February 1st, 2021 · Long Essay

Startup Year

A Review of My Year at Jupiter (YC S19)

At the end of 2019, I moved to a new place in San Francisco and started my career as a startup software engineer. I am now planning on returning to college. My 2020 in review.

The Painted Ladies against an orange sky

I started writing this piece in October 2020, around when I was beginning to transition out of my role at Jupiter. At the time, twelve months prior was October 2019.

Life was good in October 2019. Venture capital investment activity was on the cusp of hitting another record, the stock market was reaching all-time highs, and the California wildfire season had been mild so far. I was recently dropped out of college and starting my first full-time job as a startup software engineer at Jupiter, and I had just moved into a spacious new home with my coworkers in San Francisco, a few blocks away from the Painted Ladies.

Life one year later, at least on some of these dimensions, seems unsettlingly similar. The S&P 500 is up over 15% since last October, marching inexorably towards another all-time high.1

1

As of this publication, fifteen months later, the S&P 500 is up about 30%.

× Close

This fall seems frothy as ever in terms of fundraising; a company that graduated from YC just this past August recently raised a mammoth $16 million at a $75 million valuation (YC invested three months prior at a $2 million valuation). One founder described the venture market right now as “pants on head insane […] money is flying around and valuations don’t really have concrete backing.”

Yet — this year’s wildfires were the worst in California history, the air filling with so much smoke that the sun didn’t rise one day, and that’s not even getting to the hell of a year otherwise — pandemic, economic crisis, protests, racial injustice — that 2020 has been so far for all of us.

From its highest highs to its lowest lows, it’s been quite the year. Between all that’s been going on, I’m thankful to have had the fortune (I’d love to say foresight, but I can’t) to choose to leave school last year to work for a startup, and I’m thankful to have been generally spared from the overall crisis. I hope that things get better for all of us in the weeks and months ahead.


I’m writing this now because — amidst the whirlwind of changes happening in the world right now — there has also been a big change in my personal life: I’ve left my job at Jupiter.

Over the past year, I’ve learned from and thought a lot about this startup experience, from:

  1. specific skills I’ve developed on the job as a software engineer
  2. to new perspectives on starting my career at an early-stage company
  3. to certain feelings about life that I think only living independently in San Francisco could impart

In my review below, I’ll go over the detailed pros and cons of each of these aspects of my experience from the past year. I’ll start with an evaluation of my job in itself, followed by observations applicable to my career overall, followed by thoughts on “living”, in a more general sense, that I think are worthwhile to share.

I’m excited for what this next year will bring, but in the meantime, I hope this review can serve as a useful chronicle of my startup experience in 2020, inform those who may be thinking about taking the startup route as well, and most importantly, express my thanks to the team that decided to take a chance on me one year ago.

I. On the Job

Pros

Autonomy

From the start, I had the autonomy to choose the entire tech stack — from the TypeScript frontend all the way to the Kubernetes infrastructure — and the result was that no matter what I was working on, I’d be using a technology that I was personally invested and interested in and thought was the best tool for the job.

I was able to get a better sense of how to make good technical decisions going forward, as certain tools turned out to be great choices and others, a little bit less than great.

Choosing Kubernetes, Rollup, and Bazel

424 words

  1. I’d initially chosen to put our infrastructure on Amazon’s proprietary container orchestration service, Fargate, since I had prior experience with AWS and some of our other infrastructure was already on AWS. I figured there would be good interoperability. However, community support for Fargate was lacking (and we weren’t ready to pay for premium AWS support) and documentation was poor, leading to a bunch of hard-to-debug issues. We moved to Google Cloud’s GKE instead. Kubernetes’ thriving open-source community was invaluable in helping us get set up and debug issues, and I learned the importance of developer mindshare and the power of open-source.
  2. Using Webpack to bundle our code seemed like an obvious choice at first, but as our internal shared web libraries grew in size, our bundles became huge. I was doubtful of Rollup at first since it seemed to be less well-supported (and had significantly less GitHub stars), but switching to Rollup yielded huge reductions in bundle size, and I’ve permanently added it to my toolkit as an alternative to Webpack for bundling frontend libraries I write. Although Kubernetes’ thriving community was much better for us than Fargate, I learned that I shouldn’t discount tools I consider to be “also-rans,” because they just might really be better tools for the job at hand.2

    2

    To be fair, Rollup is growing fast and isn’t really an “also-ran,” but at the time, Webpack was huge in my mind.

    × Close

  3. I initially set up Gradle as our build tool because it was the most modern JVM build tool that also had significant buy-in from the ecosystem (the second constraint reduced the scope of potential options to essentially Gradle or Maven). However, we ran into a lot of trouble with Gradle’s protobuf compilation. Because Google invented protobufs, I prototyped Google’s Bazel build tool instead (which I was initially wary of because of the lack of a strong ecosystem outside of Google-supported projects), and it worked like a charm. Nonetheless, since implementing it, we’ve had problems with Bazel that we’ve only been able to resolve by combing through BUILD files in Google open source repositories — problems that could have been avoided if Bazel had a stronger community. Ultimately, it’s all about tradeoffs, and if the team ends up switching back to Gradle, I’d be disappointed that Bazel didn’t end up working out but not completely surprised.

I now have better instincts for which characteristics of a tool I should weight most heavily, depending on factors like the problem at hand and the impact of adopting the tool.

