Even Cracked Kids Grow Up

Happy fourth, the startup meta, Soham

---

Happy July 4th from Vermont! I’m definitely glad to be out of the city for the 4th this year. Even as there’s quite a lot to be sad and frustrated about in our country right now (the OBBB is just more regressive, decel, fiscally irresponsible nonsense), I find a lot of comfort in the idea that ‘there is nothing wrong with America that cannot be cured by what is right with America.’ The uniquely American spirit of optimism, humanism, and pluralism will prevail. On to the content.

---

Even Cracked Kids Grow Up

The current meta in startups is cracked kids. The younger and twitchier the better. Youth has been in since Facebook at least but now everyone wants to find the next “AI native” founder. And the consensus is that youth will translate to fluency/speed in era where velocity is everything. It might be right!

On the one hand, the narrative is great because it’s opened doors for talented young people to take big swings commensurate with their ambitions and do stuff that they wouldn’t otherwise be able to.

Young people are getting taken seriously as builders because they have some real advantages (energy, flexibility, availability) that can sometimes compensate for lack of experience, connections, etc. 

But it also reminds me of a particular risk/problem I noted at the height of the “Gen Z VC” thing a few years ago. In 2022 I wrote:

If your plan is to be valued within your firm primarily for your youth (or plan to get a job on that basis) you're eminently replaceable, increasingly so as you age. You are setting yourself up as a piece of marketing for the firm, not an investor. Youth is fine as a calling card when you have nothing else but it has to be an entree to developing something durable and unique.

Young founders will never go out of style but eventually this meta will end and some new archetypal founder will pop up. Eventually Michael Truell will be 30 too. He’ll have built Cursor - what will you have done?

I worry for those wrapping themselves in the flag of youth as so central to their identity. If your identity and value is being young, what happens when you're not? At some point, the story will shift to what you built/can do - take it from a guy who got old(er)!

H/t to Graham Carney for the inspo here.

Soham Parekh

ICYMI the big meme of the week was that one software engineer in India had managed to get jobs (and get fired from) dozens of startups simultaneously. He’d reach out, absolutely crush the interview process, not really work (too busy interviewing elsewhere) and see how long he could collect a paycheck before getting fired. Read more here.

The Soham stuff is obviously funny/great fodder for jokes. But there’s two real things in it.

1. It holds up a pretty damning mirror to the hiring and management practices at most startups. That the system was this hackable by a bad actor and someone got away with it for so long says pretty clearly that getting the job has become a skill unto itself divorced from doing the job.

2. It is a portend of what is coming for hiring and job markets overall. Soham is clearly clever and motivated but he is not a one-off. As the transaction costs for job applications go down, the systems stop working and become unmanageable. AI is making this a lot worse already with a second shoe ready to drop.

None of this is super surprising. In Feb I wrote:

AI-powered scale will make labor markets murkier and less efficient (more work for worse outcomes). Competition, infinite/global markets, sophisticated automations, and unsophisticated slop are going to kill performant channels.

Unlike sales, hiring is bidirectional and so the problems (and solutions)will be bidirectional as well.

Talent (supply) is increasingly bombarded with low quality inbound that is hard to parse. And hiring managers (demand) are overwhelmed with impossible to distinguish candidates. This is true up and down the labor/income ladder.

In the AI era, how do you source and parse talent?

1. Trusted sources and networks. Referrals and reputation will win.

2. Quality and skill assessments over resume screening.

3. High transaction costs. “Common apps” induce more, low intent demand which leads to quasi-random placements and a less efficient matchmaking process with sub optimal outcomes. Transaction costs reverse that.

And how does talent vet inbound and prioritize outbound?

1. Trusted sources and networks.

2. Talent agents in any non-permanent market (like nursing or trades). The “staffer in your pocket” will book the best jobs, instantly apply, and earn more. Don’t talk to recruiters.

Just like sales, hiring will be all about reputation, network, and authenticated attention versus mass campaigns. Spamming candidates and jobs won’t last as an efficient or effective SOP. The optimal outcomes won’t be scaled, anonymous, and efficient but hand to hand and trusted.

The age of inbound is coming to hiring.

Soham is a symptom of systems that are too easily hacked and broken by brute force.

I Read

As AI Infiltrates Call Centers, Human Workers Are Being Mistaken for Bots. I strongly suspect that the EU will require AI disclosure and a “right to speak to a human” within a few years. The US probably won’t do the same (maybe disclosure).

