Uncapped with Jack Altman — Episode #22: Greg Rosen from BoxGroup

Date: August 20, 2025
Host: Alt Capital
Guest: Greg Rosen, BoxGroup

Overview

In this episode of Uncapped, Jack Altman hosts Greg Rosen from BoxGroup for a candid and wide-ranging discussion on venture capital’s collaborative models, early-stage investing philosophies, how to avoid the biggest traps in VC, and navigating the evolving power dynamic between founders and investors. Greg shares BoxGroup’s unique approach that emphasizes breadth over ownership, the importance of seeing more deals, fostering a culture of humility, and the detailed mechanics—right down to the rituals and time audits—that differentiate their style from much of traditional, consensus-driven VC.

Key Discussion Points & Insights

The Collaborative Venture Model & Scaling BoxGroup

• Collaborative Model as Anti-Scale:

  • BoxGroup intentionally avoids leading rounds to maintain their "Switzerland" status—staying neutral and collaborative with other funds (01:27).
  • "Collaborative venture is hard and almost by definition it has to be anti scale in certain dimensions." — Greg Rosen (01:27)
  • Leading rounds can poison the well; as soon as a VC leads, other investors suspect not all opportunities are shared (02:15).
  • BoxGroup's differentiation: never lead, maintain flexibility, and always be value-additive without competitive tension.

• Check Size and Flexible Participation:

  • BoxGroup's check sizes have grown with inflation, but stay minority: $750K–$1M (03:11).
  • The goal is to make it seamless for founders to include BoxGroup, regardless of stage or round structure.

• Scaling Collaboration vs. Ownership:

  • Instead of scaling by writing bigger checks, BoxGroup scales by making more investments—70–80 per year—while still doing rigorous filtering.
  • "If we look at all the good companies that get started in a year, we think about how many of those did we see and then how many of those did we actually get right versus wrong." — Greg Rosen (24:09)

Adverse Selection, Deal Sourcing, and Pick-Win Dynamics

• Adverse Selection as the Central Problem:

  • More than just seeing more deals, the key is not getting adverse selection: not being left out of the best companies (06:08).
  • "Every strength has a corresponding weakness and vice versa. So the knock that I hear from other people knocking on our model is you need to own a ton of the companies that work in order to make venture work for us. What we would say is that is true, but also you gotta be in the right companies and it's much harder to be in the right companies." — Greg Rosen (06:12)

• The Power of the Basket vs. Index Critique:

  • BoxGroup takes "more shots on goal," aiming to be in enough $30B, not just $3B, companies; in long-run returns, this outweighs lesser ownership (06:55).

• It’s About 'Seeing' Not Picking:

  • In an increasingly competitive landscape, proprietary deals barely exist anymore (08:58).
  • “If you look at the math formula of venture, the...seeing picking and winning a deal and then equals enterprise value output...if you don't change any of those variables, what are you accepting? You're accepting that returns are going to go down.” (09:18)

• Approach to Evaluation:

  • BoxGroup processes 5,000–6,000 referred/inbound deals a year to make roughly 70–80 investments—“shots on goal” (10:55).
  • Humility is core: “We don’t know how to have a basket of 20 companies and pick the most important ones, given how competitive and hard it is.” (10:55)

Practicalities of Outbound Sourcing & “Calendar Audit”

• Active Outbound Over Passive Inbound:

  • “Inbound is like the icing on the cake that tastes really great, it's awesome. But you can't just eat icing. Right. You need the actual body of the cake and that is outbound.” — Greg Rosen (12:08)
  • Outbound includes curated events (interns, campuses, engineering leaders), high-quality networking, and value-sharing with trusted investors (12:30).
  • They’ve even built software tools for surfacing stealth founders—“if there’s a deck, it’s too late.” (13:55)

• The Grind at Earliest Stages:

  • A heavy focus on meeting people before they’ve left their current jobs, knowing most such “bad meetings” don’t go anywhere, but that’s the price of getting the first look at future founders (15:04).
  • “There is no substitute for your time. And most people, especially check writers, do not spend time with someone who hasn't left their job yet, who may never start a company.” (15:04)

Internal Process: No Partner Meetings, Single Trigger Model

• Rejecting Consensus:

  • BoxGroup has no partner meetings; consensus slows things down and raises the risk profile in the wrong direction (18:08).
  • “Consensus driven decision making at seed doesn't work, in my opinion. Back to the point on… It will just shift the level of risk that you're taking.” (25:51)

• The Middle, Not the Obvious, is Where Alpha Lives:

  • Most top winners start in the “interesting, but not obviously great” bucket: “All the alphas in the middle bucket… and more often than not, we are discussing those deals as just like we're talking about the human being for 15 minutes, and a little bit maybe two minutes on what exactly they're working on.” (19:55)

• Process Designed to Get to “Yes”:

  • Culture pushes toward supporting the champion for a founder within the firm to “squint” and advocate; the team actively helps others get to a yes, rather than poking for nos (22:17).
  • “We have to help the sort of person that's dreaming with the founder. Or squinting is sort of the term we use internally to sort of get to their own version of yes.” (21:19)

• Single Trigger:

  • One partner’s conviction can push a deal forward, eliminating groupthink or endless debate (25:49).

Picking: Taste, Intuition, and Learning

• Is Great Picking Describable?

  • It's half taste (innate, "who do I want to spend time around?") and half learning through immense repetition (28:26).
  • “You can’t teach taste... But you can get reps.” — Greg Rosen (28:26)
  • There’s value in observing your own behavior: if you take a third meeting with a founder, you probably should invest (29:33).