Choosing Kotlin

203 words

Although there was a decent amount of switching tools and prototyping early on, one choice I stand firmly behind is that of Kotlin as our main backend language, which seemed kind of risky at the time given its novelty. Nonetheless:

  1. by compiling Kotlin for the JVM, we were able to take advantage of the strength of the JVM ecosystem (with battle-tested libraries like gRPC-Java and Apache Commons)
  2. the syntax was relatively much easier to learn, especially coming from a Java- and JavaScript-heavy background (compared to say, a Scala or Clojure)
  3. overall, the language was expressive, readable, and a joy to work with

The biggest impact, though, was that it seemed to really attract new engineering candidates — people told us they specifically responded to our hiring outreach emails because we mentioned we used Kotlin on the backend.3

3

Paul Graham’s “Beating the Averages” mentions a similar phenomenon with Lisp.

× Close

For what it’s worth, the most recent Stack Overflow Developer Survey seems to support the choice, too: Kotlin was the fourth-most loved language by developers in 2020, after Rust, TypeScript, and Python.

Scope

Outside of autonomy in technical decision-making, working in such a wide technical scope also kept my days interesting: I could be optimizing React performance in the morning and then figuring out how to write a custom Terraform provider in Go in the afternoon; later that evening, I could be debugging a gRPC proxy issue. I gained a better understanding of everything that needs to come together to deploy a reliable production application.

Scope extended outside of the purely technical aspects, too: I was also given wide latitude to contribute to engineering processes and culture from scratch.

Besides instituting code review and introducing a few tests (which I unfortunately began to neglect as our product and code changed faster than I could keep up), I learned a lot during my initial research into development processes about concepts like:

  1. Continuous integration and continuous delivery (CI/CD)
  2. Git workflows: tradeoffs between rebasing vs. merge commits and trunk-based development vs. long-lived feature branches
  3. Releasing: continuously vs. daily or weekly

These were things I never really had to think about at a big company (where all of that was already set in stone for me). I was also able to create an engineering blog to help with recruiting, write a blog post about our tech stack, recruit an intern from that blog post, and do fifteen technical interviews for full-time software engineers when we were hiring.4

4

One unintended benefit is that it’s made me a lot more comfortable as an interviewee; I can kind of understand what people are looking for now.

× Close

All in all, I was able to do and learn a lot of things that I don’t think I would have been able to do in a larger organization or as a later hire.

Cons

A quick disclaimer:

This is all from my personal perspective and reflects my opinions alone.

You probably aren’t hearing the full story about working at a startup from other startup employees. I think it’s a cultural thing. All startups are challenging in one way or another; it’s just that nobody is willing to say anything remotely not positive. The fact that I feel comfortable enough to speak honestly about my experience at Jupiter — both the things I loved and the things I had difficulty with — probably speaks more highly of the founding team of Jupiter than it does of all the other startups that you don’t hear much about — so keep that in mind.

At the end of the day, I want Jupiter to succeed. Equity aligns incentives, and I own a small part of the company. If they do great, I do great.

Every situation is unique, and my experience will never map one-to-one to anyone else’s (and vice versa). I believe in the power of transparency, and I share the details of my experience with the hope that if you’re considering joining a startup, this helps you make a more informed decision, and if you just started working for one, this prepares you for the road ahead. We’re all best off when we’re in the places we can thrive the most.

Technical and Nontechnical

One of the hardest parts of the job — especially as a highly technical person — was working with a majority nontechnical founding team (one engineer and three MBAs). Much of the founders’ prior experience was with physical products and e-commerce operations, not necessarily software.

Jupiter was an ops-heavy business, so it made sense, but it was difficult in two ways.

First, the immediate priorities at Jupiter, by nature of the business, were operations-focused. Solid technical principles generally took a backseat to rapid feature development. As the weeks went on, I could feel the technical debt piling up: dependencies between packages got more and more tangled; convoluted SQL queries began to overload the database.5

5

For what it’s worth, I’ve gained a lot of respect for SQL’s capabilities as a data retrieval language. I learned so many new keywords I’d never used before.

× Close

I pushed for a senior engineering hire to help us very early on, but I think operations were a higher priority then and we didn’t get around to making a senior hire until a month before I was planning to leave.

The second challenge, perhaps related to the above, was that I wasn’t used to communicating with less-technical people about technical concepts on such a frequent basis. While not hiring an engineer as early as I suggested might’ve made sense given the priorities on the operations side, I think if I had been able to better communicate the technical challenges we were facing as an engineering team to the other founders, we might have gotten a senior engineer a bit earlier, which would’ve made work on the engineering side less chaotic and would’ve established a stronger technical foundation for us.

While I believe this experience has taught me how to better communicate between technical and nontechnical sides of the business6

6

A few simple things that improved communication: writing more approachable documentation, coming up with good analogies for technical concepts, and generally being patient (in both directions).

× Close

, I’ve also realized that I prefer engineering-driven teams and more technical products.

Team Composition

As first-time founders, I think the founding team was learning how to lead7

7

A new-grad friend, who recently started leading a team, shared this with me after reading this:

“Learning to lead has been huge for me over the past year. I think understanding that in some ways, everyone is doing everything for the first time is a helpful (albeit maybe naive) perspective.”

× Close

at the same time as I was learning how to be a good startup employee. As a result, it was sometimes difficult being essentially the only employee for the first ten months — outside of our delivery drivers and warehouse staff, the team was just me and the four founders during that time.

