” or have “agentic” qualities/features: browsing the web, deciding what modality in which to respond to a prompt, etc. are all actions taken on behalf of users. But if every product with AI is an “agent,” it’s a meaningless term and the whole argument reduces down to “is AI useful” which it obviously is.

The important debate is what it’s useful for, and where it hits natural limits that have nothing to do with scaling laws or inference costs - at least short of physical AGI.

AI agents are nascent but the overall concept is not so new that we already know where they can/will be useful: the middle ground of complexity between purely rote tasks (can do, maybe not necessary) and highly effortful/intentional personal endeavors (can’t do).

I totally believe that things that require a measure of coordination and memory are useful surface area for agents. That can include things in the real world. Finding service providers, doing your taxes, scheduling appointments: that’s the middle of the curve where an agent with sufficient context can probably make your life a little better around the margins.

But there are hard limits here. If it takes a little bit of agency you can make gains by handing off to an agent (price compare a bunch of stuff vs. buying the first result on Google). But if it takes a lot of agency (cleaning your apartment or reading a book) the agent is useless. Going to the gym with AI that helps plan a workout only matters if you get yourself to the gym and claiming that an AI that makes you a workout plan is “agentically getting you into shape” is just baby talk.

Things that require agency require YOUR agency to work.

Of course life contains work. Sometimes that’s a problem to solve but sometimes that’s just life. And when it is a problem to solve it may not be solvable by AI, let alone an agent specifically.

AI may be able to help you even if it’s not primarily an agent. AI features inside apps can be useful across the whole spectrum. The agent question is overly narrow and specific.

The confusion in the discourse is collapsing “AI is useful” into “agents are the future of consumer,” and those are completely different claims.

Real life doesn’t happen via command line

Of course people do work in their lives but most of it (the hard stuff) does not look anything like the work they do at work. Real life is rarely accessible via the command line.

This list is... something. It mashes together completely different categories of life as though they’re all the same kind of problem.

Maxis hear “agents can’t do X” and get butthurt about “doomerism” ignoring that what you can do and what you should do are not 1:1. There are limits to each of possibility and desirability, largely unrelated to one another.

On desirability: you can already outsource gift-giving with a gift card, a corporate gifting program, or an assistant. But doing that misunderstands the rich culture and history of gift giving. To give a gift is to say something about yourself, the recipient, and your relationship together through an item. An agent-chosen gift is impersonal and cold like a gift card.

On possibility: there are problems that are just beyond the capability of agents because they’re not defined by coordination and memory at all. Sometimes they come down to skills and physical manipulation, where the coordination of that process is limited in its effectuation of the outcome.

An agent can find you a handyman, but it isn’t going to oversee that entire process: noticing the problem, explaining it aloud, clearing out the space so they can work, making time in your schedule, being home, opening the door, showing them where the problem is, answering their follow-up questions, pointing out the part they missed, noticing they scratched your floor, checking that it’s actually fixed, deciding you don’t like how it looks and asking them to redo it.

It will still take work from you to do a good job even if you outsource to a handyman via an agent. And the agent certainly can’t just fix the shelf itself.

And anyone who thinks agents will make parenting or running a household easier should read More Work For Mother.

What’s at stake

I will never bet against laziness. But outsourcing your social and emotional life to an agent is bloodless at best and demonic at worst.

Having a machine take over your life to smooth out all its inconveniences and to view all life’s friction as surface area to optimize is both profoundly bleak and wrongly technomaximalist. Claude cannot give you the will to change yourself. Not all struggle is suffering.

The only way to make agents work for most of that list is to profoundly change yourself to live for the AI instead of the other way around. We already do this in small ways (saying things aloud to ensure Granola hears it, etc). In darker terms it will mean avoiding paths inaccessible to the agents: only buying what can be delivered is an atomized, limited life.

I empathize with the optimization view. I often treat things that way and wish others would take care of things for me/that I could outsource every hassle. But it’s not something to be proud of.

Planning a trip together is an opportunity for connection and collaboration. We should strive for that even when it costs us convenience. The friction is part of the experience - sometimes a really valuable part.

The point isn’t asking what an agent can do (some stuff) or how useful AI is in general (very) or even what an agent is (who knows). The point is asking what is good and desirable: a life of rich connection where technology serves you and you exercise agency to be the person you want to become.

Thanks for Chris Melamed for reading and shaping.

Positive Sum HC

Most healthcare “innovation” is zero-sum at best, negative-sum at worst. I want more positive sum HC ideas

Basically every HC idea boils down to:

• Pay providers more(drives up premiums, bad for patients)

• Cover more care (drives up premiums, bad for patients)

• Deny more care / pay less (bad for providers and patients)

Lots of shifting money around, very little to expand the pie. Much of this is just a reflection of the fact that “health insurance” isn’t really insurance; it’s catastrophic care insurance bundled with a rate card and a provider directory.

Prior auth automation/arms race is exemplary of the problem. You automate getting through hoops so payers build new hoops. Both sides spend more on admin despite the “efficiency.”

Positive-sum ideas I want to see/am rooting for:

• Gold carding/eliminating prior auth entirely.

• ICHRAs: an intermediate step to decoupling insurance and employment

• AI for low-acuity care: 80% solution at 10% of the cost; improves outcomes per dollar vs “expanding access” which just increases volume/premiums

• Care coordination: investment with a return on adherence lower readmissions

• VBC: right in theory, but admin overhead recreates the disease and blocks out small providers

• Pharma/new drugs: outside my purview but the purest positive-sum play. GLP-1s are the most important thing in the world after AI.

