Ep. #23, Do One Thing Well with Timothy Chen of Essence VC

Rachel Chalmers: Today, I am very excited to have Timothy Chen on the show. Tim is the managing partner at Essence VC with a decade of experience leading engineering and enterprise infra and open source communities and companies.

Prior to Essence, Tim was SVB of Engineering at Cosmos, a popular open source blockchain, SDK. Prior to Cosmos, Tim Co-founded Hyperpilot with Stanford Professor Christos Kozyrakis, leveraging decades of research to disrupt the enterprise infrastructure space, which later exited to Cloudera.

Prior to Hyperpilot, Tim was early employee at Mesosphere and CloudFoundry. Tim's also active in the open source space as an Apache member.

Thanks so much for coming on the show. It's great to have you.

Timothy "Tim" Chen: I'm super-excited, yeah.

Rachel: You have described Essence VC as being the investment firm you wish you'd had on your cap tables when you were a founder. Tell us more about that. Tell us how the fund came into being.

Tim: Yeah, that's how I describe myself and often to other founders too. But not to say any of my investors are bad at all.

Disclaimer here, they're all great. You know, obviously we've heard so many investors that could do bad behaviors and stuff like that. My experience with my own investors and you know, now being an investor, I feel like most people's intentions are generally good.

But to answer your question, what does that mean? Is the kind of investor, I wish I had my own cap table? It really comes down to who I am and what kind of company I was starting, right?

So you talk about my background, I was mostly engineering. You know, I built products. I joined early stage startups, first 10 employees twice.

And I thought I knew a lot about companies and startups when I started my own company. But I realize I don't. As a founder, I've never been responsible for sales. I never really had to think outside the box for so many different things to make a company work.

And I just realized once we raised our rounds, you know, raised a seed round back in 2016, first time fundraising, you know, and learned a lot, everybody sells what they do, right? That's what VCs do with founders.

You know, I went to Sand Hill Road, went to San Francisco, everybody says, this is all the stuff we do. And once we have a few of them on our cap table, what I realized when I say when I wish I had my cap table is really about talking about the people that I feel like are selling all these services, I realized it's really not specific to me, it's more generalized.

It's really about I can do the same thing to everybody and this everybody includes every kind of startup, right? Because most investors don't just back a particular space on a particular stage. Most of them, you know, even the ones that are cap table have, even that the ones that only do seed, they quite often do a huge range of SaaSs and different kinds of companies.

So their services are so broad, we help you hire, we help you find customers, we help you do things like that. And in reality, when you're just starting raising your first capital, your product most likely isn't there yet or you're still building it or you thought you have a product ready, but everything is a hypothesis at that point.

Nothing really is proven out. And what I found is so difficult to help us at this sort of pre-product market fit phase is that, you know, most help from VCs are like, "Tell us exactly what you need. Tell us exactly what is the specifics you're looking for and tell us exactly how you're looking to sell. And we're going to open up our big portfolio to just a couple because you know, we don't waste everyone's time."

And we give a very generic list of hires, tou can give a big spreadsheet. Everybody kind of is shared for everyone and none of that works for me. Yeah, they're just so generic.

We're looking for very specific people for specific mindsets with specific culture when it comes to hiring. And when it comes to finding customers, our product was not even built yet. We're still building it and I don't even know if it's going to work or not, in the first product.

Back then we had a lot of confidence of course, because we have to, to raise money saying this is the thing to do, this hot stuff to be building. But we learned quickly. I even wrote a blog post about our journey a little bit.

Basically, we had to go through three different products, even find the one that actually worked and took so much time.

And so, during that part of a journey, I realized the help we really needed wasn't really about generic pipeline hiring or just random introductions to certain kind of companies.

It's really trying to reduce the amount of mistakes we might make and also give us more insights about the things we might run into specifically for us and also help us able to navigate these unknowns, unknowns. And that's very difficult.

I think investors by default, they're always great at picking, you know, you should be a great investor, you have to be great at picking to some degree.

Rachel: But that's table stakes.

Tim: Yeah, but that's their side, right? It's not really our side. From our point of view, what can you actually do to help us? And you know, now of course now I'm on this side. Now I fully understand the best companies don't need any help, right, you know?

Rachel: Yes. Well I don't know that that's true. I mean some of my best outcomes, I really rolled up my sleeves and worked hard on them.