For instance, at the start, the four founders already had an established relationship. It took me a while to figure out their working style — I shied away from overt conflict8

8

This was something that came up in my exit interview. I appreciate that the founders took the time to give me candid feedback, and the things they shared with me are things that I will definitely keep in mind to be effective in future jobs. I had a chance to share some of my own feedback during the interview, too. I’m not sure how common exit interviews are in general, but it was very valuable for both sides. I think all startups should do them.

× Close

for a long time since I didn’t realize that was how they communicated. There were no other employees I could consult with.

Nonetheless, I’ve learned to prioritize explicit communication and to say what I really think as early as I can, so progress and discussion can happen as soon as possible. This was not something I was automatically accustomed to doing before working at Jupiter.

Pace of Learning

Learning was the main reason I joined the company. After maybe six months or so, however, I felt the learning begin to slow down; after the initial technical architecture was set up, there was a lot of “aligning divs”/optimization-type work that wasn’t as exciting as the initial parts. While I didn’t expect the work to always be something novel and exciting (I think it’s unreasonable to expect that of anything), it came to a point where it began to feel tedious.

Part of this was that I felt that we were starting to take a much more scientific, Lean Startup-style approach even though I felt that a more radical, Steve-Jobs-style, “people don’t know what they want until you show it to them” attitude was necessary to reach product-market fit. It felt like I spent the last few months at the company doing a lot of conversion rate optimization. In retrospect, I’m not sure what approach would’ve been better, and maybe I would’ve liked a Steve-Jobs-style approach only because it would’ve been more “exciting” but it wouldn’t have really made a difference.

I think the root cause was probably that the company was growing slower than we had projected. In addition to affecting the type of work I was doing, psychologically, I think that also took a toll on how interesting and fast-paced the work felt. I know I will need to get more used to this if I want to continue a career in startups.


Altogether, the greatest difficulty, but also the greatest opportunity of the job — which I hope to have grown into a little bit better over the past year — was that I was the sole employee, mostly treated like another founder responsibility-wise. For those who suggest joining a startup for lightning-fast career growth, I am 100% sure that that happened to me.

That brings me to:

II. On Career

Pros

New Perspectives

Working at Jupiter gave me concrete evidence that I didn’t need a college degree to have a career in startups. When I was first in talks with the founders to join, they cared only that I could bring value. For a nineteen-year-old kid, the cash compensation was excellent.9

9

$100,000. Compared on an annualized basis, this was actually less than what I was getting at Google as an intern — which should give you an idea of big tech’s compensation philosophy and how crazy the market for software engineers was at the time (and probably still is).

× Close

I really appreciate that this was the founders’ attitude and that they were willing to take a chance on me like that.

More broadly, their indifference to my lack of a degree was a refreshing change of perspective from what I had been taught all throughout school, and it’s shaped my decision about what I’m going to do when I go back to college: I’m planning on focusing more on studying what I think is interesting than anything decidedly pre-professional.

I have avid entrepreneurial aspirations, and by joining so early, before we were too siloed, there was also a decent bit of osmosis of nontechnical, general business and startup concepts like unit economics, contribution margin, the need for general liability insurance, and sending investor updates that I absorbed simply from seeing the founders get everything set up. There are still lots of things I’m missing — like I don’t have a great understanding of the specifics of equity financing since I joined after they raised their seed round — but I think everything I’ve picked up will still be really good preparation if I start my own thing, and it’ll give me a unique edge and perspective versus if I was just a solid engineer with no business experience.

YC Community

Joining a YC company specifically was also a huge boon. For one, I was also able to meet many other YC founders and employees through the network. Everyone I met was exceptionally friendly, ambitious, and open to helping or mentoring me. These are people I know I can call up for advice, tips, and maybe even an investment. Some of these people, especially those closer to my age, have ended up becoming great friends, too, and — between spontaneous trips to Tahoe, late-night ice skating, and rooftop birthday parties — made my time in San Francisco a lot more collegiate than I would’ve expected when I first left Ann Arbor.

Another YC benefit is access to the YC partners, who are typically former founders or early startup employees and who have seen enough companies go through the program that they can give some pretty good advice. I was able to tag along to a few of the partner meetings, and each time I learned a lot, from what kind of company culture is best for the seed stage to interesting stories from the partners’ previous ventures. One of my favorite insights was from Jared Friedman, whose explanation of how “move fast and break things” was a great culture for Facebook early on, but a poor culture now given its global impact today helped me internalize how company cultures need to evolve.

Company Cultures

Speaking of company cultures, the culture at Jupiter was different — to say the least — from the one I experienced over the summer at Google, and working in such a dissimilar environment helped me get a better idea of things I’m looking for in a company and what kind of culture I’d want to be a part of. I’ve learned that I do have a slight preference for earlier-stage, but the chaos at Jupiter was sometimes a little overwhelming. Now that I’ve tried a public company and a seed-stage company, one thing I’d like to do going forward is to maybe try a Series A or B company, which might provide a more manageable balance of chaos and stability. I’d also maybe like to try a company with a radically different corporate culture; none in particular, but a company with a culture as distinctive as Netflix’s “We’re a team, not a family” or Amazon’s “It is always Day 1.” As Jeff Bezos said in his 2015 letter, there are lots of high-performing corporate cultures, and the best way for me to know which one I like the most is to try a bunch; through Jupiter, I’ve gotten a little closer to figuring that out.