My personal blackpill: there are very few novel ideas for positive-sum systemic change in the biggest market in the biggest economy in the world.

Slow Security

Super stoked for this: we’re hosting founders, operators, security leaders for a ≈100 person cyber security mini conference in NY next month

We’ll have Anthropic’s head cyber and NatSec policy for a fireside chat along with panels with people building and backing security companies. More to come soon.

Sign up to join us. Space is limited and we want to prioritize builders and buyers.

Code as execution, not a moat

After writing about the SaaSpocalypse and the headwinds on software defensibility I got some pushback from a founder (h/t Comron):

1. You’ve been able to outsource software development to India for cheap for a very long time. If you had an idea and a spec, you could ship it overseas and get something back for a fraction of the cost. Why wasn’t offshoring the death of the software industry?

2. If code was already commodity enough to be a punchline (”all vertical SaaS is just Salesforce with extra fields”), what moat just died, exactly?

TLDR - code was always execution (not a moat) and LLMs are different in both kind and degree from outsourcing.

The outsourcing question

This is the obvious objection: if cheaper code was going to kill software, outsourcing would have done it already.

But outsourcing was a low-tech, shitty attempt to commodify code production. Huge barriers to entry, high cost to manage, bad results. It didn’t actually relieve the bottleneck.

AI code generation is a difference in both kind and degree.

• The difference in degree: code generation tools are so much cheaper, so much faster, and roughly at par with anything you would get from an outsourced team. That massively expands the market.

• The difference in kind: Outsourcing is basically useless to (and even a burden on) really talented engineers, whereas code gen/LLMs are a massive accelerant/multiplier on their ability to run fast and far.

Code becomes free at the low end (savvy non-technical operators solving problems with code) and amplified at the high end (great engineers being able to do more than ever before). Collectively, this compresses the value of code as a differentiator.

Software is more valuable than ever in that it’s more important. But the ability to do it well is no longer enough to separate you from the pack.

Code is/was a chokepoint on execution, not a moat

There are only a few moats and execution is not one of them (nor is speed), even if it matters a ton and makes some businesses better than the rest.

Writing application software was never a moat but it was more valuable (harder to replicate) as a form of execution. The reason it felt like a moat was that a limited number of people could do it well. That talent constraint made software development look like a barrier to entry when it was actually just a bottleneck on execution.

AI relieved the bottleneck, changing nothing about classical moats. More people can functionally do it well AND those who can do it really well can do much more of it.

Moats and execution are fully independent variables

Execution in the field of writing software is not fundamentally different than execution in the fields of hiring, selling, finance, etc. It matters enormously to how well a business runs, but it doesn’t create a durable barrier to entry or guarantee margins on its own.

Google Search is the definitive illustration. It is probably the best business of all time because it has multiple strong moats: data, network effects, brand, economies of scale. But the execution is notoriously terrible: no product direction, low velocity, bloated, kills products on a whim.

Moats determine the ceiling on how much value you can capture. Execution determines how close to that ceiling you actually get. They are fully independent variables.

The companies in trouble are the ones that confused the execution bottleneck for a moat. AI didn’t kill your moat if you never had one. It just relieved the chokepoint that made it hard for someone else to execute as well as you did.

Everything shifts down one tier

Execution on application-software-shaped problems gets easier, which means you can afford to focus on harder problems. Every company used to have to solve two things: the software problem and their actual hard problem. AI is knocking out the software layer, so now all your effort goes to the hard part. That makes each category one step more tractable.

For pure application software companies, there is no harder problem underneath. Software was the whole thing. When you knock out that layer, there’s nothing left to create barriers to entry. That’s why they look more and more like services businesses: easy to start, hard to differentiate, lots of competition.

But for everyone else, this is a gift. The infrastructure business that used to spend real effort on the application layer can now put all its resources toward systems engineering and reliability. The hardware company can focus entirely on the physical and mechanical problems that actually differentiate it. Same logic, repeated across every tier.

We’re seeing basically everything shifting down click one in terms of risk and feasibility. Pure software companies look more like services companies. Infrastructure companies look like software companies. Hard tech companies look like infrastructure companies in terms of difficulty, specialization, and the ability to start and fund them.

Even as application software becomes non-viable on a TV basis, the rest of the world and the TAM of software writ large is growing exponentially.

From now on there are only 4 jos

There’s a very real possibility that the only jobs in tech companies are going to be:

1. product eng/vibe coder/PM/slop cannon: self explanatory. this is the high velocity, high tool use generalist. they are obviously not restricted to product and eng roles. Anyone can be commercial and product minded.

2. security/SRE/infra: we’re going to be producing so much STUFF across every org that there’s going to need to be really really good people stitching it together, making it stable, secure, and robust.

3. hot people. You will find hot people in roles ranging from sales, to people, to CX. There will always be an important place for those who present an easy UX to the world and are pleasant to be around. Remember, there are many ways to be hot.

4. grown ups: sometimes you need an adult in the room to just say “hey, come on.” They are effectively a much needed governor on an otherwise accelerating organization. You will find them across roles but there are obvious places like legal and finance. They are basically the non technical equivalents of #2 and might even be able to be bucketed with that.

the latent traits have always cut across job titles and orgs.

Our Cyber Conference: Slow Security NYC

For the last year we’ve been going deep on how AI is changing the threat landscape by creating new surface areas to defend and transforming attacker economics.

At the end of April we’re hosting ≈100 founders, security operators, buyers, and investors in NYC to discuss and learn.