But I love how you frame that, because when I think about the real contributions I've made there, one was just like I was totally convinced that those founders were going to make it, I was just brought in from day one and I think just believing in them and cheering them up when they were having a bad day was pretty significant for them.

The other thing was I love working with technical founders on customer discovery, because as you say, product market fit is the thing. It's the most important thing that you could be working on.

And so de-risking by just increasing the surface area you have with potential ICPs is really fun and really worthwhile.

Tim: Yeah, yeah, and so I think that's really the most interesting part is because investor has so broad experience, they always use that as their advantage to some degree, because we can, we sing SaaSs, we sing, you know, all kind kinds of company including maybe the ones we heard of like Twilio and some others.

So therefore we can give you good advice. I know none of the advice actually worked at all, but I definitely have found angels are really useful.

And when I kind of reflect back why they are so much more impactful, one is I think, you know, I find at least some of the angels we work with, they just have more intimate knowledge and cadence, they work with teams than some of the other investors are.

So they kind of like just know, "Oh I should work with that these teams are similar to your space just recently," right? You know, here's some specifics so that's helpful, right?

Or, and they have much more specific context on certain people and just frankly just be available. It's already like a pretty big win versus not, 'cause you know, oftentimes--

I guess to some degree if you're good looking team, we can pick investors, but when you have investors you haven't really established this trust or friendship or whatever. And also we learn quickly investors really selling them and getting into round once you're in, you know.

Rachel: Yeah, you're yesterday's news.

Tim: Yeah, I've done my job almost, right? You know, my sales is done. I'm just on life support.

Kind of like, you know, you still have to have this relationship you need to build, to feel comfortable asking whatever you think that's really specific or just questions.

And so the sort of relationship-building phase that we are able to do are only people can meet regularly. And also with the people that have so much context for our space, it really made pretty big difference. Obviously relatively bigger difference than any other investor at our cap table.

So that's kind of what I meant when I say want to be someone I want to be on my own cap table is really mimicking sort of the best ones and also have our own thoughts to it.

What I see on my own cap table now becoming Essence VC, you know, I never knew how to be a VC whatsoever. You know, this is my first time doing it. I never worked at a VC fund, I never even been in financial, I never took an MBA, none of that stuff.

So there's so much to learn on a lot of other aspects of VC. But one thing I always feel like I know really well is working with founders, because I've been one and it's not that long ago and I'm not going to be a perfect VC for everybody for sure.

But I think for anybody that looks, not looks like me but have similar backgrounds with similar kind of space, I always felt very relatively confident that we should be so much more useful than anyone else.

And so far seems to be true. You know, you never know, you cannot be 100% right, but we feel like we've done more good things than bad so far.

Rachel: I love the approach and I love the genuine founder friendliness. But you are incredibly prolific. You have so many logos on your website, how do you manage your time across that portfolio? How do you scale Tim?

Tim: You know, and people literally use that question to me, or scaling Tim and the more I thought about it it's like this is just not scalable.

I would say that the approach we want to take is unscalable by design, 'cause if it's scalable it means everybody else can do pretty much the same thing.

So you know, we're a boutique fund with a very boutique value and it's complete unscalable, right? So we have to have approach what kind of companies we work with and how we work with them and how do we make sure we have enough time and energy, but also understand how we make sure that our founders understand what we do really well.

So, you know, I talked to so many other GP friends and you know, this is a very common thing. I also heard from other friends that are, you know, going from really smaller funds and a little bit larger.

They're much larger than me. But scaling the portfolio size I guess, you know, your time by default will be spread in much more thinner.

That means we have to be so much more specific what we tell founders what we do, 'cause you know, when I started it was just like, "Hey I'm Tim, you know, I never heard of Essence VC. It's okay, you know, as a founder we do everything, come anytime, text me anything, you know," and we still--

I still text and answer every question and everything we could, but it's also good to set expectations to founders like, "Hey, we're probably not the best at introducing so many customers to you, because we're not here to be the most tough funnel customer introductions for you."

Some funds only focus on that, right? So we're going to introduce quite a few, you know, most of our companies, we have definitely introduced quite a few companies and design partners, but we're not going to be always the number one compared to everyone else, right?