Professional Network

Although not specific to working at a startup, being in San Francisco and being surrounded by so many other people working in startups and technology pushed me to be better at staying in touch with people and “networking”.10

10

I feel like the term “networking” has sort of a bad rap so I sometimes hesitate to use it, but it’s probably the most precise term for what I’m describing here.

× Close

For instance, it was easy to keep in touch with my Google intern mentor since we could have lunch at the San Francisco office once every few months — former interns typically don’t get the chance to do this since they have to go back to school. I met a lot of cool new people by virtue of being in the city.

One of the most interesting “networking” tips I learned was from another YC founder who I met after leaving Google and starting at Jupiter. He told me he took notes after every new meeting he had. I remember thinking that the idea was kind of crazy, but as more and more meetings moved to Zoom and started blending together, I decided to give it a whirl. I created a Google Sheet with columns for the date of the meeting, the name of the person, from whom I got the intro, and my notes. It’s turned out to be super helpful, and I love being able to reference previous conversations when I talk to someone again; it helps keep things interesting and moving forward and I think it’s helped me maintain a greater number of relationships than I would’ve been able to without.11

11

Some people are very opposed to the whole “people on spreadsheets” idea. I like to think that my spreadsheet is just another tool, like Google or Wikipedia or the Contacts app: it’s hard to remember URLs and facts and phone numbers, and it’s sometimes similarly hard to remember the specific details of calls and meetings. I also think it shows I care more, since I care enough to take notes. Did people in the early 2000s think saving their contact information in your phone was bad since you couldn’t remember their phone number? Or is this information stuff that I should really care enough about to just remember outright?

× Close

On a more meta level, seeing how “take notes after every conversation” moved from being something bizarre to something that’s become an essential part of my process has helped me think a little bit more critically about the fortune-cookie-style life and startup advice I see on Twitter from tech people. What might work for a busy founder or VC likely doesn’t work for me now, but I sometimes try and think from first principles about how that tip might work if a circumstance in my life changes, because at a later stage the advice might actually be useful (sometimes, though, the advice just makes no sense).

Cons

It’s common among venture capitalists and founders to advise young people that joining a startup is the fastest way to accelerate their career. I can say firsthand that it’s probably true: I was given unparalleled amounts of responsibility, I learned so much, and the people I’ve met along the way have created and shared so many cool opportunities with me. Still, this “join a startup” advice is often prescribed as the “best choice,” given on Twitter with little room for nuance, and there are definitely downsides — or at the very least, complications — to starting one’s career at a startup that are worth mentioning.

Bad Habits

For instance, I think the fact that my first Silicon Valley software engineering experience was at Google was really helpful. It gave me context into what code review is like; what good processes look like; how not only code, but also information and documentation could be organized at scale; what having a good mentor at work was like; and a very liberal sense of “work-life balance” that made me feel OK about taking more ownership of my personal time. In general, processes were concrete and people had high standards for their work and that of others.

At Jupiter, many processes had to be figured out from scratch. On the engineering side, where it was just me and the CTO, we weren’t altogether sure about how to set up a CI pipeline, how to set up a process for communicating between business and tech, how to put together a product roadmap, and how to divvy up work.12

12

We just kind of worked on what we wanted to work on at the start. Once we realized the need for some sort of project management system, we starting playing around with GitHub Issues. That got messy pretty quickly too since we didn’t have a good system for organizing and assigning issues. At the time I left, we had been using Linear for a while.

× Close

We wrote messy code and tested in production maybe a little too much. Our Google Drive had important files all over the place. I kind of began to see the appeal of a PM-type person to help us stay on track.

I feel like without some of the context I picked up from Google, I could’ve picked up some very bad habits.

Work-Life Balance

Two key values of startup culture are “acting like an owner” and being “mission-driven.” No matter how well-structured or intentioned, I think this incentivizes founders and venture capitalists to maintain a sort of startup “hustle” culture, potentially to the detriment of the employee.

For me especially, as the sole employee amongst four founders for a while, I felt like a founder — and felt like I had to live those values — since it was just the five of us. That sometimes made it tough to balance my needs with those of the founders. Especially being so early-career and never having worked at a “Silicon Valley startup” before, I wasn’t sure how to think about my role as an employee versus that of the founders.

I worked a lot, especially at the beginning, and maybe a bit more than was healthy, because of that. I started out doing twelve-hour days13

13

The hours got a little bit more flexible once we moved to remote work. But when we were in the office people were typically at their desks from 9am to 9pm. I’d wake up every morning at around 7am to run in Golden Gate Park.

To be honest, I was mostly OK with this at the beginning since it felt less demanding than my schedule back in college.

× Close

, six days a week, and we didn’t officially move to a five-day workweek until about a month before I left, after we hired a second engineer. I wanted to work on a few personal side projects outside of work, but it was hard to always feel OK doing that (even when the founders said it was OK, which I appreciated) when everyone else I was working with was devoting 100% of their time to the mission14

14

Cf. the data surrounding “unlimited vacation” policies.

× Close

.

Although I’ve definitely improved between when I left college and when I left Jupiter, I still want to get better at protecting personal time and establishing boundaries with my work.