That Dropped Call With Customer Service? It Was on Purpose. More call center takes on the ways that companies incentivize frontline workers to create administrative and operational barriers in the hope that you’ll just give up.

For some people, that humanity gets trained out of them. For others, the threat of punishment suppresses it. To keep bosses happy, Tenumah explained, agents develop tricks. If your average handle time is creeping up, hanging up on someone can bring it back down. If you’ve escalated too many times that day, you might “accidentally” transfer a caller back into the queue. Choices higher up the chain also add helpful friction, Tenumah said: Not hiring enough agents leads to longer wait times, which in turn weeds out a percentage of callers. Choosing cheaper telecom carriers leads to poor connection with offshore contact centers; many of the calls disconnect on their own.

China's Clean Energy Boom Could Win the Race to Power the Future. This is just so damning. The US is ignoring the technologies of the future and committing to a path of expensive, dirty energy while also ceding soft power and influence to China. As I’ve said before, Trump is a degrowth, decel, anti-abundance president.

‘Thrilled’ California Developers Cheer Rollback of State’s Environmental Law. This is shit headline. What happened is that CA has expanded by-right permitting for infill development in urban areas. This is just great news on the housing front and should, over time, be a meaningful lever to drive down housing prices in the state.

DraftGPT: The Brave New World of AI Hits the NBA - The Ringer.

Teams have begun incorporating machine learning into their player evaluations, including to scout this year’s draft class. One model even claims to be able to predict NBA success based solely on player sound bites.

Superdial's $15M Series A

Building the agentic present and AI clearinghouse future for healthcare

---

Note from Yoni: thanks all for the well wishes and congratulations on my impending nuptials (mentioned last week). I had a great weekend with my buddies upstate. Now I have another bachelor weekend starting today (my friend who’s getting married the weekend after me). Pray for me.

---

In May of last year I wrote:

[AI agents] will be a clever hack to launch new marketplaces and platform and software businesses that wind up looking rather like their “Pre-AI” counterparts.

Agents talking to agents or people is dumb and expensive, but unavoidable; it shouldn’t be the end goal and it doesn’t have to be. Instead LLMs should be a bridge to turn more of the world into APIs and transaction layers.

I described it as “Jevon’s paradox for conversational AI” whereby LLMs would

1. dramatically lower the cost of calls to customer service (just get the agent to do it for you), which will

2. increase calls volumes proportionally, which in turn would

3. necessitate an agent on the other side to handle all that increased volume so that you could eventually

4. replace agent <—> agent calls with a data exchange or clearinghouse

Basically you can brute force your way to a marketplace or data product, kind of like Doordash did with restaurants (just calling in orders manually before restaurants signed up for the service to handle the volume).

Now Superdial has raised a $15M Series A from SignalFire to build the agentic present and AI clearinghouse future for healthcare providers and insurers.

Today, Superdial is helping providers and insurers automate phone calls. In the short term, life gets a lot better for providers and their admins when they outsource rote and tiresome outbound phone calls. They save money, get paid more, and can spend more time on patient care.

With enough density and closed loop agent-to-agent calls, it can flip to a network model and obviate voice calls altogether. That work is just beginning.

We seeded Superdial in 2022 and the company has now raised ≈$20M in total. The business has grown dramatically since then by helping customers improve workflows, reduce cost, and drive quality around health insurance admin.

Superdial is laser focused on doing vertical AI the right way: using it as a way to win customers and ultimately build a deeper, nfx-driven business rather than maxing out as a point solution.

Superdial was founded by Sam Schwager and Harrison Caruthers who are exceptionally gritty, high velocity, high agency. They’ve become great leaders and it’s been super fun to have a front row seat.

On a personal note, this was the first investment I made at Slow and this is now the third time I’ve backed the company. It’s been a long road and the future is always high risk/highly uncertain, but this is a great milestone.

They are hiring across engineering, sales, and design for their NY-based team.

You can watch Sam’s appearance on TPBN here - fun time starting at 2:50. And you can read more about Superdial in Fierce Healthcare, Forbes, WSJ Pro, and Axios.

I Read

Young People Face a Hiring Crisis. AI Is Making It Worse. - Derek Thompson

• “The most dramatic takeaway from these conversations wasn’t that AI was clearly destroying jobs. It was something I wasn’t expecting to hear at all: AI is shattering the process of looking for jobs.” The whole article is worth reading but this has been increasingly obvious for some time.