I basically had to set expectations to founders like, "This is what we are so boutique good at. We know the space well. We know so many different companies went through exactly your phase, including myself.

So it's less about the particular hiring or getting customers. It's really about our touch points and understanding how do we be open to talking about the journey you're on and the thought process you have, because what we are really good at so far is almost like company momentum driving and positioning.

Fundamentally how do you pick and choose what to do every week? We are really good at help you brainstorming some of that, you know, in all aspects to it." It's also not very specific, because I realized every company has a little different need to push different things.

So it's not particular to just hiring and just so it's still a bit of a unique feel to it. But yeah, so the most important thing is being available to some degree pre-product market fit.

So oftentimes when you have a sales team or you have repeatable sales, we don't have to do biweekly or you know, or weekly sync up anymore. You're okay. You know, you have more folks that has so much more to offer ahead of sales playbooks, runtime, you know GTM playbooks, you know.

We're still helpful, we've seen a lot, but as well, I want to really focus on your pre-product or pretty much your pre-product market phase, 'cause that phase, nobody helps the best, right?

And that's really some of the most scariest time as a first-time founder or first-time going to any space. I tell founders we focus on that phase. You're busy, you got sales teams go for it. We can drop our syncs, but hey, if you have no sales teams, you barely make any sales.

We got to do the syncs. You know, we got to chats, we got to talk, 'cause that's the only way we show our value and help.

We basically have a time window approach basically, for founders going through pre-product markets and that could actually take years.

Unfortunately, it could take years. Some people actually only take months.

Rachel: Yeah, but they are the exceptions, yeah.

Tim: "We have teams raising three rounds in six months." "Okay sir, I will, I either get to chat with you that much, but you know, you're good sir. You know, you can do whatever you want."

Rachel: Go with God.

Tim: So it's sort of like a natural selection. We don't by design retire on people in a cohort or we don't have a time window cohort approach. We only have a sales sort of momentum approach where if you're in this repeatable sales, you know, we can be more hands-off.

Rachel: I do think this is one of the great benefits of the proliferation of smaller firms. We don't all have to be Kleiner Perkins anymore. We can specialize and we can pick the thing that we're best at.

Tim: Yeah.

Rachel: And startups can have a diverse cap table at different investors they can call on for different things.

Tim: And different personality. You know, I found some people I feel way more comfortable talking about certain things too.

But some people I probably won't talk about some certain things, like people issues, probably for people that have been manager and more empathetic, you know, having a hard time actually closing some certain deals and I don't know how to drive certain things.

There's certain people I probably feel like you're probably better at that, right? So I feel like personality and almost mindset plays in a big part of that too. Beyond just your harder skills, your softer skills often plays that, at play.

Rachel: How do you look at the investment landscape for AI? Not just gen AI but all kinds of ML, what are you most excited about?

Tim: Well, given every company is AI and ML, at this point, it's very difficult sometimes now to say what I'm actually more excited and less excited, because like I said, it's everything is AI.

So you can't not look at a AI company. And I guess most people know it's funny, every company I talk to, even though you're not doing AI-specific, AI for sure has to be in our deck, because even founders are getting, every investor telling them you got to have AI.

I don't even want to talk to you, if you have no AI. It's so funny, even the dev tool infra company doing networking, doing security, you know, that doesn't specifically just for AI.

Big 70% of the deck is all AI everywhere, you know, but we keep hearing from investors telling us "AI is so important, you know, that's why we have AI workload production, AI this, AI that."

And I understand why, you know. So to back to your question, why I'm most excited? I feel like because I'm back really early stage, I've been backing people pre-idea quite a few times as well. Pre-product is almost almost a given to some degree. But pre-idea is sometimes a little scarier.

And do I do AI, not do AI? To me doesn't really make much difference. I feel like I'm much more excited to back people. I found somebody actually really, really good and they have really good team with a really good mindset, have a deep enough knowledge of certain space, have a really open willingness to be, take feedback, right?

Have a unique opinion about certain things. I'm so excited always to work with teams like that. And sometimes it doesn't even matter what they do.

It could be a, you know, we backed a debt collection company. You know, I don't know much about debt collection pretty much at all, I have to learn about it. But the team is so great at knowing how to navigate that space.

