Blogging about technology, government, and life, for fun and (no) profit.

Final Reflections For “Entrepreneurship in Silicon Valley”

These are some of the final reflections I took away from my “Entrepreneurship In Silicon Valley” January term course in 2014.

The most common thing that I think we heard was that in entrepreneurship (especially entrepreneurship in technology), ideas don’t mean anything. It’s easy to have good ideas. Everyone has good ideas for a business. The hard part, the part that differentiates a successful startup, is the execution of the idea. It’s worthless to keep your ideas secret, or to make people sign non-disclosure agreements. Even if somebody “steals” your idea, chances are, they won’t succeed in making it a successful business. In fact, someone has already probably had your idea, but wouldn’t or couldn’t execute it. That’s what distinguishes an entrepreneur.

I learned a great deal about the financial side of founding a startup. As a tech person, I’ve never paid particularly close attention to the financial side, but obviously that process is just as critical for success. We heard some interesting opinions about the system of venture capitalists in Silicon Valley. The part that surprised me was that some people are very wary of accepting venture capital. I had assumed that accepting funding was a no-brainer; why would I turn down money that might mean the difference between the success and the failure of my company? However, several people we met with suggested that in many cases, the right move is to turn down the capital. Usually a funding round from a venture capitalist means allowing the VC firm some control over the company. This might be in the form of a controlling interest of shares, or a seat on the board of directors. Any time you’re forced to give up some control, you incur the risk of conflicts. Perhaps I have a different vision for the future of my product, one that may not generate quite as much revenue, or might take longer to realize. Of course, there’s always the chance that I could be wrong about the way I want to do things.

The most interesting non-technical meeting we had was with Ann Winblad, of Hummer Winblad Venture Partners. She has very strong opinions on the way that startups should run, and the direction that technology is going. Most interesting, however, was her approach to the risk in funding technology startups, especially in comparison to inventors and scientists like Phillip Alvelda. Ann spent a lot of time discussing what startups should be doing in terms of building a product; she strongly believes that most future development will be based on building blocks of other software. One of her examples was the growth of the Maven software repository. Maven allows developers to include libraries and functionalities that have been written by other people in very simple and robust way. Of course, not reinventing the wheel is a very common theme in computer science. However, Ann really takes it a step further by arguing that a startup really shouldn’t be writing that much code, but should rather be connected existing software together.

I didn’t realize it at the time, but later, when we met Phillip Alvelda, Ann Winblad’s approach to startups really started to make sense. She’s a venture capitalist: her goal is to make money. To do that, she needs to invest in low-risk ventures that have a high potential payoff. The way to minimize risk in development is by using components that have already been developed and are known to work well. Designing software that builds on already existing software components takes less time and programmer skill (which in turn means lower risk) and consequently has a better chance of being successful and repaying the investment. However, this approach isn’t a good way to truly innovate. As Phillip pointed out, true innovation is building something that’s completely new. It’s generally pretty difficult to do something completely new without building large parts of it from the ground up. This is much riskier. There are many more places for things to go wrong. Phillip used the example of Elon Musk’s proposed hyperloop transportation system. The hyperloop isn’t software, but the analogy applies just as well to software development. Under Ann’s philosophy, a project like the hyperloop wouldn’t get funded. It depends on innovation and theoretical research, instead of just leveraging existing technology. If everyone thinks like Ann does, there won’t be much real improvement in technology!

One of things we heard from PerCognate really stuck with me: the so-called “four hats of entrepreneurship”. This was fascinating, because it really made me start thinking about the stuff I’m good at, the stuff I’m not so good at, and what kinds of people I want to work with. Their four hats are the builder or hacker, the designer, the visionary, and the closer or hustler. I’ve realized that I’m certainly a builder. I like to write code, and make technology work. I’m not very good at sales or marketing, so I’m definitely not a closer. I’m not a very good designer, so I wouldn’t be a good choice to design the user experience. I might be somewhat of a visionary, but I generally prefer to let others do that sort of planning. If I ever decide to start a company, I’m definitely going to be much more aware of the balance of skills, and whether all of the important roles are filled.

Learning about the way that entrepreneurs pick a market and position themselves was also really interesting. For example, PerCognate’s e-discovery technology could easily be applied to many areas. However, they’re running out of “runway”: they aren’t yet making any revenue, and are running out of money to keep working. What this means is that they’ve made a very conscious decision to target their work toward lawyers and the evidence discovery process. They’ve decided to do that one thing really well, because lawyers will be able to pay them well. If they spent time using their technology to try and solve other problems, that would distract them. They would end up as a company that was okay at doing several things, rather than really good at doing one thing. Capsule also discovered this. Their product could be applicable to many different use cases, but they picked one market (weddings) and went after it directly.