Join us

Etiquette School

This is gonna be fun. We’re bringing Slow’s etiquette school to NY in a few weeks to teach founders and builders in NY how to show up the right way. We’re gonna have a bunch of speakers and teachers to help you learn how to present like an adult.

Apply to join us

A few different notes and ideas this week about the interface between AI, power, and the law. And I’m big mad about a stupid law being floated in NY.

OpenAI’s obvious endgame: becoming a GSE

OpenAI raised $110B in the largest private fundraise in history. Same day Trump threatened to kill Anthropic and OpenAI stepped over the body to take the Pentagon contract (with some equivocating).

This is one event.

Foundation models are obviously incredible products but might be terrible businesses.

Training is a cost of revenue, not an R&D cost. Stop training, everyone churns and moves on to your competitor or a distilled Chinese version. The labs are perpetually investing unbelievable sums in building these fast depreciating assets (capex and training). And each new SOTA model costs more than the last; we’re rapidly approaching >$1B training runs.

There’s only so many places to get $100B. The largest IPO in history, Saudi Aramco, was “just” $30B. The capital markets are running out of road. Retail and public markets can’t easily absorb or finance the build-out.

The natural next dollars are from the U.S. government. The only parts of the economy that are growing are the AI capex trade and the great national nursing home. Without AI capex we would be in a recession and the Trump admin can’t take that risk. The DoD contract fight was a preview for truly picking winners and Chinese style market intervention.

The endgame: OpenAI will become a GSE/SOE, at least in fact if not in law. The mechanism – equity, debt, guarantees, some weird structure – is unclear and unimportant.

It’s very likely the taxpayer ends up as the bagholder.

Legal AI Is Pro Consumer

New York is considering a law banning chatbots from doing anything that might be construed as legal work (SB S7263).

Banning AI legal advice is stupid and anti-consumer.

Obviously, there is the potential for consumer harm in chatbots providing legal advice, but the benefits far outweigh the risk. The biggest benefit, very clearly, is leveling the playing field.

All the lawyers are obviously gonna have AI, and the potential for lawfare and flooding the zone with abusive lawsuits is extraordinarily high. Contracts are gonna get longer. Legal fights will get more complicated. We’re all gonna be drowning in paper because of AI.

If the lawyers are the only people that can use legal AI, then only the people with lawyers on standby and large legal budgets will have access to those tools by proxy.

Without legal AI for consumers, we’ll see the most massive re-weighting of the American legal system towards landlords, corporations, and employers in history. Consumers and small businesses will suffer massively when they are going up against a constant barrage of frivolous lawsuits, long contracts, and punitive agreements.

We are moving to an increasingly hostile legal environment, and consumers need access to good tools to have a chance.

This law being proposed in New York is reflective of the legal cartels’ scarcity mindset and anti-competitive position.

Clawd - Dean W Bell

Dean Bell was an AI advisor to Trump and writes an extremely convincing, and terribly bleak, essay about how the fight between Trump and Anthropic represents an acceleration in the death of the American project/republican.

The gist of it is that Congress has become so calcified and incapable that lawmaking now happens on an almost entirely arbitrary executive basis with no recourse. That power is awesome and terrifying and is now wielded by people too stupid and reckless to take it seriously, let alone operate with constraints:

Even if Secretary Hegseth backs down and narrows his extremely broad threat against Anthropic, great damage has been done. Even in the narrowest supply-chain risk designation, the government has still said that they will treat you like a foreign adversary—indeed, they will treat you in some ways worse than a foreign adversary—simply for refusing to capitulate to their terms of business. Simply for having different ideas, expressing those ideas in speech, and actualizing that speech in decisions about how to deploy and not deploy one’s property. Each of these things is fundamental to our republic, and each was assaulted—not anything like for the first time but nonetheless in novel ways—by the Department of War last week. Most corporations, political actors, and others will have to operate under the assumption that the logic of the tribe will now reign.

There is something deeper about the damage done by the government, too. The Anthropic-DoW skirmish is the first major public debate that is truly about where the proper locus of control over frontier AI should be. Our public institutions behaved erratically, maliciously, and without strategic clarity. Our political leaders conveyed little understanding of their own actions, to say nothing of the technology and its stakes. They got off on an extraordinarily bad footing, and it is hard to imagine them ever recovering, because they do not seem to care about improvement. They are a cartoonish depiction of the American political elite, but sadly their failings have been the prototype of American political elites from both parties for much of my life now. “The same as before, but now noticeably worse” has been the theme of American politics for 20 years.

As I have said many times before, neither liberal democracy nor capitalism works without strong property rights. Trump doesn’t respect either.

AI Safety Has 12 Months Left - Michael Dempsey

Dempsey with a good summary and POV of how dire things are for any prospect of safety being a serious thing which matters to anyone beyond the EA cult.

There is roughly a twelve-month window to embed [safety] into technical and social infrastructure before IPOs and competitive dynamics make it permanently impossible. Every mechanism that currently constrains how labs behave besides compute is either already dead or expiring, and what gets built in that window has to survive years of market pressure on the other side. The people who care about this have a rapidly shrinking amount of leverage to work with but perhaps in the past ~week have been handed a specific opportunity to reclaim some level of importance or opportunity.

I’ve long wavered on the idea of AI safety because I’m not a believer in fast takeoff superintelligent AGI - Skynet is not on my radar of threats. But it seems clear that there is increasingly nothing governing the use of AI but technical limitations and Dario’s benevolence.

Hiring

I’ve been helping a number of portfolio companies with hiring, which has been really fun. In talking to these founders about their hiring plans and talking to the candidates about what they want, it’s become very clear that the paradigm of good early-stage startup employees has shifted significantly and rapidly over the past few years.

