Latest / The Learning Corner by Precursor / Episode #44: Founding Engineers Are Impossible To Hire, Regulators Come For AI Therapy Chatbots, What VCs Get Wrong About Hiring

Transcript

Episode #44: Founding Engineers Are Impossible To Hire, Regulators Come For AI Therapy Chatbots, What VCs Get Wrong About Hiring · The Learning Corner by Precursor

• 0:00
  The close cousin of responsibility is liability.
• 0:02
  And most companies do not want the liability for what LLMs spit out because it's
• 0:08
  probabilistic and not deterministic.
• 0:10
  And they don't actually know what's gonna come out when you put in a chat.
• 0:13
  But at the same time, I don't know that the companies have gone as far as they could have
• 0:17
  on safeguards.
• 0:19
  It's the week of August 18.
• 0:22
  Welcome to the Learning Corner, a Precursor Ventures podcast.
• 0:29
  Welcome to the Learning Corner by Precursor.
• 0:31
  I'm Mia Farnam, and I'm joined by RGP and managing partner Charles Hudson, where we will
• 0:36
  be talking through the articles and Goodreads that caught our teen attention over the past
• 0:41
  week.
• 0:42
  Our first topic is a piece from the San Francisco Standard titled, The Hot Girl at the
• 0:46
  Club, How Founding Engineers Became Silicon Valley's Biggest Catch, with top engineering
• 0:52
  talent being snapped up by capital rich foundational model companies or big tech.
• 0:57
  And with the role of a founding engineer at an AI company evolving, this search for the right
• 1:03
  candidate has become nearly impossible.
• 1:06
  The article highlights how in the pre AI era, a founding engineer's responsibilities centered
• 1:11
  more on coding and technical skills.
• 1:15
  Now with AI tools offloading much of the coding work or being perceived to do so, founders
• 1:21
  expect far more from their founding engineers.
• 1:23
  Enticing equity stakes and generous compensation packages are being offered to
• 1:28
  attract top talent, and having access to strong candidates has become almost like a valuable
• 1:33
  currency exchange between VCs and startups alike.
• 1:37
  Add in the fact that VCs are preempting technical talent earlier and earlier to become
• 1:43
  founders off the bat with way more equity than a traditional founding engineer typically had
• 1:48
  in the pre AI era, and we've really got ourselves in nonstop scramble for talent across
• 1:54
  the ecosystem.
• 1:55
  So along with so many other dynamics across venture and start up today, this feels like
• 2:00
  such a whirlwind shift from the pre AI days.
• 2:03
  Wanted to start with your thoughts first on everything mentioned in the article.
• 2:07
  I read the article and I was like, wow.
• 2:09
  People are really going to great lengths to find talent.
• 2:12
  And I think there's a lot of interesting things happening all at once.
• 2:16
  One is the supply of people who actually have real world experience.
• 2:20
  At the frontier of AI is limited and we don't make a lot of new ones every day, so the
• 2:25
  competition's fierce.
• 2:26
  Second, a lot of these companies have tons of money.
• 2:29
  And in a world where all of your competitors have money, unless you're Meta who can make,
• 2:33
  you know, nine digit offers to candidates, for a difficult startup, you are a well funded,
• 2:39
  well resourced startup competing against other well financed, well resourced startups.
• 2:43
  So money probably isn't the currency.
• 2:45
  So you have to get more creative in terms of how you woo these people.
• 2:49
  And the third thing is you look at the Windsurf character AI world and the value prop for being
• 2:58
  an early employee versus a founder in recent memory hasn't been so great necessarily.
• 3:04
  And so it's threading this needle of finding people who are good enough to work for you and
• 3:08
  who are talented and who don't wanna be founders and have the skill set, but don't
• 3:12
  wanna be founders and start companies themselves and be in charge.
• 3:14
  And, you know, when I read this article, my instant reaction was, to the extent we are in
• 3:20
  an AI bubble, and I believe we are, but, like, there are always tech bubbles, this is one of
• 3:25
  those moments we'll look back on and say, maybe that was a sign that things had maybe gotten a
• 3:30
  little out of hand.