Upside

Related to the above is the low formal equity ownership typically granted to early startup employees. Even if you’re employee #1 and a rockstar, you’ll likely get at most 5% equity (10% and above is considered co-founder level). My compensation package included 1%, which seems par for the course for an early employee. However, for comparison, the four founders (assuming a 10% employee option pool, so 90% remaining) each owned 90% / 4 = 22.5% of the company — 22.5x more.15

15

Technically, we were all on four-year vesting schedules, but the relative proportions were still the same since I joined less than a year after founding. When I hit my cliff and vested 0.25%, the founders had each vested a little more than 5.625% each.

× Close

While we can debate the risk/reward calculus and the unique burdens and responsibilities of a founder that may merit the increased magnitude of equity, in either case, one thing that was hard for me was always maintaining a good sense of those different burdens and responsibilities. I felt I was often stressing myself out and thinking about things that shouldn’t have bothered myself with. Given my comparatively lower economic ownership, I wasn’t always sure how to interpret “act like an owner.” It was especially tough since I joined so early — the company was founded six months before I joined, and they pivoted a few weeks before Demo Day at YC. I wrote the first line of code in the new repository and probably about half the total lines of code by the time I left.

I have limited context on what fairness looks like here, though, and if anything my takeaway from this is that I need to discuss this more with people in the industry — VCs, early employees, founders, etc. For instance, one counterargument to a greater employee equity stake is that not being a “founder” in economic terms gave me the flexibility to leave to go back to school, which I’m doing now. If things go as planned, the company is the founders’ life for the next ten years.

Remote Work

The COVID pandemic in itself also provided some career insights. For instance, I now do not believe remote work is a good fit for early-stage (maybe more specifically, pre-product-market-fit) companies. At Jupiter, before the pandemic, we were accustomed to relying on synchronous communication and a lot of great ideas came out of in-person discussions. It was hard to replicate that serendipity online, even when we used virtual office software. Also, our personal relationships seemed to deteriorate online. I sometimes got really frustrated when people would ask the same question over and over again in Slack instead of using the search bar.

That being said, if I was starting a company now, I would try my best to start in-person. I’m not even sure how willing I’d be to join a remote company as an employee going forward if I didn’t have to.


As a counterpoint to all of the above, maybe my Google experience made me too “soft.” VCs nowadays seem to think that people just go there to coast, and perhaps I got into big tech too early. I still think Google taught me a lot of good things and was a net positive, but it probably also made it harder for me to transition to the chaos and hustle of early-stage. Regardless, my expectations and context have undeniably expanded as a result of my year at Jupiter, and going forward, having both experiences — public company and startup — under my belt will help me adapt better to any situation. At this point, that’s what really matters to me.

III. On Living

Pros

San Francisco

I’m really glad that San Francisco was the first real city I lived in. Part of it might be that I grew up the East Coast (California weather is phenomenal compared to anything you get over there). But San Francisco is also a uniquely beautiful, culturally distinctive city. Our “Paris of the West” has breathtaking views all over the city, over 5,000 acres of open space, and the most photographed bridge in the world. People nerd out about housing policy, bike lanes, and a bunch of other issues16

16

There’s even an unofficial API for our transit system’s stored-value card.

× Close

I never paid attention to in my hometown. Every year on 4/20 tens of thousands of people descend on Golden Gate Park to smoke weed and the police don’t care. On warm summer days, Dolores Park fills with picnickers and it’s like a huge party. At least before the pandemic, I felt a real sense of community.

If I lived in New York or Philadelphia first and stayed close to my East Coast roots, perhaps I wouldn’t know what I was missing out on.17

17

To be fair, I haven’t lived in either of those cities yet, so maybe I don’t know what I’m missing out on over there. Definitely want to give New York a try. But San Francisco seems to be especially unique and special, a city with an outsize cultural influence much greater than its small size might imply.

× Close

The culture of the city influenced me to be especially open to political advocacy, community building, spending time outdoors, and being independent in my thought and action. Even if I leave San Francisco, I hope to bring those unique aspects of the city wherever I go because they made life there interesting and fulfilling.

I don’t know what life in San Francisco will look like after the pandemic, but its undeniable status as a startup hub when I started in 2019 which resulted in me living there was a huge benefit of starting my career at a startup.

Life Skills

The skills weren’t all necessarily startup-specific, but it was exciting to build general life skills early.

Learning Personal Finance, Cooking, etc.

370 words

  1. Getting a regular paycheck has allowed me to developed a deep understanding of personal finance. I have six credit cards, a “Very Good” FICO, and maxed out my Roth IRA for the past three tax years. I enjoy explaining all this stuff to my friends because it helps me understand it better and I also can get referral bonuses.
  2. I’ve learned basic cooking skills. I never realized how hard it was to cut a pineapple with all the little brown fibrous divots on the sides. Slicing a part of my finger off (I was chopping bok choy and not being careful) was a great lesson to be more cautious in the future. I am now comfortable handling raw meat and have become a huge Instant Pot fan. I never had to do any cooking while working at Google (free food), so maybe this is truly a startup-specific skill.
  3. I got much better at blocking out personal time from work time than when I was in school. It was way too easy to let problem sets and projects creep into time that should been mine and mine alone at college. In the working world, there’s no homework. Even though we all worked hard at Jupiter, we did expect to have some degree of a life outside of the company. We rarely worked on Saturdays or outside of working hours. It was easier to make time for myself, and as a result, in 2020, I ran four times as many miles than I did my freshman year.
  4. I’ve developed a “bias for action”. At a small startup, either you do something or it doesn’t get done, and I think this mindset has shaped how I go about my daily life. This change was most apparent to me when I started planning vacations for myself. I remember feeling so cool booking my first flight and thinking, Damn, I can be anywhere in the world in less than 24 hours. The thing is, anybody can. But working at a startup has made it much easier for me to just pull the trigger and take the trip.