The best early stage hires are radically different from a few years ago: multi-hyphenate, commercial generalists.

The engineers want to talk to customers and the business people write code.

High agency and AI native instead of heads down 10x performers.

Some exciting roles/companies to join:

• Phoebe (network of agents for healthcare labor) is hiring engineers and salespeople in NYC

• Atomic is hiring rust engineers to build an AI-native replacement for git

• [Stealth] is hiring engineers in NY to build security and observability for agents

Building network effects after agents

Software margins are compressing and pure application software is becoming a fundamentally different, worse category.

But network effects remain among the most reliable paths to durable software businesses. Switching costs are the real driver of pricing power and consequently margins and long term earnings. Network effects are one of the best ways to create them.

Aggregation is not a network effect

A lot of what people call “network effects” are actually aggregation advantages. Unless my experience gets better because you use the same system these are single player experiences without strong nfx. The value is just “we put all the options in one place.” That’s a feature an agent replicates trivially.

The Citrini note on DoorDash exemplifies the confusion, ”you’re hungry, you’re lazy, this is the app on your home screen.”If that’s all it is, then yes, agents destroy it. A vibe-coded delivery app can show you restaurants.

But merely surfacing all the restaurants in an area is not particularly hard with agents. Managing reputation, mediating disputes, routing and bundling orders across a network of drivers: all of these deliver value by having multiple parties transacting at once. More drivers makes routing better. More orders makes bundling possible. More transactions makes reputation meaningful.

Finding and integrating suppliers is not the hardest problem anymore.

DoorDash is a sin eater enabled by multiplayer mode: it absorbs logistics risk, quality risk, labor coordination, etc. and quite literally handles the (long) first and last mile.

Agents commoditize aggregation, not network effects. Aggregation creates value. Network effects capture it. It’s easy to assume they go hand in hand because they usually showed up together. Agents break that coupling.

The OG network effects business (telecommunication) has no aggregation effects.

Where agents create new network effects

Agents don’t just leave existing network effects intact. They expand the surface area where network effects can form in three ways:

1. Networks of agents. Agents themselves become nodes in networks.

2. Agents as an on-ramp. Lower onboarding costs and easier participation.

3. Using aggregation to bootstrap nfx. Starting in single player mode and delivering demand side economies of scale later.

1. Networks of agents.

Agents themselves become nodes in networks. Not “agents help humans use networks better” but agents transacting with other agents, creating coordination layers that didn’t previously exist. When an agent representing a worker interacts with an agent representing an employer, or an agent representing a patient interacts with an agent representing an insurer, you get a new topology: a network where the participants are machines acting on behalf of people.

These networks get more valuable with density the same way human networks do. More agent-to-agent interactions means better matching, richer data, and higher quality outcomes. The difference is that agents can participate at volumes and speeds that humans never could, which means the network effects compound faster.

2. Agents as a new UI, expanding who can participate.

Some networks should exist but don’t because the interaction cost was too high for humans to bother with. The UI was too complex, the onboarding too heavy, the workflow too manual. Agents lower that cost to near zero, so dormant network effects activate.

People can interact with networks via agents instead of a UI. This means markets that were previously bottlenecked by complexity can now unlock participation. The surface area to build network-effective-driven businesses/products expands because you can onboard people invisibly.

3. Using aggregation to bootstrap your way to network effects.

Aggregation is getting easier to replicate. That makes it cheap, not useless. And cheap aggregation is a powerful bootstrapping mechanism for building real network effects.

You can brute force your way to a marketplace, kind of like DoorDash did with restaurants by calling in orders manually before restaurants signed up for the service. You start with an agent that delivers valuable single-player utility, and it backflips into a multiplayer network as a byproduct of individual agent actions. The single-player mode is genuinely useful on its own, and the network forms as a byproduct, not as a prerequisite. This solves the cold start problem that has killed marketplaces for decades.

Three Slow portfolio companies are each doing some version(s) of these: Phoebe in home care, Ando in hourly workforce, and Superdial in healthcare administration/payments. They’re building networks of agents, using agents to expand participation, and leveraging easy aggregation to bootstrap toward density and nfx.

I’m long network effects

Even as software gets squeezed on margins, network effects are still a reliable way to differentiate the commodity code and build durable, valuable software companies.

The generic conventional wisdom says “agents destroy moats.” The reality is more specific: agents destroy aggregation moats. True network effects, where the product improves with more participants, not only survive but become viable in more categories/modalities.

Citrini is wrong and intelligence will be a tailwind for NFX even as it wipes out commodity aggregation.

We’ll see more businesses building for network effects because there are both new surface areas (agents themselves as nodes in networks) and new modalities (UI-less network effects, invisible onboarding) and new GTM approaches (single-player-to-multiplayer backflips, cheap aggregation as a wedge into deeper network plays).

Read more:

• Modeling software after SaaS: The software companies that succeed in this brave new world won’t be pure application software and they certainly won’t be traditional SaaS. But they will be bigger and throw off huge amounts of cash to their shareholders... at permanently lower margins.

• Creating margins after SaaS: Software is definitely not dead. Nor is SaaS. But pure software will be a much different (worse) business than it is today.