• 3:31
  Which isn't to say that founding engineers aren't incredibly valuable.
• 3:34
  I've seen them have huge impact in our portfolio companies, particularly ones where
• 3:39
  there isn't a strong CTO in place or there is a strong CTO in place, but they're horsepower
• 3:44
  capped.
• 3:44
  So I don't know.
• 3:45
  I I hope that at some point that we get off the boil on this stuff.
• 3:48
  I think it's getting it's getting a little crazy.
• 3:51
  Question for you because a lot of the conversation in the article centered around
• 3:55
  attractive equity packages that they're extending to founding engineers.
• 3:59
  Most of the time for founding engineers, 'd see maybe attractive cash comp bonus structures,
• 4:04
  but their equity is a lot smaller than a founder.
• 4:06
  Now we're seeing really large amounts of equity being extended to these founding engineers.
• 4:11
  Can you walk through some of the potential consequences of something like that?
• 4:15
  Yeah.
• 4:15
  I talked to somebody who gave a large founding engineering grant to someone.
• 4:19
  I said, how did you get to that number?
• 4:21
  And he said, well, we're gonna hire fewer engineers.
• 4:23
  So he's like, before I would have had this equity pool that I would have had to spread
• 4:27
  across a larger number of people.
• 4:29
  Since we're hiring fewer people, I can give this person more.
• 4:32
  And I was like, that actually is consistent.
• 4:35
  That tracks.
• 4:36
  I think the tricky thing is you gotta make sure that you've got investing in place and that the
• 4:41
  folks are gonna stay.
• 4:42
  Otherwise, you end up with someone who comes to your company for, we'll say, twelve months in a
• 4:47
  day and leaves with a fairly large equity grant.
• 4:51
  And now you've got this hole on your cap table where you've given equity to someone who has
• 4:55
  certainly provided service to the company but is no longer there.
• 4:58
  And I've seen so many cases, founding engineer or not, where founders have looked back and
• 5:02
  said, wow, if I'd known we were only gonna have access to that person for a year, I would have
• 5:07
  made a different offer.
• 5:08
  So I think we will see the downstream effects of this in a few years probably.
• 5:12
  But the flip side is giving a founding engineer a huge equity grant can also be a huge
• 5:16
  motivator and signal to that individual about how much you value their contributions.
• 5:20
  So I always felt like the whole point of the founding engineer thing was you're not a
• 5:24
  technical cofounder, so I don't need to give you 25 to 50% of the company.
• 5:28
  But I wanna also find a way to give you more than 50 basis points of equity.
• 5:33
  So I'm gonna create this new title called founding engineer, which sounds like a founder,
• 5:36
  but it isn't a founder.
• 5:37
  And I so can rationalize giving you more equity than I'd give you as a regular employee, but
• 5:41
  not as much as a cofounder.
• 5:42
  I was talking with a friend who recently made their first founding engineer hire, and she was
• 5:47
  mentioning how she's nervous all the time about a potential big tech company, a larger startup
• 5:53
  with more cash coming in and and swooping in and taking them.
• 5:56
  And her question was, at what point will I be able to stop worrying about that?
• 5:59
  I actually wanted to ask you that because I could go either way with where I'm landing.
• 6:03
  I had the almost opposite conversation with somebody this weekend who said, Meta and
• 6:08
  OpenEye have not tried to poach our engineers.
• 6:10
  Does that mean they're not good?
• 6:11
  And I was like, don't think that about your own team if you think your team is good and they're
• 6:15
  producing.
• 6:16
  But I think it's created this other worry of people which is like, if I have good people,
• 6:19
  how do I keep them?
• 6:20
  And if I think I have good people and no one's trying to poach them, do I in fact have good
• 6:24
  people?
• 6:24
  And so I don't know.
• 6:25
  I've always felt like as a manager, some of these things are outside of your control.
• 6:29
  All you can really control is like, do you create an environment where the people you want
• 6:32
  to stay wanna be there?