• People, Not Ideas:

  • For BoxGroup, founder evaluation trumps idea evaluation, especially in the earliest stages. Ideas aren't precious; the right founder will evolve to a great market (30:37).

Time Audits & Most Common VC Time Wasters

• “Calendar Audit” Philosophy:

  • VCs are particularly prone to wasting time because the job is amorphous (31:46).
  • Unproductive time includes excessive networking with other VCs, attending events without founders, and hiding behind “necessary” board work or investor meetings (33:13).
  • “You are only as good as your last deal... Why are you spending all this time networking with other investors? Why don't you go meet founders?” (32:16–33:13)

• Choose Your Archetype, Stick to It:

  • Everyone should focus on their competitive advantage—being a hyper-networker, a philosopher/media presence, or a true specialist (34:25).
  • “You are better off just focusing on the thing you are outlier in... and just focus all your energy and effort on being amazing there.” (34:48)

• Audit & Cut Out Empty Calories:

  • Say “no” to requests that don’t align, sift high-quality inbounds versus prioritizing outbound events and proactive relationship-building (36:56).

Notable Quotes & Insights

• On Early Stage Alpha:
  “All the alphas in the middle bucket ... it is not obvious. It is like we're talking about it in an obvious way, but these are decisions that are so on the margin.” — Greg Rosen (20:01)

• On Collaborative Models:
  “If we’re collaborative...and we only lead two or three seeds of that, now, when I'm sharing a deal...you're going to be thinking...is it the one or two deals that they're leading and actually most excited about?” (02:14)

• On Venture Humility:
  “This is our Venture Humility score, which is we don't know how to have a basket of 20 companies in a fund or in a year and pick the most important ones, given how competitive and hard it is.” (10:47)

• On Founders vs. Investors:
  “We pitch zero. Like, founders are the reason why companies work, not investors.” (40:38)

• On Picking Taste:
  “Certain people just like have inherent deal taste. And what is that? It's like the human taste.” (28:26)

Power Shifts and Relationships in Venture

• The Broadening and Narrowing of VC 'Cliques':
  • A decade ago, founder power was low and rounds were orchestrated by VC alliances. The rise of founder power has decentralized the landscape—but even now, new trusted networks are forming, though it's now the founder (not the VC) who selects their partners (37:46–39:37).
  • “I always thought it was like bad to play favorites. And now it's like, I'm absolutely playing favorites on the recommendation. And then ultimately it's on the founder to go and take that recommendation. And if we have high trust, like, hopefully it means a lot.” (38:31)

Future of AI, Code, and Neuralink

• AI & Coding Platforms:

  • Greg draws a parallel to the early mobile era, where it took years for truly native apps to emerge. In AI code gen, we're still at the “Pandora on mobile” stage—native forms are coming, but not fully realized (41:43–44:01).
  • “What Warp is doing with the ADE and the agentic development environment is...maybe the right native expression is not opening up a code file and hitting tab 100 times...You should just be telling the computer or machine what you want it to do and then auditing all the thought and steps throughout it.” (43:11)

• Brain-Computer Interfaces:

  • Greg speculates that direct brain I/O may become a more efficient way to disrupt productivity than building Westworld-like robotics, but admits this is speculative/fanciful (44:50).
  • “Maybe what's actually more fun and interesting is just if we find a way to have input output to our human brain...the path of least resistance rather than that utopia on Earth might just be we're all, like, hooked up to the Matrix.” (44:50)

Timestamps for Important Segments

• [01:27] — Collaborative Venture Model Explained
• [03:11] — Why BoxGroup Anxiously Avoids Leading Rounds
• [06:08] — The Real Adverse Selection Problem
• [09:18] — Why 'See' Trumps 'Pick'
• [12:08] — Outbound Sourcing and the 'Cake' Analogy
• [15:04] — The Necessity of 'Bad' Meetings and Time Investment
• [18:08] — Why BoxGroup Does Not Use Partner Meetings
• [21:19] — Creating a Team Culture that Gets to 'Yes'
• [25:51] — Structural Necessity of the Single Trigger Model
• [28:26] — Is ‘Picking Taste’ Teachable?
• [29:33] — The 'Third Meeting' Rule
• [31:46] — The Calendar Audit & Time Waste in VC
• [34:25] — The Three VC Archetypes: Hyper-Networker, Philosopher, Specialist
• [37:46] — The Changing Nature of Relationships in Venture
• [41:43] — The Future of Code, AI Agents, and True Platform Shifts
• [44:50] — Brain-Computer Interfaces, the Next Utopia?

Memorable Moments

• On deal intuition:
  "If you looked back anytime you took a third meeting with a founder, you should have done it." — Greg Rosen (29:33)
• On humility in VC:
  "We don't earn the right to pick or win unless we see it." (25:26)
• On founder-centricity:
  “We pitch zero [value add]. Like, founders are the reason why companies work, not investors.” (40:38)
• On professional FOMO:
  “Ideas are not precious. You can work on anything. And so now we've sort of iterated away from that and said like, okay, that's actually a sign that we're doing our job and getting to there early.” (16:00)
• On time management:
  “Say no all the time. Yeah, like to coming back on this podcast.” (37:20)

This episode is packed with practical yet contrarian wisdom about building a modern, collaborative, high-velocity venture practice. Greg and Jack reveal the hard truths of early-stage investing: undervalued grind work, how culture eats process, and that the path to seeing greatness isn't easier, just earlier. A must-listen—or read—for anyone who wants to understand the mechanics of modern seed investing, or anyone considering how to build enduring institutional advantage in VC.