Elsewhere

The 2026 Global Intelligence Crisis - Citadel Securities

Pretty remarkable to see this response, which is itself remarkable

Displacing white collar work would require orders of magnitude more compute intensity than the current level utilization. If automation expands rapidly, demand for compute definitionally rises, pushing up its marginal cost. If the marginal cost of compute rises above the marginal cost of human labor for certain tasks, substitution will not occur, creating a natural economic boundary. This dynamic contrasts sharply with narratives assuming frictionless replication of intelligence. Even if algorithms improve recursively, economic deployment remains bounded by physical capital, energy availability, regulatory approvals, and organizational change. Recursive capability does not imply recursive adoption

Remember, diffusion and implementation are massive tasks and huge opportunities. Evidently OpenAI agrees!

Are You “agentic enough?” - Wired

I talked to Max from Wired about the newly in-vogue/paradigmatic startup employee: multi-hyphenate, commercial generalists. Everyone wants high agency, AI native employees. The engineers want to talk to customers and the business people write code.

High agency and AI native instead of heads down 10x performers.

Abundance agenda 2.0

Against the backdrop of yesterday’s absolutely insane, fever dream Trump-Mamdani confab 2.0, Abundance NY is hosting a great event on Tuesday with the Wagner School.

This political moment needs policy ambition to match and ANY is continuing their great work to make it so.

Last week I wrote about the big SaaS sell-off in public markets and the doom it might spell for pure application software companies. My central claim is that AI undermines the three historic pillars of software businesses:

1. Zero marginal cost of reproduction (same code, infinite users). Inference introduces real CoGS.

2. Non-ephemeral value (software doesn’t depreciate like media, so you keep selling across time). Shorter lifetimes as tooling improves faster.

3. High switching costs (mission-critical or highly differentiated products drive retention and willingness to pay). Software becomes less differentiated/easier to replicate, and easier to switch between when it’s an agent not a person using it.

So even if SaaS isn’t dead (and again, it’s not and the public names won’t all get wiped despite the correction), the economics are getting permanently worse. Every line of the income statement gets hit.

The threat to software is not demand destruction. It’s inference costs, competition, and commodification. There will be some marginal TAM compression from seat-based pricing erosion and homebrewed software. But the biggest source of pain will be the contribution margin challenges posed by less efficient S&M (owing to both competition and channel decay), inference (baseline assumptions should no longer be zero CoGS), and shorter LTVs (easier to swap out tools when they become agent end points).

It’s not all doom and gloom. I remain very enthusiastic about the potential to decommodify software to drive high switching costs and earn high margins even if pure SaaS has a challenging future. And more optimistically, as software can do more, the end markets will get much bigger.

The big trade in software will be size for simplicity.

You can get into bigger TAMs with larger contracts if you can tolerate more complex operating models, higher cost structure, and lower margins. Scaled, successful software companies won’t default to pure SaaS but they will be bigger and throw off more cash (in absolute terms) than the prior generation.

You can also run a bootstrapped or vibe-coded software business very profitably on a per-employee basis but likely not at scale. This is the other side of the trade: cash flow for size instead of size for margin. Numerous, easy to start, usually not very big, little or no need for outside capital at inception. A perfectly valid path that looks more like a traditional small services business or consultancy in terms of earnings and terminal value.

Vertical AI / BPO Replacement (from VSaaS)

The bull case for vertical AI is basically the BPO replacement trade whereby you can earn your way into much bigger TAMs/contracts because you sell into a services/headcount line rather than a software line item.

Slow port cos Phoebe, Superdial, and Ando are doing this really well by building toward a combination of nfx and high liability workflows. Each starts/started with high liability “single player” workflows that open up opportunities to get into two sided businesses that are hard to dislodge or replicate. That’s a key feature of this upside case: doing something valuable out of the box to earn access and build defensibility over time. Remember, startups almost definitionally can’t have moats early or they couldn’t start themselves!

But more broadly, vertical AI is an undifferentiated business in its default state. You have little pricing power because you’re selling commodity labor replacement and replicable code. Not a lot of long-term margin in that unless you can command higher prices through leverage elsewhere in the business.

This is probably the most challenging/tenuous option because this will be the easiest to compete with on a pure (vibe) code basis. Without some combination of NFX or “sin-eating” (doing something dirty/risky/liability generating that people want off their hands), you will trend to look more like the SaaS bear case from last week.

Services to differentiate software (forward deployment)

The forward-deployed model uses services as a loss leader to earn the right to bigger, longer-term software contracts. You don’t really make money on the services even if there’s revenue in it for you.

The point of forward deploying (often but not always engineers) will be unlocking/earning bigger and more retentive contracts at high margins (even if still a bit lower than today’s baseline because of inference). It’s reasonable to think of this as re-allocated, margin-generating marketing for the software business. And because you’re selling something clearer (”we’ll do the thing for you and it’ll work”) and you’re ultimately responsible for success, not them, the sales cycle is more efficient per dollar of revenue even though it’s consultative.

There’s some financial juggling in these businesses where product and sales headcount functionally gets reallocated into CoGS (note: Palantir doesn’t report this way and instead pushes CoGS into G&A and R&D). But the net effect is higher ACVs, higher LTVs, and a bigger albeit more complicated business.

In aggregate the business earns better because the services piece creates switching costs and value to customers that pure software can’t.

And of course the bear case here is that you just wind up as a consulting company doing custom work with no repeatability, no long-term contracts, and no real durable enterprise value. This is a big risk right now in the diffusion trade (hiring AI researchers to do custom work and calling it a lab).

Vertically Integrated Services (software differentiating services)

This is the inverse of the FDE model. You start with a services business and invest heavily in software to make the services line more profitable. Easy to say and hard to do.