• 6:34
  Do you try to listen to what they need and meet whatever needs you can?
• 6:38
  And ultimately, like, I've had people who work for me who needed to make more money and I'm
• 6:41
  like, I I cannot pay that.
• 6:42
  Yep.
• 6:43
  Or needed to move across the country to be with a partner or had other things happen in life.
• 6:47
  Whereas I feel like I've done everything I can to make this the right place to be for you, but
• 6:51
  it isn't based on what you want now.
• 6:53
  Yeah.
• 6:54
  It's funny.
• 6:55
  On the one hand, you have all this conversation about RTO and manager strike back and employers
• 7:00
  flexing their muscles on everything other than founding engineers and AI researchers.
• 7:04
  Yep.
• 7:05
  They're still in the, like, COVID 2020 work from home halo employees have power era.
• 7:11
  To your earlier point about founders questioning if their people are good, if none
• 7:15
  of the bigger companies are willing to poach them, it's almost this weird conundrum of you
• 7:18
  want your talent to be wanted by other people, but not enough that they actually leave.
• 7:22
  And I think if you get stuck in that headspace forever, you probably wind up making some
• 7:26
  really, really odd decisions internally if you're just Mhmm.
• 7:29
  Constantly looking around and wanting external validation, which is, I guess, like such a
• 7:33
  large portion of venture and being the founder in general is sitting with those thoughts.
• 7:38
  Last question.
• 7:38
  Do you think that these CEOs are in the right for asking for different qualities and skill
• 7:44
  sets for their founding engineers?
• 7:45
  So I'm talking about them wanting most of these founding engineers to be higher level thinkers,
• 7:49
  communicators, product architects, and this was something that didn't really exist in the pre
• 7:53
  AI era?
• 7:54
  Are they kind of over clubbing a little bit?
• 7:56
  I think we don't really know.
• 7:58
  And I think this is an incredibly difficult time to be the CEO of a tech company,
• 8:02
  particularly a startup.
• 8:03
  Because on the one hand, you're like, well, do I hire technically brilliant people to be my
• 8:06
  founding engineers when I know that Cursor, Replit, Cloud Code, Codex is gonna produce a
• 8:12
  significant amount of the output?
• 8:13
  So do I want someone who can work with slash supervise a coding agent?
• 8:17
  Or do I want someone who can envision and architect a product and maybe help build quasi
• 8:22
  vibe code the first version.
• 8:24
  But it's a person who knows what tools to grab off the shelf and is a fast builder because
• 8:28
  speed is really at a premium, maybe even over quality right now.
• 8:32
  And I don't think anybody knows, and I'm seeing people ask the same questions about product
• 8:37
  managers.
• 8:37
  Do I wanna classically trained product manager at my company, or do I want someone who's,
• 8:41
  like, really good with, like, lovable Replit and these other tools who can, like, conceive
• 8:45
  of the product, design it, and bring a quasi functional, even a functional version to
• 8:50
  engineering and talk to them about how we build something big.
• 8:53
  So I think everyone's rethinking these roles, and the problem is when you're the founder of a
• 8:57
  CEO of a company, you can't go off on the mountain and say, I'm gonna come back in six
• 9:00
  months with a plan.
• 9:01
  You have to live in the now.
• 9:02
  Yep.
• 9:03
  And I think a lot of people are trying to figure out, in the now, what do I need?
• 9:06
  And in the now, it feels more like people who have facility with the tooling and know how to
• 9:11
  coordinate this stuff, so the product archetype makes sense to me, but I hope people don't
• 9:16
  overcorrect because you still do need the technical ability to know what good looks like.
• 9:21
  Do you need the ability to make architecture decisions?
• 9:23
  It's not that simple.
• 9:24
  And the problem with living in the I don't know quite yet what I'm looking for added into the
• 9:29
  AI talent war is you will not know, then you're going to be pushing insane compensation
• 9:34
  packages and equity packages without framing up exactly what you're looking for.
• 9:38
  And so you're buying something and possibly maybe overpaying or not even overpaying, but
• 9:42
  you just have no idea what you've just purchased coming into your company because you
• 9:46
  haven't really set the plans and it's not your fault because the ecosystem is constantly
• 9:50
  shifting.