And I wasn't looking for one, but I got excited, because this is the kind of team I like to back, that they have very passionate take on a space that they spent years for in a space that I feel like has not been competition yet and there's a real shot for them to do something.

And so investors always say that they're excited about certain things and we have to. There's certain areas where's some are competitive, you need to have such a thematic way to look at our thesis coming in.

But we're also very opportunistic as well, because if we see a great team, I'm always going to be excited about the great teams, but even on their ideas and areas I don't know super-well.

But on the flip-side on some of these particularly interesting or exciting things I'm always looking for or I have particular ideas for, you know, for a while I've been very excited about finding people doing multimodal LLMs.

You know, just since a year or two ago, almost three, four years ago actually we backed Rosebud, one of the earliest AI design AI sort of image generator almost three, four years ago. You know aesthetics, AI generative design company almost two and a half years ago.

So this is all pre-LLMs, pre-everything. Again, we're just taking off. So I was super-excited to back these early and going into today everything is LLM and AIs and multimodal is just taking off like music all kind of stuff is going on.

But I feel like from an infra tooling standpoint, multimodal sort of like infra is still open. You know, there's still a lot of things to do around not just specifically the workflows and the toolings, but the whole stack actually.

You know, video, audio generation, real-time latency stuff, you know, all kind of things require. I think that given the change of the content, it does require a pretty big shift of all the sort of infrastructure required to produce, to serve, to iterate, to do a lot of, you know, production type of value, application, billing stuff.

So I'm actually always excited to see somebody going after something that unlocks a different kind of data for you to use AI, you know, Unstructured is one company for example we did that kind of fits that bill.

Rachel: Which parts of AI risk being over-hyped do you think? Where is there more noise than signal?

Tim: Maybe the first thing I would think of is agents, because I think agents is what everybody's talking about. Back in the blockchain days it was always a hype thing, right?

And it kind of sort of died. I worked on blockchain for about two years and you know, back in the day was of course, Bitcoin, Ethereum and then you know you get NFTs and then DAOs right now.

Now I feel so weird even mentioning these things anymore, right? But at the point of time there was this new, almost shiny tools or shiny stars you can collect like, "Oh wow, you got to do an NFT company, right? You got to do a DAO company, because that's the thing that's going to change everything," right?

And in the end, these things change some things, right? It's not like they're completely useless, but certainly didn't make that big of a difference, because it's a tool, it's not a actual paradigm shift. And I think AI agents, I kind of put into some level of camp like that where I don't think AI agents is like a paradigm shift.

It's almost basically like an abstraction or a tool to some degree as well. It's very, very useful. I think a lot of companies are building agents to empower lots of things.

And so it's just funny like last year nobody mentioned the word agents at all. Now it's not just LLMs. We're building the agent for insurance, we're building the agents for that, right? It was used to be GPT for insurance. Now it's agent for Y.

But if you can easily just change a word and basically doing, you're just kind of following what the state of the art is at that point in time means the state of the art will change rapidly, quickly next as well.

I think agent probably will stay just like NFT DAOs was staying, but it's probably not going to be this exact same way you look at it today, right? It's a choice. It's a choice why you want to use agents and how you want to use agents.

It's just probably over-hyped to a point where every agent will become huge and every agent will become, you know, the 007 that just assassinates every market. I think it's just much more marketing than actual realness to it, 'cause agent doesn't work well as it is today.

There's so many stuff that doesn't really able to actually produce real value yet. So you can't just use agents, it's part of your solution, but you get so much more you have to do here.

So I think agents for now is a good market to look at. I actually do believe in agents. We were backing an agent framework for example, but from a high value, is agent the answer for all things in life? Probably not.

You know, not the overall category thing you want to stick on everything and that's the thing that will make your company actually work. Just because you agents for insurance, doesn't mean your company for insurance doesn't mean your product would definitely stand up.

So to me, agents as probably overly used or overly hyped used word for now. You know for folks that we understand the space, we're always looking for specific nuances way beyond just using agents, right?

Rachel: How do you see AI making engineers lives easier or harder?

Tim: You know, we do dev tool investments a lot. AI is everywhere. Chatbots, Copilots, the whole nine yards, right? AI SREs, AI whatever.

My own company, Hyperpilot, we were actually using AI ML to help recommend instance types. We're actually doing runtime isolation for workloads using ML as well back in 2016.