The baseline is a services company with low gross and net margins. If you can build a product/product org that radically transforms the margins, you can bear that new corporate cost, afford to aggressively invest in growth, and own all the upside from this transformed P&L. GM expansion is the main thing and requires you to fundamentally be a product company.

You can build these companies either organically (from zero) or inorganically through acquisitive means, a growth buyout (GBO) where you buy services companies and transform them with software while owning the upside. The inorganic version may ultimately look the same at steady state/scale but with more balance sheet complexity along the way (debt to borrow and service, integration, etc.).

Here the bear case is simply that you fail to pull off the economic transformation. The best way to mitigate that risk is by building product first, in both the organic and inorganic case. Jumping straight to providing services before you can do it in a differentiated way is a capital intensive way to own an undifferentiated asset you have to run forever.

Other options

Left unexplored here is the hardware + software model that creates margin and differentiation through supply chain, physical switching costs, proprietary data via sensors and devices. It’s very much in vogue right now as investors look for the downside and durability of industrials with the upside of software. That’s basically the entirety of American dynamism: defense, industrials, robotics. Claude is not going to go start building houses or tractors.

There’s also a whole host of companies that will also have radically different income statements by investing preposterous sums in data centers and/or foundation models. The foundation model companies exemplify the size for simplicity trade by being bigger than we thought possible but very obviously lower margin, especially when you factor training into CoGS, which you obviously should since the models depreciate so quickly.

I’m not specifically modeling NFX here either because that’s a feature of many businesses, not a business model or company structure unto itself. You can/will see network effects in each of the types above (including hardware and AI capex companies).

What else do you think I might be missing?

Conclusion

The SaaS era trained everyone to value simplicity; SaaS was “the best business ever conceived” in no small part because it felt/was so easy to model and understand. Zero marginal cost, high margins, retentive revenue, clean P&Ls. The next era of (application) software, as the code itself commoditizes, rewards the opposite trade. The biggest opportunities belong to companies that decommodify their code to access TAMs and ACVs that pure software never could.

Software will be more important over the next decade than it’s been over the last: cheaper to produce, more powerful to use, responsible for orchestrating and accelerating more GDP.

The income statements will look “worse” by SaaS standards and better by every other standard. The software companies that succeed in this brave new world won’t be pure application software and they certainly won’t be traditional SaaS. But they will be bigger and throw off huge amounts of cash to their shareholders... at permanently lower margins.

As a professional venture capitalist, it’s my obligation to write about software valuations and the big SaaS sell-off in the public markets. This is my “end of SaaS” post. There are many like it, but this one is mine.

Software after SaaS: sin eaters, hybrids, and NFX.

Software is definitely not dead. Nor is SaaS. But pure software will be a much different (worse) business than it is today.

The TAMs will shrink or at least stop growing. Pricing power will decline. Margins will compress. Multiples will stay depressed.

Margins in SaaS depended on three assumptions:

1. Zero marginal cost of reproduction (same code, infinite users)

2. Non-ephemeral value (software doesn’t depreciate like media, so you keep selling across time)

3. High switching costs (mission-critical or highly differentiated products drive retention and willingness to pay)

AI degrades all three. Marginal costs go up with inference. Value depreciates faster as new tools improve rapidly. And switching costs collapse as software becomes less differentiated/easier to replicate, and easier to switch between (does the agent care what software it’s using?).

The economics of software businesses are getting permanently dented.

The net result: a business that used to earn 25% income margins might earn 10-11% going forward. In the future/near-present, you can think of pure application software having similar dynamics to traditional services business today: very numerous, easy to start, usually not very big, little or no need for outside capital at inception.

It’s definitely still a business but it’s a fundamentally different kind of business. The assumptions of perpetual revenue growth and stable high profits that undergird SaaS multiples go away for many/most of the household names that can’t successfully find defensible positions.

With all that said, the TAMs won’t go to zero and public equities won’t get permawiped, because no serious person/leader/company that actually wants to win is going to homebrew everything, switch vendors constantly, or buy from tiny upstarts.

• Leaders want a throat to choke and purchasing decision they can defend (“no one gets fired for buying IBM”)

• You want to crowdsource best practices from other customers/you don’t want to discover every commodity from first principles

• No one wants to be responsible for maintaining undifferentiated cost centers, even if they can build an 80% solution with relative ease.

Everything is a capital allocation decision evaluated in times of relative IRR, even if only implicitly. Businesses want to focus only on the things that they do best and allocate their resources to the places where they can produce the most differentiated returns. Even if codegen changes that math it does so largely by making pure software companies more competitive to run, not by introducing tons of internally vibe coded competition.

This is why Anthropic and OpenAI don’t rebuild Slack internally. It’s a pain in the ass and doesn’t drive their business. And it’s why white shoe professional services like consulting and law won’t go away. You’re rarely paying purely for the outputs - especially not when they seem rote.

So: software isn’t going away. But SaaS as a business model, defined by 80-90% gross margins, is not the default anymore.

Ultimately the solution to this new problem for software lies in #3 (switching costs) not in ephemerality or replicability.

Switching costs are the real driver of pricing power and consequently margins and long term earnings. If you can dictate pricing substantially above your costs without customers leaving, you have pricing power. It shows up in the P&L as margins. Whether those switching costs derive from brand, network effects, product value, monopoly, or usage patterns doesn’t matter.

”Software is a business tool, not a business model.” - Sam Lessin

The future of software after SaaS:

Build for network effects. Create switching costs through networks and multiplayer dynamics such that your value gets better with scale and the cost to compete goes up for any new entrant. There are specific places where network effects might erode in the face of agents but overall this is part of the physics of business and will generally persist.