• 9:50
  I think there's gonna be some people a year from now who are gonna look back on the
• 9:54
  decisions that they made on role design and team architecture, and they're gonna say, we
• 9:59
  got it wrong.
• 10:00
  And I think everybody's trying to make the best decisions they can with a high degree of
• 10:04
  uncertainty.
• 10:05
  Well, our next piece comes from Platformer News titled The States Are Coming for AI Therapy
• 10:10
  Bots.
• 10:11
  The article explores ongoing political debates and efforts to rein in AI tools that market
• 10:16
  themselves as offering therapy, quote unquote.
• 10:19
  Notably, this has become a bipartisan issue with lawmakers on both sides actually working
• 10:23
  to restrict the claims these tools can make and the influence they can exert on consumers from
• 10:28
  a therapeutic standpoint.
• 10:29
  The urgency really has grown in light of some troubling incidents such as Meta's chatbots
• 10:34
  falsely claiming to be licensed therapists and young adults turning to bots when considering
• 10:39
  suicidal thoughts without any human intervention in the loop.
• 10:42
  These events really have intensified calls for accountability.
• 10:45
  At the same time, while there are significant concerns about chatbots substituting for
• 10:50
  clinical care, the technology has created more accessible avenues to receive support and
• 10:55
  introduce a new, possibly more approachable format for delivering therapy in quotation
• 11:01
  marks.
• 11:01
  And we're gonna dive into that problematic wording as well.
• 11:04
  Ultimately, the piece concludes with the author Casey Newton arguing that no matter how quote
• 11:08
  unquote therapy like a chatbot may sound, it should never be granted the accreditation of a
• 11:13
  licensed therapist.
• 11:14
  This raises the question, where should responsibility lie, and how should it be
• 11:19
  distributed among the different players in the space?
• 11:22
  So I wanted to ask you kind of where you're thinking these days.
• 11:25
  I'm gonna get this wrong, but there's an old, I see it every now and then, IBM thing that says
• 11:30
  computers should never make decisions because computers can never be held responsible.
• 11:34
  And it's interesting.
• 11:35
  I I when I read the article, it was like, you know, for a long time, a lot of people said,
• 11:39
  hey.
• 11:39
  We shouldn't have state by state regulation because it becomes impossible for startups to
• 11:45
  operate.
• 11:46
  I'd say, well, we already have that in health care.
• 11:47
  So I would say that it makes it impossible for startups to operate is hyperbole.
• 11:52
  All the regulated industries have usually federal and state regulation.
• 11:55
  So it is possible.
• 11:56
  I will not say it is friction free, but it's possible.
• 11:58
  And I think the alternative is if you don't want state by state licensing, you have to at
• 12:02
  least take the posture that you're being responsible.
• 12:05
  And I think this is an area where I think when it's all said and done, you know, there's
• 12:09
  always the we talked about this in our team meeting to prevent some of these things.
• 12:13
  And in every interview I've seen lately with Sam Altman, he's mentioned that a lot of people
• 12:17
  are talking to ChatGPT as if it were a friend or a therapist.
• 12:21
  I know that there are lots of people who are using tools like Character AI to have, like,
• 12:26
  extended conversations and explore topics.
• 12:28
  So I think in the best case framing of this, we have a shortage of clinicians who are skilled
• 12:34
  and able to meet the demand for mental health services.
• 12:37
  So the best case scenario is that the lighter weight versions of mental health problems, not
• 12:43
  the things that require either medication or, like, serious clinical care, that AI becomes a
• 12:49
  safe way to offload some of that burden from the system so that the limited supply of
• 12:54
  clinicians we have can attend to the most severe people.
• 12:58
  The worst case scenario I can envision is that people just start to treat these things as if
• 13:02
  they're licensed therapists and listen to what they say.
• 13:05
  And I think the thing that always makes me nervous is LLMs are pretty good at telling you
• 13:09
  what you wanna hear.
