Frankly Speaking, 5/17/19 -- "Work" as a VC
A weekly(-ish) newsletter on random thoughts in tech and research. I am an investor at Dell Technologies Capital and a recovering academic. I am interested in security, blockchain, and devops.
WEEKLY TECH THOUGHT
In lieu of my normal blog post, I will discuss “work” as a venture capitalist. I was inspired by my friend Jean Yang, who wrote an interesting post about her “evolution on work” where she described her transition from PhD to professor to startup founder.
In the past, I blogged about what “work” is for a PhD student. As a computer scientist and former PhD student, I’m trained to think of work as producing tangible output like writing a computer program or writing a paper. However, most work as a VC can be boiled down to three main categories: meetings, emails, and creative time.
Meetings
This takes up 80–90 percent of my week.
Startups: This is pretty self-explanatory. We listen to pitches from companies needing to raise money as well as do diligence (understand product, sales, market, etc.) on companies we’re seriously considering.
Board Meetings: My board meeting docket is not that full yet. We take large ownership stakes in companies, so we also take board seats. These meetings happen 3–4 times a year and are a good way for us to get a pulse check on the company.
Portfolio Support: We are active board members and try to add value to our companies outside of our capital investment. I introduce our companies to customers, advisors, potential hires, etc. I also help them with brainstorming, go-to-market, and technical strategy (because of my PhD).
Networking: In order to provide effective portfolio support, we need to have a strong network to make connections and better understand market trends. I spend a good amount of time networking with executives at companies as well as subject experts like professors and engineers. Ideally, if a company wants to talk to someone about a particular topic, we can easily create that connection.
Partner Sync-up: Most of the work I described above is pretty individualistic, so communication and coordination are key. We meet to sync up on both Mondays and Fridays to discuss deals and action items. We have to force this communication because we, otherwise, don’t spend too much time as a group, and it’s easy to work in isolation. It’s important to know how others think about deals as well as receive feedback and help on companies, both existing and future. We also use these meetings to discuss and work on administrative and logistical matters for the firm.
Emails
When I started in VC, one key piece of advice is that “you have to be good at email.” It is our main form of communication outside of calls. Calls are usually reserved for urgent or in-depth matters, but emails are a good way to get in touch with VCs or send a non-urgent note over for us to read. On a slow day, I write about 20 emails, but I can write and respond up to 100 emails a day.
Creative Time
As a PhD student, I had way more creative time, and I think this time is very underappreciated. As a VC, I have to force myself to have this time because otherwise, I would be dragged into another meeting.
Reading: I read a fair amount of blog posts and market research, but I also read Hacker News and research papers. I also find that reading fiction or non-fiction sparks creativity. I keep my scope of reading broad because I think it’s important to read and consider various perspectives.
Blog posts: The PhD forced writing upon me, but I find that many times, writing helps me find gaps in my reasoning and helps me focus and prioritize my thoughts. I find blog posts are a nice, quick way to do writing because they tend to be short and easy to find later on.
In conclusion, most of my time is spent in meetings, but I really enjoy interacting with entrepreneurs and learning about new topics. I have learned that having more creative time allows me to become more efficient in meetings and provide better feedback as I am able to ask more relevant questions.
Finally, at the end of the day, VCs provide a service to entrepreneurs. We work for them, not the other way around!
WEEKLY TWEET
WEEKLY FRANK THOUGHT
This is the last on my "rant" on biases in data, which stem from my frustrations around people using data as absolute fact without any rigor. I talked about a variety of sources of bias. The last main source of bias is evaluation bias.
Evaluation bias happens when the data used to benchmark or evaluate the machine learning model does not represent the target population. Training data is used to create models, but benchmarks that don't properly represent the target population can bias the model into working poorly for certain subsets of the population. For example, commercial facial recognition algorithms didn't work well on dark-skinned females.
I covered five main sources of bias. Again, this is based on the following paper, and shows the need to have a rigorous treatment of data and be skeptical until we better understand the context around the methodology that generated the data.
FUN NEWS & LINKS
#securityvclogic
"How much do we pay Gartner to create a new security category for our struggling security portco?"
#research
I/O is faster than CPU -- Let's partition resources and eliminate (most) OS abstractions.
This is a pretty interesting paper, and I plan to talk about it more in the upcoming weeks. In CS, there's always been the fundamental assumption that I/O operations are much slower than CPU operations. However, with programmable NICs and faster non-volatile memory in servers, I/O operations are almost just as fast, so we need to rethink about abstractions in the OS, which have high overhead.