• I wrote in Feb “Just like with sales, AI-powered scale will make labor markets murkier and less efficient (more work for worse outcomes).”

10 thoughts on the NYC Mayoral Election - Better Cities

• Not my own but sums up my thinking reasonably well. Catherine and Ryder also wrote it up well in the Abundance NY Substack.

OpenAI and Microsoft Duel Over AGI in High-Stakes Negotiation

• Really well reported in The Information. It’s gonna be hilarious when a court has to decide whether or not we’ve achieved AGI as part of a byzantine licensening deal.

The rollups are rolling

Plus my NYC Ranked Choice ballot

---

Yoni’s Note: I’m at my bachelor party right now! While the newsletter doesn’t take a Friday off, I sometimes do. So this week is a bit lighter on my writing and much heavier on some hot takes and stories from around the web.

Read to the end for my ranked choice ballot.

---

AI Roll-Up Deals Accelerate as Top VC Firms Edge Towards Private Equity - Madeline Renbarger

Long Excerpt followed by a take from me:

Large VC firms like General Catalyst, Thrive Capital, and 8VC are increasingly running the private equity playbook of buying legacy businesses and trying to improve their performance. The special sauce is AI, these investors believe, and half-a-dozen top-tier firms have now backed efforts to combine established companies in the same market and supercharge them with artificial intelligence [...]

Slow Ventures has backed three roll-ups still in stealth — one is buying private healthcare practices, another is rolling up system integrators, and a third is for medical billing, according to sources familiar with the deals. Slow on its website cites “the future of healthcare is private practice” as one of its signature theses, and its new deal in that sector has also landed checks from 8VC and Better Tomorrow Ventures.

I’m just honored we made the list as “notable” investors!

More skeptical investors pointed out that PE is a different business with much lower expected returns than traditional VC deals. It makes sense for the big funds to be branching out in this direction, but it will undoubtedly prove to have many challenges of its own.¹

Now look, if you’ve followed me long enough you’ll know what my take is (long form deck). In short:

1. Software will eat the world but the best business model won’t always be SaaS. Sometimes it makes sense to buy and transform an asset yourself.

2. Buying companies is expensive (duh) but it’s also hard. Lots of soft costs in source, diligence, negotiate, finance, integrate, etc.

3. You need to focus first on value creation (product) before M&A. Let the main thing be the main thing and only buy if the product absolutely rips (generates tons of value). You wouldn’t hire a sales team before building a product and M&A is just a go to market motion here so build then buy.

4. And when you buy, you need to buy increasingly large assets vs an increasing number of small ones (unless, like Teamshares, the act of buying is core product/IP). If there are no medium and large assets to buy, the strategy risks topping out too soon.

5. Use debt. The point is to be maximally equity efficient on the belief that selling software is often inefficient. So don’t light equity on fire and undo that.

So will many of these plays ultimately look more like PE? Yes. And is that because funds are so big they need to find ways to deploy, even at low returns? Yes. But that’s not all or perhaps not even the majority.

There’s room for venture/risk capital in the early phases of these companies if you do them right. The first tranches of capital are (should be) relatively small, highly uncertain, with uncapped upside. That’s venture risk! And in that case/if you stick to this plan, we think the returns will be there in the long run.

Early rounds may be more dilutive than normal software venture rounds (you’re pulling forward capital) but what matters is terminal size and ownership. The later rounds here aren’t venture-y but neither are the later rounds into software and AI companies!

Read more on this topic:

• The PE x VC Gigacycle is Fully Underway

• The VC x PE Supercycle (since upgraded to a gigacycle)

• Our Buyout Thesis

• PE is the great new villain

• Fragmentation and market structure

• Our Private Practice Thesis

My NYC Mayoral Ballot

NY has an election on Tuesday June 24 (early voting is already open!) and it’s ranked choice. Instead of picking one person, you can rank up to five in order.

1. Brad Lander: Brad Lander has been my #1 choice since he announced his candidacy for Mayor. He has the right ideas as a pragmatic progressive and has a track record of accomplishments as both Comptroller and on the City Council. He has the vision for how to improve the city on housing, safety, homelessness, etc. and the managerial ability to actually back it up. Lander is not a lightweight and as Comptroller has been extremely tight on technical matters of policy. He has the most ambitious plan for homelessness and among the most ambitious on housing and infrastructure. He has been a dedicated public servant for years and is well liked across government. Brad Lander is a man with a plan who can get things done for New York. FWIW he is also the top choice of the NYT and Abundance NY.