Having to figure stuff out all on my own also served as a sort of meta-learning which I think will help me navigate more complex challenges in the future.

Comfort Zone

In general, I was consistently out of my comfort zone. Outside of the early-stage startup scene, one moving aspect of this came from living in San Francisco and passing homeless people every day. It was a very different environment from my hometown, and it was uncomfortable. In high school, volunteering at the suburban food bank seemed really abstract; I didn’t personally know a single person who experienced hunger. Living in San Francisco helped make that more concrete and instilled in me a greater awareness of the urgency and impact of giving back to the community.

Cons

Rather than pure cons, here are some thoughts more in the vein of caveats — things I’m keeping in mind and considering carefully as I decide on a future career in startups and whether I’d want to start one.

Personal Bubble

I’m much deeper into the tech bubble than I ever was. Most of my friends are technical or involved in startups in some way. They consider $1 million to be a “small” seed round. I have four different scooter apps on my phone and I get texts as soon as any of the Founders Fund guys tweets anything provocative.

Especially so because of COVID, it’s apparent that the tech industry is in a whole different world. Google employees’ biggest gripe with the new normal seems to be having to actually pay for food. In general, most of those I knew in tech were barely affected by the pandemic.

I want to stay pragmatic (realistic? cynical?) about my work in technology and stay aware of what’s outside the bubble. There’s a lot of talk about mission-driven this and mission-driven that, and I hope to maintain my ability to see through bullshit as much as I can. If I’m going somewhere to make money, I want to be able to tell myself “I am here for the money” and be OK with that. At the same time, I’ve also begun to think more about actual impact (I’ve actually started to more seriously consider law school18

18

At the current tuition rate, though, law school kind of seems like a huge waste of money. Maybe I’ll do what Kim K is doing.

× Close

, but I have to get my bachelor’s degree first).

Entrepreneurial Aspirations

I used to want to be a successful dropout startup founder. My direct manager at Jupiter, the CTO, was a Stanford dropout who was a year older than me. He had a lot of responsibility. I guess I’ve sort of been in the trenches with him, and for better or for worse, that experience has tempered my immediate entrepreneurial aspirations. I honestly don’t think I’m ready to commit to running a startup of my own right at this moment.

Paul Graham puts it well in his essay “Before the Startup”:

Do not start a startup in college. How to start a startup is just a subset of a bigger problem you’re trying to solve: how to have a good life. And though starting a startup can be part of a good life for a lot of ambitious people, age 20 is not the optimal time to do it. Starting a startup is like a brutally fast depth-first search. Most people should still be searching breadth-first at 20.

You can do things in your early 20s that you can’t do as well before or after, like plunge deeply into projects on a whim and travel super cheaply with no sense of a deadline. For unambitious people, this sort of thing is the dreaded “failure to launch,” but for the ambitious ones it can be an incomparably valuable sort of exploration. If you start a startup at 20 and you’re sufficiently successful, you’ll never get to do it.

I want to do some more breadth-first search first.

Macro Risk

From a more macro perspective, I’m also curious if the current economic tech bubble is ending. It seemed vaguely symbolic when the Creamery — the cafe known as “deal central” of the early 2010s tech boom, right by the Caltrain station where venture capitalists from Palo Alto would alight to meet founders based in the city — closed due to the pandemic in August, and journalists and VCs alike pined for better days. With the growing acceptance of remote work, potential antitrust action, and vanishing tailwinds, I’m not sure if working in tech will be as lucrative as it has been over the past few decades.

Moreover, the tech industry, with its zero marginal cost of replication and high gross margins, seems to be a primary driver of the growing concentration of wealth in a few cities in the hands of a few people. I think as a society we’re increasingly seeing this as a problem, and I wouldn’t be surprised to see sweeping — much more than just antitrust — policy changes proposed to address this in the coming years. Will these changes mean that the same favorable economics that have blessed those who’ve worked in tech over the past few decades will not be the same for me if I end up pursuing a long-term career in the industry? I’m curious how to think about my life and career ambitions in that scenario. I would like to be financially independent as early as possible19

19

Million by 30!

× Close

, but is the path to that goal — or even that goal itself — something I should reevaluate given the macro context?

Still, I enjoy writing code and do it in my free time. If I could be involved in the tech industry in some way for the rest of my life and not worry about money, I would probably be happy. And I probably could. Despite the potential macro headwinds, realistically, as a young person who knows how to code, I get the overall sense that personal financial instability will be the lesser of my worries.

At the same time, if 2020 has shown us anything, it’s that we live in uncertain times, and it feels irresponsible not to at least consider these possibilities.

On the Sepulcher Mountain Trail in Yellowstone National Park

IV. What’s Next?

I want to reiterate this, about everything you read above:

This is all from my personal perspective and reflects my opinions alone.