This is something that intrigued me, because I could easily see myself falling in to the trap of trying to do everything. If I had invented Capsule, I’m not sure I would have been able to let go of the original vision of catering to any event or trip, and just focus on one market. In fact, I’m not sure the idea would have even occurred to me. Talking to the founders of PerCognate and to Cyrus from Capsule really made me question the assumptions I had about how to market a product. If you’d asked me before this trip, I probably would have said that the best way to market something like Capsule would be to target as many use cases as possible, rather than focusing on the core money-maker.

This trip really drove home the point that in the business world, connections are incredibly important. I’ve never liked the phrase “it’s not what you know, but who you know”, but it a lot of ways it really does apply. Obviously, most of the people we met were very intelligent and capable in their respective fields, but over and over again, it’s the connections they had and people they knew that got them to where they are. That means that it’s worth cultivating those connections. I collected a number of business cards as well as making some online connections, and I strongly believe that some of those connections that I gathered on this trip will help me some time in the future.

I definitely had an interest in entrepreneurship and startups before going on this trip, but after visiting so many different companies doing so many different things, my interest in working for or even founding a company has really been solidified. I think this might be the most important thing I took from the course: I want to work in a place where everyone is able to contribute meaningfully, where I can constantly be learning new technologies and ideas, where I can have true ownership of work I’m doing, and where I’m truly passionate about the products I’m developing. I think that place is a startup. Seeing the passion that people like Eric Bieller (of Sqwiggle) have for the work they’re doing really spoke to me. I’d rather not get stuck in an environment where I don’t have a choice about the tools I use. I don’t want to work in a place where I have to dress up in order to fit in.

While hearing about all the “failures” that many entrepreneurs faced was definitely interesting and often inspiring, the idea of starting my own company is still terrifying. It would certainly be incredibly rewarding, but how do I balance the risk of needing to have a job to support myself with that thrill? I think at this point in my life, I wouldn’t be comfortable founding a company as a full time venture. I’d much rather work as an employee at a small but still somewhat established startup. Perhaps that’ll give me the experience I need to actually take the leap of starting my own!

Fullscreen, I Really Wanted To Hate You…

On Friday, we visited Fullscreen, a “global YouTube network for creators and brands.” We were connected to them because they have a number of Luther alumni working there, but the only thing Brad told me about Fullscreen before the visit was that they create YouTube video ads. Obviously, I’m going to be a little predisposed to hate them…

What I found was actually something pretty different. It’s true that Fullscreen is definitely involved in video advertising on YouTube, but they do a lot more.

Their primary business model is that they’re hired or contracted to manage content for thousands of of channels. Their primary goal is to increase viewership and revenue for the content creators so that the creators can actually make a full time living from their content.

Most of Fullscreen’s clients were people (“content creators”) who were already doing content on YouTube (the content ranges from music videos to teenage girls complaining about their lives). Fullscreen works with them to monetize.

What Fullscreen does depends on the needs and wants of each client. In some cases, they simply connect their clients with advertisers. In other cases, they help their clients design the entire ad campaign.

They’ve got between 26,000 and 30,000 channels under the Fullscreen umbrella. Most of their channels had at least 1000 subscribers before they signed a deal with Fullscreen. Some of their clients include Lindsey Stirling (“dubstep violinist”), lonelygirl15, Jenna Marbles, and Philip DeFranco (phillyd).

A lot of these people now make enough money from just their YouTube content to live on full-time (in the case of Jenna Marbles, much more than enough to live on full-time).

Fullscreen is also aggressivly trying to expand their product and service lineup. They’re working on tools that help the creators create content, and streamlining the process so that clients can take more of a self-service approach to monetizing their content. They’re also working on networking between creators, who can share ideas and tips via some forum-like software.

One of the questions I had going in to Fullscreen was how their business was affected by ad-blocking plugins that are very effective at suppressing YouTube advertisments. It sounds like they aren’t particularly worried: less than 5% of the world audience uses ad-blockers, and they don’t see that increasing tremendously, as ad-blockers are generally only installed by technically-saavy users. (I asked one of the developers at Fullscreen whether he personally used an ad-blocker: he didn’t seem to have any dilemma in using one!)

So, walking away from Fullscreen, I realized that I can’t fault them for their business. The people who make YouTube interesting deserve to make money, and Fullscreen has gotten pretty good at helping them do so in a fair way.

Creating Something Out Of Nothing

Yesterday, we visited Coloft, a co-working space in Santa Monica. We met with one of the cofounders of Coloft, Avesta Rasouli. Avesta is a serial entrepreneur who had some really interesting thoughts and advice on startups, especially in Silicon Valley. I’m just going to talk about of few of the points he made that stood out at me, in no particular order.