2. Zohran Mamdani: Polls have it very tight/a tossup and Cuomo is absolutely unfit for office (more on him later). So while I disagree with Zohran on many of his key policies, he strikes me as an energetic person of high character who, crucially, is open-minded and thoughtful on policy. He has changed his tune on crucial issues, telling the NYT that the role of the private sector in housing affordability is the biggest thing he was wrong about in the past. He’s also been a quick study on issues near and dear to me like government efficiency and supply side liberalism/abundance. My hope is that Zohran’s worst ideas would be constrained by political realities and his best qualities would be channeled through competent lieutenants (probably Lander as Deputy Mayor). Lander and Zohran have cross-endorsed and so have I (for what it’s worth).

3. Zellnor Myrie: Zellnor has BY FAR the most ambitious plan for housing and I love that. He wants to build 1 million new homes in ten years and it’s not just a number. He’s in the weeds on how to make that happen. The city needs more leadership like that. It’s a very long shot campaign but I hope he sticks around in city politics and continues to push NY to dream for more.

4. Whitney Tilson: look, he’s 4th for a reason. He has little to no personal appeal to me BUT I think he would do a reasonably good job on housing which is my #1 issue. Beyond that, Tilson would probably be fine. NYC deserves far more than fine but fine is also far better than awful.

5. Michael Blake: he was the only candidate to say that firing Thibs was the right call which tells me he’s not afraid of making hard, even unpopular decisions for the good of the city. That earns him my 5th and final spot ;)

Bottom line: whoever you rank, do not rank Andrew Cuomo. Even putting aside Cuomo’s sordid personal history of sex abuse AND his mishandling of the pandemic AND his criminal coverup of both his sex crimes and covid disaster, he is totally unfit for office.

But also don’t put those aside. They’re legitimately disqualifying!

Cuomo’s #1 claim is that he’s a competent, no-nonsense wrangler of government. But everyone in government hates him and his track record as a people manager and leader is atrocious. As governor, he pushed out all the best civil servants and let feuds get in the way of policy. If elected, Cuomo would use the mayoralty as a vehicle for retribution and animus, not improving the city. He is morally and managerially unfit to run the hugely complex bureaucracy of the city and lacks any ambitious plans for what to do if elected. Do not be fooled: political moderation is not the same thing as pragmatism. He has very little to say substantively on matters of policy, no meaningful plans for the city, and will have very little ability to execute.

I Read

• They Asked an A.I. Chatbot Questions. The Answers Sent Them Spiraling. There is a ghost in the machine for those ill or susceptible enough to find it. I’m very receptive to the idea that “social media algos are unaligned AI.” Viewed through that lens, this looks pretty similar to some of the pernicious ratholes people fall down online - but worse.

• Octopus Energy’s software platform is emerging from parent’s shadow. Really cool company that built an operating business in energy to validate a product before spinning out the software. Adjacent to some of the ideas I’ve been writing, thinking, talking about. Just fantastic execution of a super impressive approach.

• Substack in Talks to Raise Financing as It Taps Into the Zeitgeist of Trump II Substack is doing much, much better than I would have thought. $450M of GMV with $45M of revenue. That’s not nothing! I still wish I could just pay them for hosting this rather than worrying they’ll boot me for abusing their generosity but it does feel like a smaller risk if they’re successful. I’ve been writing on substack nearly since the company’s beginning and am rooting for them.

1. I found an even longer list here: https://www.ai-rollup.fyi/companies

Six Themes for 2025

Where I'm investing next, Nuclear Waste, 23andMe, Su*c*de Rounds

At the end of last year I wrote “when to call me” - some theses and themes I was/am interested in. That stuff all still matters but a lot of it was quite backward looking on what I ~had~ done.

Here’s what I’m thinking about now and am going to do going forward:

1. AI Extrapolations: the first order impacts of AI are pretty obvious (lower cost digital stuff, personalization, observation, and coordination). But so what? What happens next? Rather than debating the inputs, we need to be tracking the outputs and the ripples from there. It’s at the 2nd and 3rd order impacts that things gets important and unique opportunities open up. Stop asking what will happen - we know what will happen - start asking what happens next.