Become a sin-eater. If you can absorb and make smooth the most jagged, ugly part of a business, you allow them to wash their hands of something they desperately want off their plate, which is very often risk. The code isn’t the product; the code is the interface for high-stakes, operationally intensive coordination. Think: flows that are too high volume/intricate to do manually but too high stakes to do generically. This is why security, infra, fintech, etc all still have a future. Nihar calls this “embracing liability.”

Roll hybrid models. Decommodify code by bundling it:

• Software + services: use services to differentiate software. This is why “forward deployed” businesses are so in vogue right now in the enterprise.

• Hardware + software: physical switching costs. Generate lock-in or create proprietary data through sensors, devices, and physical infrastructure.

• Services + software: use software to differentiate services. This presents as vertical integration, whether as a de novo business or an acquisition platform.

There are other advantages that can create durable businesses: brand, trust, regulatory capture, cost of capital, distribution. But those have to be earned over time and you generally can’t plan for them at inception (or some people can but YOU probably can’t).

Software will be more important over the next decade than it’s been over the last: cheaper to produce, more powerful to use, responsible for orchestrating and accelerating more GDP. The challenge/opportunity/profit is in decommodifying the code and building good software businesses: align with rather than against the foundation models such that their improving capabilities are a boon not a headwind to you. The latter option (rooting against the models/betting on their limits) only makes sense within some narrow domains, software engineering not being one of them.

As the models improve software capabilities even more than they already have, the SaaS TAMs will be smaller but the software TAMs overall will be MUCH bigger. There will be more software companies and the most successful of them will be hugely profitable and valuable - not just in financial terms but in terms of impact.

Connecting the foundation models to high value complex real-world use cases is the biggest opportunity for wealth creation in recent memory. It will happen through new (often hybrid) business models and products rather than re-running the cloud playbook.

The software companies that succeed in this brave new world won’t be pure application software and they certainly won’t be traditional SaaS. The physics are moving too much, too hard, and too fast against the business models and the products having an economic future.

Read more

• The end of pricing power (2025)

• Implementation Is the Task of Our Time (2025)

• The Future of Software Looks Like Services (2023)

Libbing Out with Eric Newcomer

I sat down with my friend Eric Newcomer on his podcast to talk about AI rollups/growth buyouts, our creator fund, and Trump/the tech right.

Some choice quotes from that conversation:

On the criticality of democratic institutions for capitalism

A lot of people don’t take seriously the extent to which our system of entrepreneurial capitalism rests on strong institutions. What Trump is doing to our democracy, he is also doing to our market system...

It should be extremely concerning to anyone who calls himself a capitalist to watch him start nationalizing companies, which is what he has done in the steel industry and in the chips industry. The systems of democracy, capitalism, property rights, institutions and independent judiciary, these are all actually one thing. In countries that have one, they generally tend to have all of them. In countries where one fails, the rest fail shortly thereafter...

Even in the most strictly amoral terms, if you don’t care about any of the ideas about the liberal world order, we’re going to wind up looking like Turkey or Argentina or Russia where there is no meaningfully free market system, there is no central bank independence, there are no strong institutions...

He’s doing third world style corruption in the US and is substantially undermining the excellence of American institutions and the excellence of American capitalism and we are all going to wind up paying for that.

More on kingmaking and consensus

The problem with kingmaking is it’s not putting the crown on the head of the crown prince upon the death of his father. It’s putting a crown on a baby out of the womb and believing that the crown itself confers legitimacy instead of the legitimacy conferring the crown.

OpenAI was so profoundly non-consensus from its start, it wasn’t even a for-profit company. Anthropic people thought was an insane thing to even attempt. Cognition started as a video conferencing app. Clay wandered in the desert for 10 years. Ramp, everyone was like what the hell are they doing? Don’t they know about Brex? It’s worth five times Brex now.

Winners always look inevitable in retrospect. I just don’t believe in that as the reality on the ground.

VC-Backed Startups are Low Status

A couple times a year, Mike Dempsey writes something really excellent. I really think he’s one of the best thinkers/writers in venture.

When the institutions are optimized to fund the legible thing and the individuals are optimized to build the legible thing, the identity of the founder itself fades. Being a founder used to carry with it the implication that you had seen something others hadn’t, that you were willing to be wrong in public about an unlikely future you believed in, and take a risk of banging your head against a wall with high career risk for a long time. Now it is almost eye-roll inducing in many non-tech circles as people struggle to distinguish Another Founder with Another Launch Video building Another (insert zeitgeist here) Startup.

[…]

Investment banking had 2008. An easy to see crisis that gave people moral vocabulary to dismiss it and started the march downwards of social status throughout the 2010s. You could say “I don’t want to be part of that” (even if it was painting a broad brush stroke on the entire finance industry) and point at something concrete.

The venture-backed startup path doesn’t have its 2008, but instead has has something more diffuse and possibly harder to reverse: cultural exhaustion.

Read the whole piece. I quite liked it.

Will Manidis wrote a great piece on becoming “legible to Capital.”

Think of an idea being legible to Capital as being magnetically charged to the capital markets. Dollars are just attracted to you: allocators discuss it at cocktail parties, funds market it at their AGMs, and twitter discusses novel financing schemes to get more dollars behind the idea.

He’s certainly right in identifying and naming the phenomenon. But he’s largely making a retrospective descriptive claim, not a prospectively normative one. That is, it’s obvious when someone has achieved it. It’s much less obvious how or what you should do.