Picking a partner

Avesta really stressed the importance of picking (a) partner(s) who have/has complementary skill sets. If you’re really good at coding, it’s pretty important to have someone who can manage a business and make sales (this really echoes the concept of the “4 hats of entrepreneurship” that we heard from the guys at PerCognate).

According to Avestar, once you have people with complementary skill sets to start a venture with, you have to be willing to stay out of each other’s business. You’ve got to be willing to trust that the other person (who has skills you don’t) knows what their doing. If you try to micromanage everything, the venture is probably going to fail.

Solve other people’s problems, not yours

This one was interesting for me: Avestar said that it’s often important to resist the urge to start something that solves your problems. You need to make sure their’s a market for what you’re doing (which involves, obviously, talking to people). If you talk to four people, and three of them have the same problem, that might be a good market to go after. If you have a problem you can solve with a product, don’t spend a ton of time assuming that other’s have that problem: you have to make sure you aren’t wasting your time.

I think this is one that I could struggle with. I’m obviously not going out and starting a business right now, but when I start projects, they’re generally aimed to solve a need I have – most of what I do wouldn’t be very useful to other people.

Ideas don’t mean shit

We’ve heard this one a number of times: ideas are a dime a dozen. Everyone has good ideas. It’s the execution that takes skill and effort. On a related note, Avestar was pretty dismissive of non-disclosure agreements. He won’t sign them in exchange for listening to an idea, and he said that most people in Silicon Valley won’t either. This connects to the fact that ideas are cheap: if you have a good idea, telling everyone probably won’t make a difference. In fact, someone has already probably had your idea, but wouldn’t or couldn’t execute it. That’s what distinguishes an entrepreneur.

The loudest people are often not the smartest

This is pretty self-explanatory. It’s easy to get caught up in how many people follow you on Twitter, or to talk up your ideas and convince people. But the smartest people are the ones who don’t waste time with stuff like that: they’re actually out building stuff.

Sportvision and Fenwick & West

We’ve been on a train now for about nine hours, so you can probably imagine that I’m about ready to be done. We should be arriving in Santa Monica in a few hours.

Yesterday was our last day in Menlo Park/Mountain View, and we visited two companies.

The first was Sportvision. Sportvision is one of those companies that you’ve probably never heard of, but you’ve definitely seen their work. Their technology is responsible for all kinds of “virtual reality” images on sports broadcast. Their most famous product is the yellow “first and ten” line that you see superimposed on the field during pro and college football games. They also do graphics and visualizations for hockey, baseball, soccer, basketball, auto racing, and even sometimes skiing. Their biggest area of growth, however, is in data collection. They are putting sensors on players or using cameras to track players, and packaging that data in various different ways to sell to various clients. (The simplest client would be a football team buying the movement and position data for all their players during a game so they can later go back and replay and analyze every play.)

Technologically, this was one of the most sophisticated business we’ve seen. All of their work is based on computer vision and motion tracking, so most of their engineers aren’t just computer scientists – they have to have tons of physics and math.

One of the coolest things we saw was a piece of software that can track a baseball (moving at 80-90MPH), and based on it’s trajectory, calculate it’s spin rate. This is not for the faint at heart.

Even the simple “first and ten” line isn’t nearly as simple as it seems: how do you teach a computer to superimpose the line on the field but below any people or objects on the field? How do you automatically compensate for changing lighting conditions and variable field colors?

We got a great introduction to their business and technology from Michael King, who introduced himself as their director of football, but obviously has a pretty comprehensive view of the whole company.

The other visit we made was to the law firm Fenwick & West, in downtown Mountain View. F & W is a fairly large law firm, but we were meeting with specifically the patent litigation team. They have a ton of large clients, including Twitter, Amazon, and Google.

We had a chance to talk with five lawyers from the patent litigation team, who gave us an extremely well prepared introduction to the basics of patent law, and especially to the problems around patent trolls (TODO add link).

The problems with patent trolls have been well documented, so I won’t spend a ton of time talking about them, but I did want to highlight what I thought was the most interesting point made by the F & W team.

The most common proposal for eliminating the problem of patent trolls involves legislation that bans so-called “non-practicing entities” from owning or legislating based on patents. NPEs are companies who own patents but aren’t involved in a business. Obviously, this describes patent trolls: they exist only to purchase patents and sue other companies based on those patents. They don’t provide any other product or service.

The problem with such proposed legislation is that, technically, universities are non-practicing entities as well. Nobody wants to argue that what research universities are doing isn’t valuable and shouldn’t be afforded patent protection. How do you draw the line, though? It’s difficult or impossible to draw the line in such a way that it excludes patent trolls while still fairly including universities and other non-profit but still patent-producing institutions.