2. Age of Inbound: The cold email and autodialer are going the way of the banner ad. The Great Ignore has begun. AI is killing traditional outbound sales channels and it’ll only get worse. Too many competitors, too much slop, too easy to scale what works. The distribution models of the future will all be based on some combination of captive/proprietary distribution, virality, and organic growth, and inbound demand. The age of inbound is upon us.

3. Going early and long NY: as much of the world becomes more formulaic and scale oriented, the opportunity to be a positive-sum truth teller and supporter to/for the earliest companies gets bigger. When capital is cheap, sell time and thoughtful effort. This is increasingly important to me, especially for NYC founders and companies. I want to ~be the first money and more importantly the first partner to NY companies.

4. Illegible to undeniable/no more gradient stages: Nominal stages (seed, A, B etc.) don’t really matter. Instead, the only important distinction is between risk and scale. Everything else is nomenclature and posturing along a gradient. So the typical distinctions of a seed fund are basically meaningless. What you really need to do is commit to being a risk-on early stage fund finding good risk/supralinear bets (where if you’re right, you’re right) no matter the named stage. Use your money to flip illegible bets into undeniable stories.

5. Long apps/user: looking at really new technologies I want to bet on the apps, users, and beneficiaries. The foundational level is too capital intensive and the “middleware” doesn’t produce the biggest winners. But the “app layer” is where we’ll find our winners: those who can use new technologies to pioneer new products, business models, markets taking the risks we like to underwrite. This is most obvious in AI but we see the same sorts of opportunities in robotics, deep tech, and stable coins.

6. The complexity crisis: as software does more (more use cases), gets cheaper to build (more competitors), and gets upgraded/updated more frequently (more features), enterprise environments will get way more complex to manage. That will create tons of downstream challenges and big opportunities for consultants/SIs, integration platforms, vendor management, and identity infrastructure. The way software is purchased, deployed, and managed is going to completely turn over.

The existing themes and areas of interest (SMB platforms, real world businesses, buyouts, healthcare) aren’t OVER for me by any means. But some of them do feel like bets I’ve already made and categories I’ve covered vs more unexplored unexplored ground.

Nuclear Waste

Nuclear power is obviously going to expand globally and be a key pillar of the energy transition. That’s already underway and we can all see that. But what happens with the waste is at this point much less clear.

Angele Sahrouie, an analyst at Slow who covers deep tech and critical industries, put together a great deck on the challenges and opportunities on the horizon in the nuclear waste supply chain. Read more here.

I Read

• Let Chi Ossé Explain. Chi is very young but has an extremely bright and promising career as a New York Democrat. He is very often right on the issues and has been exceptional at making them communicable in the right way - oftentimes across otherwise challenging ideological lines. Really a gifted communicator with an eye on solutions, not just problems.

• Oregon is taking a hard stand against PE ownership in healthcare. This is overall really good news. It’s possible that private equity investors will find ways around this but it’s clearly part of an emerging regulatory consensus against consolidation. Consolidation in HC leads to worse outcomes, higher costs, and more physician burnout. Read our private practice thesis here.

• Trump's "beautiful" bill wrecks our energy future. The “BBB” is an absolute disaster. It doesn’t just roll back the IRA, it rolls back a decade plus of energy transition progress. It’ll make energy bills higher, our air quality worse, and our climate issues more dire. And don’t forget the massive cuts to healthcare funding!

• Why OnlyFans May Sell for 75% Less Than It’s Worth. It’s the porn, obviously.

I Wrote

• Six Themes for 2025 (and beyond)

• Why the 23andMe sale/lawsuit does (and doesn’t) matter

  • Short term: it doesn’t matter. The data isn’t valuable today. We know that because if it were valuable, then 23&Me would be valuable instead of going bankrupt.

  • Medium term: biological/genetic discrimination is probably going to be a real thing fairly soon.

  • Long term: your DNA will be important for identity verification and proof of life/humanity. Unfortunately it’ll probably be more like an immutable password than a public key.

  • Longest term: can't forget targeted bioweapons! A virus or chemical just for me bc I made the CCP mad

• Suicide round: noun (derogatory). Raising at the max price with min capital. The valuation sets a high bar without enough capital to clear it. Here’s the steelman case for raising this way. Best argument I can make.

Abundance

Now that the Abundance NY Voter Guide is out (remember, it’s not just a mayoral race) I’ve been thinking more about my own ballot.