• 13:10
  They can be a little pleasey at times, and that's not always the right angle or approach
• 13:15
  when you're trying to work through a difficult issue.
• 13:17
  Sometimes a good therapist or a good advisor will tell you something you don't wanna hear
• 13:22
  Yeah.
• 13:22
  Or will hold a mirror up and help you see something that's, like, uncomfortable to see.
• 13:26
  And it's not clear to me that this version of LLMs and AI powered therapy products and look,
• 13:31
  I know there's some some newly funded or heavily funded companies that have promoted a
• 13:36
  more responsible approach, or they're gonna build LLMs that are specific for this.
• 13:39
  You you and I and the team have met a ton of companies that are using AI for therapy, and
• 13:44
  all of my questions when I meet these founders are not technical.
• 13:47
  It's like, where is your moral compass?
• 13:48
  Where do you think these products should and there's this tension.
• 13:51
  The people I've met with the strongest moral compass are worried that somebody with a lower
• 13:55
  ethical bar will grow faster and will then earn the right to sort of set the terms of what the
• 14:00
  ethical bar or standard for these products should be.
• 14:03
  And so I don't know.
• 14:04
  It's very difficult.
• 14:05
  And I think state by state regulation was inevitable as these stories about individual
• 14:11
  people getting themselves into trouble came out.
• 14:14
  And we've seen this across health care and digital health across time, I feel like.
• 14:18
  When Headspace came out for the first time, they went through their severe ups and downs on
• 14:21
  privacy, security, etcetera.
• 14:23
  Yep.
• 14:23
  I think sometimes I have to remind myself I live in a venture bubble where I use AI more
• 14:28
  than the normal person.
• 14:29
  And so I think you said this two days ago and it stuck with me of I used to use AI or or
• 14:34
  Tratchability constantly just because that's just the nature of my job.
• 14:38
  It's it's helpful for me getting familiar with the tools.
• 14:40
  But because I've used it so much, I've learned how it engages with me how and it doesn't push
• 14:45
  back on me.
• 14:45
  So because of that, I've started to pull back.
• 14:48
  And I know where my boundaries are of where I can extract value.
• 14:52
  The normal average American does not use AI that much.
• 14:57
  And so I don't think at this current rate can make that distinction of when a chatbot is
• 15:02
  adequately creating guardrails around what's actually something that they can deliver from a
• 15:06
  therapeutic standpoint, which is where I I have so many quitting concerns here.
• 15:10
  Also, like, the moral compass standpoint, it did make me also think about who the players
• 15:14
  are in this argument of who has responsibility.
• 15:17
  And I was gonna ask you if I'm missing anyone on this list or your perspectives on in
• 15:22
  general.
• 15:22
  But I guess in my mind, I was thinking it's gonna be the founders of these foundational LLM
• 15:27
  companies.
• 15:27
  Yep.
• 15:28
  The consumers themselves.
• 15:30
  Yep.
• 15:30
  The parents, I assume of underage Yep.
• 15:32
  Kids who are gonna be using these similar to social media, policymakers, investors.
• 15:38
  The list really goes on.
• 15:39
  Like, is there anyone I'm missing there?
• 15:41
  Or what are your thoughts there?
• 15:42
  Shout out to James Cham who, like, turned me onto this book, The Unaccountability Machine,
• 15:46
  the TLDR.
• 15:47
  It's a big book on systems design, but it talks a lot about how systems can be designed such
• 15:52
  that really nobody has authority or responsibility or accountability.
• 15:56
  Their classic example is something happens when you're at the airline, you go talk to the gate
• 16:00
  agent, the gate agent's like, I can't do that.
• 16:02
  You go to customer service, customer service is like, I can't do that.
• 16:05
  And ultimately, you realize the only person who can fix my problem is the CEO of this airline
• 16:09
  because nobody else has the authority or responsibility to fix it.
• 16:12
  And I worry what's gonna happen is the parents are gonna point to the LLM providers and say,
• 16:16
  you should make it so that your tool doesn't do this.