You probably aren’t hearing the full story about working at a startup from other startup employees. I think it’s a cultural thing. All startups are challenging in one way or another; it’s just that nobody is willing to say anything remotely not positive. The fact that I feel comfortable enough to speak honestly about my experience at Jupiter — both the things I loved and the things I had difficulty with — probably speaks more highly of the founding team of Jupiter than it does of all the other startups that you don’t hear much about — so keep that in mind.

At the end of the day, I want Jupiter to succeed. Equity aligns incentives, and I own a small part of the company. If they do great, I do great.

Every situation is unique, and my experience will never map one-to-one to anyone else’s (and vice versa). I believe in the power of transparency, and I share the details of my experience with the hope that if you’re considering joining a startup, this helps you make a more informed decision, and if you just started working for one, this prepares you for the road ahead. We’re all best off when we’re in the places we can thrive the most.

Doing It Again

Before joining Jupiter, over the summer, I was informally talking to a few other early-stage companies about job opportunities too. I was most excited about Jupiter since they were going through YC right at that time, in the Summer 2019 batch, but I wanted to keep my options open, and regardless, it was exciting to meet other founders.

One pleasant June evening, I was walking over from my place in the Mission to a Thai restaurant in SoMa called Manora’s to talk with an aspiring founder over dinner. He wasn’t there when I arrived, so I stood on the corner and watched the Google shuttles zoom by before he rolled up on a road bike a few minutes later.

I got the pad see ew, and we had a good chat. He explained that he was approaching two years at Facebook and was planning on quitting soon to work on an idea he had for a finance startup.

Over the next few months, we kept in touch on-and-off. I met him in-person just one more time — at a dinky coworking space on Market Street which he had just started renting to begin working on his app — before taking the offer at Jupiter.

A year later, I found out that his company went through YC’s Summer 2020 (S20) batch. At the end of the program, he raised $16 million from Andreessen at a $75 million valuation.

I wonder if he has a Ferrari now.

Just kidding (I think he still rides around on the bike). But still, there were times that I thought to myself that if I had pursued that opportunity a bit more intently I could’ve gotten a front-row seat on a rocket ship, watched him go through YC, and experienced the thrill of hypergrowth.

Yet again, if it was October 2019 and I had the same offer from Jupiter with what I know now, I think I would still take their offer. Each and every one of the difficulties I experienced was also an exceptional learning opportunity. Considering the typical tenure of a software engineer in this town, a two-year tour at Facebook essentially made the S20 founder a tech veteran. At Jupiter, to a certain extent I had the opportunity to be the “tech veteran” and make the big engineering decisions. Joining a team with relatively little prior startup experience, we had an opportunity to all learn together. That type of learning was exactly what I was looking for when I left school, and for that, Jupiter was an excellent fit.

In any case, the whole experience has shown me that Silicon Valley works in mysterious ways, and with what I’ve learned, if it’s still something I want to do, it might even be me, a few summers down the line, going through YC.

Leaving Jupiter

I’m planning on returning to college in fall 2021, and I wanted some time to focus on transfer prerequisite courses and transfer applications that were due in early 2021.

As a result, my last day as a full-time employee at Jupiter was this past October. I continued working part-time, on a contract basis, for the next four weeks into mid-November — hanging out in Slack, answering questions, fixing bugs, and writing documentation for knowledge transfer.

The specific date I left was suspiciously close to my vesting cliff, and to be honest, that was part of my decision too.20

20

If you’re a startup employee thinking about leaving and you’re close to your cliff, you shouldn’t feel bad about waiting it out and pushing through the date. In my opinion, it was part of the deal you and the founders agreed to and you’ve earned your equity by staying for so close to a year.

The main purposes of cliff vesting in startups are to mitigate the risk of a bad hire (in which case they’d be fired early) and to mitigate the risk of people leaving really early, before they’re onboarded enough to meaningfully contribute. If it’s getting close to month twelve and you’re performing fine, neither of those risks apply to you.

× Close

Still, the timing worked out — I was about ready to start figuring out my next chapter by then.

For over a year, my mind was always racing with work, and as I reduced my full-time responsibilities, it was relaxing to be thinking about different things for a change. I left the city, rented a house in Montana with my brother and a friend from Michigan, and bummed around in the mountains for a bit.

San Francisco feels very familiar to me by now, and it was refreshing to live somewhere new and figure out new routines again, from reaccustoming myself to being dependent on a car to shooting a gun for the first time. Exploring new places and trying new things continues to show me how much more there is to life than the bubbles I sometimes get too comfortable in.

Thank You

After Montana, I returned to the Philadelphia suburbs to spend time with family for the holidays. I’ve been revisiting my earlier work more frequently as I’ve been putting together my college transfer applications.

It’s May now, and I land in San Francisco in about thirty minutes.

“The Internship”, May 11th, 2019

From my first steps in San Francisco to my last flight — for now — out of SFO, it’s hard to believe that it’s been almost two full years. I am especially grateful for this last one.

To the founders of Jupiter who let me ride along — Chad, Anuraag, Anna, and Will — thank you.


  1. As of this publication, fifteen months later, the S&P 500 is up about 30%.

  2. To be fair, Rollup is growing fast and isn’t really an “also-ran,” but at the time, Webpack was huge in my mind.

  3. Paul Graham’s “Beating the Averages” mentions a similar phenomenon with Lisp.

  4. One unintended benefit is that it’s made me a lot more comfortable as an interviewee; I can kind of understand what people are looking for now.