In particular I want to highlight the Comptroller race between Mark Levine and Justin Brannan. Simply put, the current Manhattan Borough President Mark Levine is the guy for Comptroller. Whoever wins the race for Mayor, we’ll need a super competent, no bullshit comptroller to keep them in check and Mark Levine has a great track record on the issues that matter, namely housing. He’s a public school teacher who started a credit union, a great choice to audit public agencies and keep the government in check.

OTOH Justin Brannan is a NIMBY who exempted his own district from the City of Yes rezoning AND has been highly critical of congestion pricing. He’s unfit for citywide office on that basis alone.

One Cool Job

Long Journey Ventures is hiring a “gardener”. Basically an associate to help them track (and buy-up in) all the tiny checks, scout investments, etc. I’ve invested with them and like them. This is a job I would have LOVED 5 years ago. Definitely worth checking out

The Complexity Crisis

AI will drown the enterprise in complexity + the GIGACYLE is here

AI will drown the enterprise.

As software gets better, enterprise environments and IT systems are going to get way more complex and harder to manage.

• Software/AI can do more things ▶️ more use cases

• Cost to produce code goes down ▶️ more options/vendors for each use case

• Products get better faster ▶️ more features/releases to manage

Collectively, that means the future looks like way more apps and systems from more vendors swapped out more frequently with more and smaller vendors to evaluate each time you do. At the same time, MCP, operator models, etc. create entirely new surface areas and vectors for complexity, integrations, and management.

Everyone who gets paid to manage that complexity will be a winner because rising need and demand:

• Integration and abstraction platforms

• System integrators (SIs) and implementation consultants

• Vendor discovery and management

• New takes on SSO/security

I’m excited to back each of those types of opportunities. Let’s explore them individually.

Abstraction Platforms

Every company in every industry has a system of record that winds up as both an app and a database. It’s usually a specialized ERP, but sometimes it becomes so specialized it becomes its own thing all together: real estate (PMS), airlines (PSS), mortgage (LoS), healthcare (EMR), etc.

All the important information and workflow lives within this SoR, so any other apps need to integrate with or go through it. The more integrations and external apps/point solutions companies use, the more locked in they become. These things generate their own gravity.

These ERP-type tools are generally old, kind of shitty, and incredibly complex – often far too complex for any single startup to come in and replace, at least for enterprise customers. They've had years (decades) to build for every imaginable corner case and oddity their clients have presented. It’s feature-complete with high switching costs.

So, companies are completely locked into their SoR with no path out and rising costs (hard dollar and soft): delays/poor UX on the one hand, and IT consultants, integration specialists, specialized tools, high prices, etc. on the other.

But rather than building direct 1:1 replacements, startups should focus on vertical IPaaS or vertical Zapier to abstract away from the system of record and turn it into a dumb database vs a core app and workflow tool.

Sometimes this may even look as simple as an agent that sits between the SoR and the user such that users interact with one interface which then writes directly to the database, at least as a starting point.

System Integrators and Consultants

The new breed enterprise companies/vendors will often be too small to provide the services their customers need (training, complex onboarding, managed run, etc.). And the larger ones still want to offload much of that none-core work.

So even as the cost to build goes down, the cost to market, implement, integrate, etc. stays the same or goes up (more competition). You have to fit together more pieces, each with different interfaces, APIs, assumptions, and dependencies. Mapping alone becomes a beast.

That is ultimately a huge boon to the consultants, system integrators, and VARs who get paid to manage complexity and charge a toll fee for GTM to the vendors and/or implementation and management fees to the buyers.

The consultants and SIs will be beneficiaries on the demand side (already happening - Accenture and similar shops are huge winners today) but also as users/buyers themselves making use of new capabilities. These will always be services businesses, and some end clients will still need more service than code, but much of the work of an SI or implementation consultant can now be significantly accelerated / augmented (generating custom code, mapping data/fields, interviewing users, etc.).

There are going to be a bunch of big companies built around this - either selling software products to consultants/integrators or building/running them as margin-enhanced service providers.

Vendor discovery

If SIs and consultants will serve the top and middle of the market, how will the lower market deal with the same issues around discovery, onboarding, and integration?

Obviously PLG and prosumer works but only for a very narrow band of products and customers. For everyone else and anyone who needs a higher touch/consideration sale, there’s limited options for growth beyond spamming customers with outbound (which won’t work for long).

How do you find and compare options? How do you manage integrations and migrations? How do you find the small consultants to support you? Once you need more than documentation but less than a full blown lift and shift, what do you do?