• 16:19
  And the tool providers are gonna say, we can't really anticipate every use case that a
• 16:24
  person's going to use and it's not our responsibility.
• 16:27
  The policy makers are gonna say, we're gonna go to these companies and threaten them with
• 16:31
  policy or else, but policy is a slow tool.
• 16:34
  Yep.
• 16:34
  And kids are gonna be kids.
• 16:36
  And I think we've seen this movie before.
• 16:38
  It was called Social Media.
• 16:39
  And I'm not saying that we will end up in the same place, but I think that was a world in
• 16:43
  which the same set of players were largely around the table.
• 16:48
  There was great enthusiasm and some inkling about how things could go wrong.
• 16:53
  And in the end, like, nobody was really responsible.
• 16:56
  I think the stakes here are different.
• 16:57
  I won't say higher or lower.
• 16:58
  They're just different.
• 16:59
  And I feel like we're still living with the repercussions of social media today.
• 17:03
  Like, as parents, I can't speak to it, but you're a parent, so maybe you can.
• 17:06
  And you and I have talked a lot about different companies that are aspiring to do wonderful
• 17:09
  things to create guardrails.
• 17:10
  And you and I had a good conversation around, yes, there's wonderful tools that you can use
• 17:13
  to help create guardrails and stopgates of being able to know when too much is too much
• 17:18
  time interacting with social Yeah.
• 17:21
  But at one point, probably, you should be able to kinda create that barrier as a parent Yeah.
• 17:25
  And be able to kind of jump in and have, like, open conversations about that.
• 17:28
  The problem, I think, the differences here between social media and ChatGPT is no one
• 17:32
  really knows what the outputs of these chatbots are, and they change every three months.
• 17:39
  And they're they're building context on you all day.
• 17:41
  And so it's so difficult to be able to keep up and then create a safe environment as a parent,
• 17:47
  I can imagine.
• 17:48
  So I don't really know how you can effectively step in right now, especially because also
• 17:53
  you're hearing it from two sides of people's mouth.
• 17:55
  You want your child to stay safe with AI and chatbots, but you also need them to be engaging
• 18:00
  with them in order to be very dangerous in the work environment in ten to fifteen years.
• 18:04
  So, okay, you want your kid to learn how to prompt and contextualize within ChatGPT, but
• 18:08
  not enough that they become overly involved.
• 18:11
  I don't know where that that stop gate is yet.
• 18:13
  No one does yet.
• 18:14
  Well, our last topic of the week is a piece by Mike Dauber from Amplify Partners titled, Why
• 18:19
  Most VC Firms Hire Associates Wrong and What We Do Instead.
• 18:23
  The article highlights the reality that most venture funds have little idea how to
• 18:28
  effectively hire their junior team members, associates.
• 18:31
  Mike argues that much of this flawed decision making stems from a fundamental
• 18:35
  misunderstanding of what the associate role should be.
• 18:37
  Mike outlines two schools of thought on how firms hire associates.
• 18:41
  The first is the leverage model, which is far more common, I've found.
• 18:45
  In this model, firms hire smart people from very predictable backgrounds.
• 18:49
  So banking, private equity, consulting, etcetera, keep them for two years and provide
• 18:53
  really no upward mobility due to the usual pyramid structure of the firm.
• 18:58
  The second is the development model, where firms bring in people early and invest in
• 19:03
  training them over time to become partners.
• 19:05
  This approach requires significant time and energy from current partners, which makes them
• 19:10
  far more selective about who they hire given the high opportunity cost.
• 19:14
  Naturally, it's clear that Mike and Amplify is in support of this second model.
• 19:18
  He closes the piece with the quote, if we keep treating the associate role as a transient
• 19:22
  chief of staff instead of a meaningful apprenticeship, we are going to end up with an
• 19:27
  industry full of people who learn the wrong lessons about what venture capital should be.
• 19:31
  Wanted your reflections on the piece as someone who has done literally all the hiring here at
• 19:36
  Precursor.
• 19:37
  It's funny.
• 19:38
  This is one of those articles I read it.
• 19:39
  I had an immediate reaction.