So I was basically doing an AI for SRE-ish company, you know, almost eight years ago.

And I realized quickly that SREs hate AI. They hate ML. All these things that sound nice on paper that could save you so much work. And they look at it like, "This is going to give me so much pain." It is not a comforting thing.

So I mean to answer your question like how is AI changing engineering? I actually always feel like the first reaction now I have is not about how much work is going to help, it's also how much pain it'll create for developers.

Just because you know, obviously we've looked at all the Copilots of the world, all the sort of Cody and there's a bunch of VS code forks, right? That adding AI into your IDEs and you know, it's going to be very useful and so they're like common cases. but it's going to make your edge cases so much damn harder to figure out.

And I think most people building tools always just highlights, "Look how good it is to debug a performance issue. Look how good it is to find one line fix in your code base."

It's like great, I'm not going to be excited just seeing somebody do that, you know, 'cause it's going to be so much false negatives, false positives, it's to a point like I don't even want to turn this on all the time.

So it's going to help a lot, but also going to create a lot of... It doesn't really replace the need to actually go down deep in every single thing yet. So there's probably some iterations required. But fundamentally, I think AI is here to stay. There's some workflows very specifically good to help engineers.

One, write code, especially common patterns, you know, very specific boilerplate stuff. This is going to be great. AI can be really great to help on a very almost, you know, mundane task kind of stuff.

Rachel: Mhmm.

Tim: I think is also very good at helping you organize information. A lot of security stuff is probably so good to use AI for, because they can help you catch so much signals and give you some level of insights for it.

And maybe even looking at your logs and insight, all the kind of like observability stuff. There's some level of truthness that you can help use AI to help you distill some information down.

But yeah, it's pretty consensus at this point that everybody believes AI is going to help do everything for developers.

I think beyond just the common things, the very specific problems you want AI to help you solve, especially almost like the SRE-type things, it's going to take quite a lot of nuance to make it great, and most people kind of focus too much on AI, too little on systems.

And I feel like the best teams got to do both. You got to understand the systems really well and understand AI really well, you know.

You can't just sprinkle well and dust to it or add an agent, you know, your agent's is going to do whatever, you know, yeah.

Rachel: You have written about the importance of founder culture to a startup. How do you see startup culture informing the tools that we build in good ways and bad ways?

Tim: That's a really good question. Founder culture, you know, if you work at startups before, you definitely realize quickly and especially we're working so much founders too as investor, but I feel like it's so much more visible if you're actually work in a company day-to-day.

That founder culture really drives everything. And what I mean by culture is basically your founder's mindset or maybe another word to use is mental health. There's some certain degree of how they think and how they decide to do things, how they choose to communicate things is how a company actually gets anything done and how they actually make decisions. And how you make decisions that a company drives products.

It drives what kind of product can be made and drives what kind of product will be delivered and how we make those decisions. And so it's probably less of a directness, founder's culture will make our dev tools use more colors or you know, uses rust versus gold, product doesn't really drive those specific choices.

It's more about the people they can actually able to hire and retain. The cadences where they actually ship and do what's with, are they good at making long-term product investment? That's sort of short-term investment bets, right?

That drives a lot of kind of the products you can actually build and building dev tool in any infra products is hard, because it takes so much work, right? People expectations always super-high, 'cause there's always a workaround.

You know, you're never going to find an infra easily that has no workarounds. There's always going to be one. It's just so much harder to use. But you're fighting with developers want to do it yourselves.

So the bar is always high, but the founder culture plays a big role, because you have to find people that are good to have a very strong sense of what good means.

You had to give it autonomy while not going to completely lose control, do whatever they want, right? And how do you make good product decisions in a company?

You know, it will show up in any product you use, you know, you use for sales products or versus you use PlanetScale or uses Neon or you know, pick any other product, Heroku even today, right?

Heroku, the culture of the company is fully reflecting Heroku's current product at this point, right? For something that's different than what you see in the past. I always probably argue some other companies, once you have fundamental shifts of teams, your product just goes totally different directions.

And that's also a reflection of a culture, 'cause people no longer care about certain things or people start to really care about certain things. It really changes what is being optimized. And I think oftentimes every product, you cannot optimize everything.

You often only can optimize a few things. So that really drives from the culture point of view. What does a founder actually want to deliver here? How do they sell the company for success? What are the things they really able to sell that we're currently doing?