  5. For what it’s worth, I’ve gained a lot of respect for SQL’s capabilities as a data retrieval language. I learned so many new keywords I’d never used before.

  6. A few simple things that improved communication: writing more approachable documentation, coming up with good analogies for technical concepts, and generally being patient (in both directions).

  7. A new-grad friend, who recently started leading a team, shared this with me after reading this:

    “Learning to lead has been huge for me over the past year. I think understanding that in some ways, everyone is doing everything for the first time is a helpful (albeit maybe naive) perspective.”

  8. This was something that came up in my exit interview. I appreciate that the founders took the time to give me candid feedback, and the things they shared with me are things that I will definitely keep in mind to be effective in future jobs. I had a chance to share some of my own feedback during the interview, too. I’m not sure how common exit interviews are in general, but it was very valuable for both sides. I think all startups should do them.

  9. $100,000. Compared on an annualized basis, this was actually less than what I was getting at Google as an intern — which should give you an idea of big tech’s compensation philosophy and how crazy the market for software engineers was at the time (and probably still is).

  10. I feel like the term “networking” has sort of a bad rap so I sometimes hesitate to use it, but it’s probably the most precise term for what I’m describing here.

  11. Some people are very opposed to the whole “people on spreadsheets” idea. I like to think that my spreadsheet is just another tool, like Google or Wikipedia or the Contacts app: it’s hard to remember URLs and facts and phone numbers, and it’s sometimes similarly hard to remember the specific details of calls and meetings. I also think it shows I care more, since I care enough to take notes. Did people in the early 2000s think saving their contact information in your phone was bad since you couldn’t remember their phone number? Or is this information stuff that I should really care enough about to just remember outright?

  12. We just kind of worked on what we wanted to work on at the start. Once we realized the need for some sort of project management system, we starting playing around with GitHub Issues. That got messy pretty quickly too since we didn’t have a good system for organizing and assigning issues. At the time I left, we had been using Linear for a while.

  13. The hours got a little bit more flexible once we moved to remote work. But when we were in the office people were typically at their desks from 9am to 9pm. I’d wake up every morning at around 7am to run in Golden Gate Park.

    To be honest, I was mostly OK with this at the beginning since it felt less demanding than my schedule back in college.

  14. Cf. the data surrounding “unlimited vacation” policies.

  15. Technically, we were all on four-year vesting schedules, but the relative proportions were still the same since I joined less than a year after founding. When I hit my cliff and vested 0.25%, the founders had each vested a little more than 5.625% each.

  16. There’s even an unofficial API for our transit system’s stored-value card.

  17. To be fair, I haven’t lived in either of those cities yet, so maybe I don’t know what I’m missing out on over there. Definitely want to give New York a try. But San Francisco seems to be especially unique and special, a city with an outsize cultural influence much greater than its small size might imply.

  18. At the current tuition rate, though, law school kind of seems like a huge waste of money. Maybe I’ll do what Kim K is doing.

  19. Million by 30!

  20. If you’re a startup employee thinking about leaving and you’re close to your cliff, you shouldn’t feel bad about waiting it out and pushing through the date. In my opinion, it was part of the deal you and the founders agreed to and you’ve earned your equity by staying for so close to a year.

    The main purposes of cliff vesting in startups are to mitigate the risk of a bad hire (in which case they’d be fired early) and to mitigate the risk of people leaving really early, before they’re onboarded enough to meaningfully contribute. If it’s getting close to month twelve and you’re performing fine, neither of those risks apply to you.

Leave a Comment  ·  Back to Home

Get Natecation Updates

Add your email to get earliest access to new posts.

Recommended Posts

If you liked "Startup Year: A Review of My Year at Jupiter (YC S19)", you might also like:

July 29th, 2020 · Long Essay

Carpe Annum

My 2020 Gap Year Playbook (for Startups and Tech)

As more and more colleges go online, I've started to get more questions from students about taking a gap year. If you're looking to take a gap year soon, willing to learn how to code, and interested in pursuing a career in tech, here's what I recommend.

February 25th, 2020 · Long Essay

Why I Dropped Out

From My Freshman Year to Google to a Startup

My 2019 in review: freshman year, leaving college, a summer at Google, and joining a startup. Some notes on 2020 and beyond.

N

Natecation

by Nathan H. Leung

Hi, I'm Nate. I'm originally from a suburb of Philadelphia, PA. I taught myself how to code before college, and I entered the University of Michigan in September 2018, where I discovered interests in math, economics, and personal writing.

Looking to get more real-world experience after my freshman year of college, I left school in May 2019. Over the summer, I interned at Google. That October, shortly after the founders went through YC's Summer '19 batch, I joined Jupiter as their first hire. Outside of establishing a strong engineering organization and codifying the company's guiding development philosophies and practices, I spent a year building end-to-end systems with TypeScript, React, Kotlin, gRPC, and Kubernetes.

During my gap year at Jupiter, I also established California residency by living, voting, and paying taxes in San Francisco. Because of my new domicile, I intend to transfer to a UC school in Fall 2021 with an anticipated graduation date of May 2023, and I officially phased out of my role at Jupiter in November 2020 in preparation for that. As of this writing, I'm spending the next few months writing transfer applications and completing transfer prerequisite courses at the College of the Canyons. I'm looking forward to being a student again in September 2021.

I started this little blog to share some of my experiences — I hope you enjoy.