• 19:40
  I was like, yes.
• 19:41
  We pretty much, at least I'd like to think, do the development model at Precursor.
• 19:45
  But then I was like, you know, the USV analyst associate program has produced some amazingly
• 19:51
  talented entrepreneurs, investors, and contributors to the venture ecosystem.
• 19:56
  And so I think some of this comes down to, like, what kind of firm do you wanna have?
• 20:02
  I could point to other examples beyond I could look at Anne Muraco's two year chief of staff
• 20:07
  program has produced some really talented, excellent people.
• 20:10
  I think, I'm not trying to be too cynical.
• 20:12
  There's a set of people who get into venture capital because it involves managing far fewer
• 20:16
  people than an operating role would entail.
• 20:19
  And for many of those people, the optimal number of people they'd like to manage is zero.
• 20:23
  And so I think under the leverage model, it's not that you don't learn.
• 20:27
  It's that you learn different.
• 20:28
  I think under the leverage model, the people who are successful, it's like a law firm or an
• 20:32
  investment bank.
• 20:33
  The people who are successful are the people who learn how to make the partner more
• 20:36
  effective.
• 20:36
  It's not so much that you learn the craft, you learn how to be a good helper.
• 20:40
  Right.
• 20:40
  The irony is in most professional services businesses, that's how you utilize junior
• 20:44
  people.
• 20:44
  Yep.
• 20:44
  That's what junior attorneys do, that's what junior bankers do, that's what junior
• 20:47
  consultants do.
• 20:47
  Yep.
• 20:48
  Your job is to amplify and provide leverage for the senior person, and then when you become
• 20:52
  senior, you will have done all of the low level tasks, and you will have task mastery so you
• 20:58
  can then train and also evaluate the work of other people.
• 21:01
  But that tends to work better in a scaled environment.
• 21:03
  And the irony is we have firms in our industry that are becoming so large and scaled that they
• 21:07
  probably could you know, Bessemer's had for a long time, the, know, associate vice president.
• 21:12
  So you you can run a venture firm like a professional services business that's scaled.
• 21:17
  You can't run a boutique venture capital fund
• 21:19
  Yep.
• 21:19
  That way.
• 21:19
  And so I think that's that's one thing to consider that I wouldn't say that it's always a
• 21:24
  bad thing.
• 21:25
  But it is a very partner centric model and probably not great for developing talent.
• 21:30
  The development model comes with a different set of costs that I've at least countered.
• 21:34
  One is limited partners will ask you why are you spending so much time on management?
• 21:38
  I'm paying you to invest and support portfolio companies.
• 21:42
  Why are you dedicating so much of your time to management?
• 21:44
  Because I do a one on one with each of you almost every week.
• 21:48
  Plus interns.
• 21:49
  Plus the interns, plus our team meetings, plus one off things that I work on with each of you
• 21:53
  as projects.
• 21:54
  And what I tell our LPs is you don't get the benefit without the investment.
• 21:58
  So my belief is if I don't spend time with our team, getting to know you all, understanding
• 22:02
  your strengths and weaknesses, trying to give you work that will challenge you and help you
• 22:07
  expand and grow, but also giving you the autonomy to write checks yourselves, That's the
• 22:11
  investment I'm making.
• 22:12
  And when people leave after we've made a big investment, it does hurt.
• 22:16
  I'm like, wow, we put a lot of time and energy into, like, developing that person.
• 22:20
  But that's that's the job.
• 22:22
  Yep.
• 22:22
  That's the job.
• 22:23
  And I think if you wanna have a boutique firm I think of precursor as a boutique or a
• 22:27
  specialist firm, and I think of Amplify as a specialist firm.
• 22:30
  I think the development model is more personally satisfying to me.
• 22:34
  The idea that we can, like, choose people, get to know them over a long period of time, give
• 22:38
  them more responsibility and autonomy, and hopefully have them grow into being partners is
• 22:44
  to me a really satisfying role because if with the high turnover leverage model, if you wanna
• 22:49
  grow your partnership, you gotta go outside of the firm and find people that you know and or
• 22:53
  like and bring them in.