I know this is still very abstract speaking-wise, but that's kind of what I learned working in startups and also my own company that so many decisions-making are really less about what the reality of the market also means.

But also kind of dependent on how the founders, what their belief system is, 'cause there's no data to it. Oftentimes, startups are not just purely based on a survey where we are adding a bunch of features. You're not doing surveys everywhere for every single decision.

It's too slow. So somebody has to make a judgment call on how to do things and that really comes down to the type of things you could actually do well.

You know, funny enough, I do care about that a lot when I even make decisions on investing on people.

Rachel: I think that the great founders build products that reflect their character and their qualities.

So Edith Harbaugh is incredibly tenacious and resilient and her software makes its customers more resilient and more creative.

I find it fascinating the way we distill our best practices into software and the software still retains those traces of its creator.

Tim: Yeah, so that's about, it's not just because I worked at, I don't know, Docker or I worked at the places, therefore I'm going to be the best, right?

It's not really just about what you did. It's sometimes the personality and the mindset of the people. Is this person or this team able to create a company that revolves on the belief system to make this work?

Rachel: I'm going to make you emperor of the solar system. For the next five years everything goes the way you think it should. What does the world look like in five years?

Tim: To emperor of the solar system? Well, that's a pretty big system beyond the Earth and everything, right?

Rachel: You've got a palace on Jupiter.

Tim: Ah, so that's an interesting question. Just out of the whim. I don't think about this at all, but if I want to have any say of anything, you know, I think the whole universe always revolves around people, and people usually... Going to VCs, and you probably feel this too. I didn't realize how much decision-making power is really lies on just a few people.

Rachel: Yes.

Tim: And I feel like the whole world is like this, you know?

Rachel: Yes, yeah.

Tim: It comes out to just a few people, even though we supposed to reflect everyone, but it's impossible.

You know, it's a chain of trust system in the end and everybody just has to trust on somebody and nobody can ever make a fully perfect decision or decision framework, right?

Every decision framework has flaws and every system has its own ups and downs and just like, I feel like the universe and just talking about the Earth and how people make decisions overall, it leaves a lot of people unheard of.

And unless your issue is so directly related to the people on the top or you have very specific, large population, very noisy, a lot of issues get unheard of.

And the reason I even bring this up, because we adopted three children ourselves and you know, adoption particularly is an issue that, you know, is heard of, you know, but at least on the Asian culture and Asian side of the world, nobody talks about it.

Not big enough issue, right? Not a big enough thing. And there's so many different kinds of issues like that.

So if the universe can be run by me, there's always going to be, it's not sure how, but I will love to find ways to empower more people to be able to solve more problems with less resources.

B ecause it always, to solve a lot of problems, you mean systematic changes and sometimes we can't make them, so I'm not sure we can actually make systematic changes for everyone, but I'm sure there's a lot more people want to solve a lot more problems, but just don't have any other means to do it.

Requires donation, requires a lot of people putting up political campaigns, this and that. But not everybody knows how to do that really well. So to solve particular issues, there's certain things that needs to be done.

I wish there's way more simpler ways for folks to solve more problems that is more localized, but can actually be a lot more aware of. I don't know, that's just my thought process.

Rachel: Yeah, that's a really thoughtful answer and I'm 100% onboard. Last question, my favorite question. If you had your own generation ship on a 100-year journey to Alpha Centauri, what would you call it?

Tim: What do I call it? I would say something that kind of resolves on importance or matters or something like that. Maybe like, I'm not really good with names. I always try to find synonymous names of something similar to adjectives I'm looking for.

Maybe I could just call it Matters, I guess. That for me, I wish people can find what's most important for them and focus on them and only focus on most important in life, you know, of anybody in the generation ship, right?

It's like just do one thing well or just do a few things well. Let's not try to solve everything and not try to do everything, you know, I wish everyone can find their true passion in life.

I know a lot of people still forever, not able to find something. They're always as passionate or able to actually have the opportunity to do what they feel like they're passionate.

And so that leaves a lot of people not able to pursue their dreams. So anyway, maybe it's something like that nature.

Rachel: I love that. Tim, it's been a delight to have you on the show. Thank you so much for taking the time.

Tim: Of course, happy to be on.