And I’ve been writing about legibility for a while now. The frame I keep coming back to is that startups can be either legible (clear, easily understood) or illegible (hard to read), and that this distinction matters more than stage, more than traction, more than most of the things investors claim to care about. The job of true seed investors is to fund illegible opportunities to the point of undeniability (in contrast to inorganic/formation stage “king making”).

Will’s framing implies legibility is a property. Some companies have it, some don’t. On an ex post facto basis, the winners always appear inevitable and obvious. You look at Ramp or Cognition and you think: of course. That was always going to work. That was always going to attract capital, talent, and customers. But... airplane meme. Legibility wasn’t a precondition of their existence. It was an active process/outcome.

So yes, legibility to capital is one of the most powerful forces in the startup ecosystem. But it is a thing to work towards, not merely a thing to have.

In the past I had made a common category error here. Despite talking about companies as breaking down into two stages - legible and illegible - I had still been treating legibility as a curve rather than a step function. I had believed that if companies just execute, hit some arbitrary revenue milestone, the world would start to care. It won’t. You can put up great numbers and still be completely invisible to downstream capital because you never built the narrative architecture that makes those numbers mean something. Remember, if startups are just stories about the future, they are formless without that architecture

Legibility is brand for capital. And brand is promises kept.

The key white space is the narrow sliver of illegibility where an important story can meet with narrative and execution to make and keep a promise that matters.

That is while legibility is not fixed, illegibility might be. Put differently, the first most important question is whether or not you’re working on something that actually matters or has a promise big enough that keeping it is important. Read more here.

The single most interesting and important function of early stage is translating weird takes on important stories as told by illegible companies into narrative momentum and legible opportunities for downstream capital. Done right, you can transmute ideas into not merely good businesses, but important companies over time.

Institutions should go direct

Given that the single job of most scaled venture funds at this point is deploying max dollars into a handful of ~7 names, I don’t get why the top LPs need to pay them fees vs going direct.

You’re telling me that a mega allocator can’t pay [lab alum] $10M/year cash for access?

When there’s no more picking why pay a picker?

It would be substantially cheaper and they’d operate with more information/conviction if the biggest institutions just spun up teams with alums from these companies and paid them cash to get direct access to the best opportunities in the biggest names.

This is basically the same dynamic as institutions paying HFs 2/20 to buy the Mag 7…

Given how persistent it is despite being economically irrational, it must be a consequence of “non-obvious” incentives (I’m using scare quotes bc the incentives are actually so obvious).

The real incentive most of the time at most institutional allocators is to track, not to beat, the benchmark in ways are socially defensible and won’t cost you your job.

If you’re backing one of a couple of names which are in turn backing one of a couple of names, you have maximum defensibility. Unless you’re really compensated on upside, you have no reason to think about upside.

Elon’s Mega Co and the Slow creator fund

Elon has now merged xAI into SpaceX after already merging Solar City into Tesla and X into xAI. Very likely he will wind up merging SpaceX into Tesla and create one single ElonCo, which will eventually subsume Boring and Neuralink.

A single Elon MegaCo has always been the most shareholder-friendly, incentive aligned thing to do and has always been coming.

> If you are investing in any Elon Musk company, what you are buying is not so much ‘cars’ or ‘rockets’ or ‘robots’ or ‘social media’ or ‘brain implants,’ but rather that ability to allocate capital to the next big thing. - Matt Levine

The problem was that keeping these as separate companies created constant conflict. When Elon shifts GPUs from Tesla to xAI, or routes engineers from one company to another, he’s borrowing from Peter to pay Paul. His various shareholder bases are/were competing for his attention and access to free/cheap capital, which are the unique assets that make all these businesses work. Rather than have that zero-sum game play out across separate cap tables, everyone is obviously better served if they’re all just one company.

For certain entrepreneurs, discrete companies are intolerable constraints. The company structure puts them out of, rather than into, alignment with their capital partners. This is bad and requires new holding company approaches to capitalize and incentive their singular entrepreneurial abilities.

At Slow, we’ve been thinking about this for years.

Our Creator Fund takes equity in personal holding companies, not the individual companies creators spawn. We’re backing people for whom the right unit of investment is the person, not any single project they might pursue. They have audiences, entrepreneurial instincts, and the ability to spin up multiple businesses over time. The holdco structure lets us bet on them, not on one idea.

Elon is doing the same thing at trillion dollar scale that we’ve been doing with creators. Some people are permanent founders, and the capital structure should reflect that.

At Slow we’re largely focused on media entrepreneurs as a particularly and painfully undercapitalized asset class but clearly this phenomenon is both broad and deep.

Outtake’s Series B

Outtake raised a $40M Series B led by Iconiq. Outtake is a really cool new take on fraud and security. They proactively crawl the entire open web to build a trust network and proactively take down malicious/deceptive sites, ads, profiles, etc.

From their OpenAI case study:

Most alternative solutions still rely on third-party contractors to manually review flagged content—a process that can be slow, inconsistent, and expensive. Outtake⁠ reimagines that system with always-on AI agents that scan millions of surface areas, such as webpages, app store listings, and ads, per minute, building a map of trustworthy and suspicious entities. That map helps security teams understand what’s happening, who’s behind it, and route resolution recommendations for expert review in a matter of hours.

Megan and Sam led the seed round and Slow remains the biggest outside shareholder in the business. I had dinner with two of the early employees this week and am totally blown away by the quality of the team/depth of talent over there.

Attackers have always-on sophisticated AI. Defenders need the same.

We’re hosting a dinner with Work-bench in New York in March to discuss how AI is breaking cyber paradigms and what comes next. You can sign up to join us here.