• 22:55
  And the comforting thing about the development model, if you get it right, is by the time
• 22:58
  people make partner, you have a lot of tape on them.
• 23:01
  And it's not just the investment tape, it's cultural.
• 23:03
  You know that they fit in at your firm.
• 23:05
  They know that you they came up with the values that your firm cares about.
• 23:08
  You've seen them over a a pretty long development arc, and there's trust.
• 23:12
  And so if you ask me, like, if the choices are bring in an outside person that you know well,
• 23:17
  but you're not really sure will fit in your environment, but save all of the management
• 23:22
  time or make a bunch of management investments that you know won't all or, you know, our LPs
• 23:26
  ask me, what are you gonna do if if the great people you've developed leave?
• 23:30
  And and to your comment about the AI founding engineers, I I worry about all the time.
• 23:34
  I'm like, we have great people here, I'm putting a lot of energy into making this a
• 23:37
  great place to be.
• 23:39
  But I expect other people will also see in them what I see in them.
• 23:42
  Yeah.
• 23:42
  So my my job is to like make this a great place to be.
• 23:45
  The piece also notes that the best early stage investors are natural collectors of people, and
• 23:50
  I like that term, building really genuine relationships over time.
• 23:53
  As a result, Mike says firms often don't need to post open positions and can instead draw
• 23:59
  talent from within their networks.
• 24:00
  I will say in our own case, nearly all of our interns have come from literally a broader
• 24:06
  internet search.
• 24:07
  And some of them, myself included, have gone on to become full time employees at the fund.
• 24:13
  So we have a similar approach as it comes to our submission form as well.
• 24:17
  So I wanted to pick your brain about the strategy.
• 24:20
  I'm helping somebody right now find a job in venture.
• 24:22
  And I told her, I said, every Venture fund is both always hiring and not hiring at every
• 24:28
  moment in time.
• 24:28
  And if you went to their website or asked around, there's probably no public role
• 24:32
  available.
• 24:33
  But every Venture firm I know is always thinking about I would say the most recent
• 24:37
  addition to our team, there wasn't really a wreck for that person.
• 24:41
  It was just like, well, we need this person on our team because she's fantastic.
• 24:43
  Yeah.
• 24:44
  We will create an opportunity here.
• 24:46
  And I think that's part of what makes getting into venture hard.
• 24:49
  Especially if you're on the outside looking in, you you sort of don't know.
• 24:53
  I tell people, there's all these clues, people who just raised a new fund or people who
• 24:56
  recently made promote there there are some, like, tells, but they're imperfect.
• 25:00
  Yes.
• 25:00
  And I think it's it makes it really hard for people outside of the industry to figure out.
• 25:05
  Those like, wow, that firm hired an associate.
• 25:07
  Didn't even know they were hiring.
• 25:08
  Was like, well, they they were.
• 25:09
  They just didn't wanna be public about it because they didn't want 5,000 resumes.
• 25:13
  Yeah.
• 25:13
  And I think that's like the the difference between I reflect on my time in big consulting
• 25:17
  versus a smaller boutique fund is if you're in a big consulting firm and you want a promotion
• 25:24
  the next year or you wanna hire more people under your team, you can't really snap your
• 25:28
  fingers and have an open discussion with the CEO of your consulting firm about that.
• 25:32
  They typically there's just there's no layers to do that.
• 25:35
  That's smaller fun.
• 25:35
  Like, we can have that discussion openly, especially around hiring of, like, what do we
• 25:40
  need today?
• 25:41
  What are we seeing within our interns, etcetera?
• 25:43
  And there is freedom and a mechanism in order to kind of be flexible there.
• 25:47
  So there's so much to be gained from that model.
• 25:49
  We are all out of time today.
• 25:50
  Thank you, Charles, for stopping by.
• 25:53
  Thank you to our listeners for popping in as well.
• 25:55
  We will talk with you all next week.
• 25:57
  Please like and subscribe.
• 25:58
  In the meantime, thanks.
• 25:59
  Bye.