These kinds of products will increasingly get bought and sold through marketplaces. A software marketplace can support the demand side with discovery, vetting, and compliance; and support the supply side with marketing and onboarding.

You can imagine a more transactional version of deep research that specs products against each other, matches your use cases, and provides some vendor visibility, interoperability, and management.

We'll move from listings businesses (G2 and Capterra never got huge anyway) to transactional marketplaces. Obviously the big cloud and ecosystem companies will be winners here (AWS, Microsoft, Appexchange, Shopify, etc.) but there will be other vertical and horizontal marketplaces that become end-to-end discovery and implementation hubs.

Identity, SSO, and Security

This one is pretty obvious, but as you’re provisioning access to way more things and managing way more apps, your SSO provider winds up in a unique position to do more than just basic access management.

With hundreds of agents and apps per company, your SSO provider becomes the map of who’s using what, when, and how. Whoever owns that map controls onboarding, offboarding, audit trails, risk scoring, and more. And they're in pole position to be:

1. the security layer for non-human access management (agents, MCP, etc)

2. an integration service between the apps themselves once credentialed

3. on/off boarding not just users but vendors (spend management)

At the same time as the need and market opportunity opens up, agents might also make it easier to compete with SSO incumbents. You can imagine standing up an Okta competitor more easily that uses operator models to just sign in for your without needing direct integrations.

Similarly, AI-first SSO could roll continuous, proactive monitoring. Instead of quarterly access reviews and manual logs, your IAM layer could continuously monitor policy drift, surface shadow accounts, and generate audit artifacts on demand.

The demand is gonna get there for way more and way more powerful SSO/access management at the same time as the technology creates a market opening.

So What?

AI will obviously make it easier to build software. But for it to actually be valuable to customers it needs to be fully usable: discoverable, installable, configurable, and manageable. As the cost to produce code and deploy agents goes down, the challenge of managing it all will go up. To the managers go the spoils.

These are the opportunities I want to invest against. If you’re building in, around, or on top of that mess, let’s talk.

---

The Gigacycle Is Underway

Venture Buyouts are an asset class now.

At Slow we’ve long believed that some founders/companies will be better served buying/owning/operating companies directly rather than selling to them. If you sufficiently transform the earning power, keep the upside.

We've been working on this strategy for years (starting with Metropolis and Teamshares) and now it's clear/finally a thing. Lots of people are doing it and literally everyone is talking about.

Despite the mania and noise, it's a real thing.

• Software has matured and there are no moats on code

• vSaaS is tapped out and "selling work" has a ceiling on ACV

• Capital markets favor control and value capture

• Founders want leverage and independence

Plus, it has undeniable sex appeal. Thus begins the Gigacyle. But there are still immutable laws of gravity and startups that apply here:

1. Let the main thing be the main thing: Product (not process) must still be your edge. Be pareto efficient

2. Start with software: Build first. Validate with customers before you scale (by buying).

3. Do big deals, not fast ones: The soft transaction costs (source, diligence, finance, integrate, etc.) don't scale linearly. Amortize them on big deals or treat lowering them as your core product. Avoid SMBs.

4. Use debt: duh.

As Will put it, you have to limit yourself to one miracle at a time.

Nearly all of the "roll ups" (high velocity, small deals) and venture-funded PE execution plays (we're smarter/better operators) will top out because transactions themselves are hard. Respect the laws of gravity.

Focus on value creation first. Go big.

Read more in our GBO deck.

Baseball!

Finn and I have a box at the Brooklyn Cyclones for founders and operators on Thursday June 12. We'll take in some top notch (minor league) ball with dogs, beers, and maybe a roller coaster ride. Everyone gets a free hat and an elite NY experience.

Abundance Voter Guide

Abundance NY has released its 2025 voter guide.

New York is in crisis. Our rents are rising, our infrastructure is crumbling, and our streets and subways feel less safe.

On June 24th, New Yorkers will vote in our primary elections to choose party nominees for offices from City Council to Mayor. The officials we elect have the power to shape how safe, affordable, and livable New York is. But voter turnout is low, and high-quality information is hard to come by.

This Voter Guide highlights a set of competitive, high-stakes races. It is meant to empower readers to make informed decisions and recommends the candidates most likely to move the city forward.

While this voting guide doesn’t PERFECTLY line up with my own (I’m not ranking Whitney Tilson) it’s generally a great resource.