I’m certainly not a lawyer, and I have no idea where to draw the line, but it was incredibly interesting to hear about some of the issues directly from lawyers who have to deal with patent trolls every day!

Using GVM (Go version manager) To Install Golang

I’m starting a project in Go. The first step was obviously to install Go, but after a half hour wrestling with an old and broken installation and a out-of-date PPA, I got fed up.

Then I found gvm. If you’re familiar with rvm for Ruby, I think you’re going to like gvm.

Setup on Ubuntu is pretty simple:

$ bash < <(curl -s

Then you’ll need to restart your terminal.

Since the install script seems to assume you’re using bash, if you’re using a shell other than bash, you’ll need to add this to your config file (e.g .zshrc, etc):

[[ -s "$HOME/.gvm/scripts/gvm" ]] && source "$HOME/.gvm/scripts/gvm"

In my case, I wanted the latest version of Go (1.2 as of this writing), so I ran:

$ gvm install go1.2
$ gvm use go1.2 --default

The --default flag sets that particular version of go as the default when you open a new terminal session.

It was fun to see that gvm was written by the devs at MoovWeb. When I visited them a few days ago I was interested that they were using Go for their backend – clearly they’re pretty invested!

When “Just Trust Us” Isn’t Good Enough

I originally wrote this paper for an interdisciplinary Ethics and Technology course, during the fall of 2013.

Net neutrality is a term frequently encountered when discussing access to information on the Internet. While it is generally intended to refer to a guarantee of certain qualities of access to the global Internet, it has been specifically defined in many different ways. I intend to ethically analyze the various aspects of net neutrality, including several common definitions. I will contend that a definition of net neutrality that allows for network traffic prioritization and shaping is ethical, but only for the purposes of managing and improving the network. Non-net neutrality is unethical if it is done for financial gain (e.g. “pay extra for the premium sites package”), if it is not transparent to the end user, or is done to discriminate against a particular class of users.

Read on →

Assimilation into the GooglePlex

Yesterday, we visited Google headquarters, in Mountain View, California. It was certainly an impressive experience, but I was left with some mixed feelings about the experience.

First of all, Google is obviously doing some really cool stuff. My computer and Internet experience would be very different without Google. I use Gmail, Calendar, Maps, Search, Chrome, and other products every day. Large parts of the Internet are financed by Google-served ads.

Google also has the deep pockets and corporate will to perform groundbreaking research. Their self-driving car is incredible, and Google Glass is pretty revolutionary (whatever you think about the benefit to humanity).

On the other hand, I found myself fairly uncomfortable at many points during our visit yesterday. One of the people we met was Craig Cornelius, a former Luther alumna who works on the internationalization team (i18n) at Google. One of the first things Craig said when he starting talking about Google was, “how much do you pay Google to use our services?” It’s a fairly innocuous question, but he clearly expected us to answer that we paid nothing.

When someone said that we paid with our personal information, he sort of acknowledged the point, but then immediately moved on to convincing us that because we don’t “cut Google a check,” we really aren’t paying anything. That unwillingness to acknowledge Google’s core money-maker (our data) leaves a bad taste in my mouth.

I give them that data willingly, in exchange for some very useful services and products. But if Craig’s perspective is common at Google, I have to feel like they’re simply trying to forget (or convince the rest of us to forget) that we are paying them.

I did enjoying seeing the campus with Charles Banta (a recent Luther grad, and someone I’ve worked with at the Luther Tech Help Desk), and getting to see some of the perks that Googlers get. Google clearly views time that employees are not on campus as wasted time. And when you can eat, drink, sleep, get a haircut, get your car repaired, do your laundry, see a doctor, and work out on campus, what’s the point of going home?

Again, it would be fun to work at Google, and I know that Charles is really enjoying it. But I can’t help but think that your life would quickly become Google-centric. It’s one thing to love your job and want to work a lot (a situation I hope to find myself in when I enter the job market), but the word “assimilation” kept popping in to my head during the entire visit. (For a fictional take on this, I highly recommend reading The Circle by Dave Eggers. Just replace “The Circle” with “Google” and you’ll have a good idea how I felt visting Google.)

Two of the technical talks we heard, however, were fantastic. Craig Cornelius gave a lecture on the challenges of internationalization of Google products. It didn’t sound that interesting at first, but his team has to solve so incredibly difficult projects. How you design projects so that they work equally well with left-to-right AND right-to-left languages? How do you handle keyboard input in a language that has thousands of “letters”? How do you deal with one of the hundreds of ways a single date and time can be expressed?

We also had a chance to hear a little bit from Brian, an engineer on Google’s self-driving car team (unfortunately, I didn’t manage to catch his last name). We didn’t have a chance to get as technical as I would have liked due to time issues, but he had a great introduction to the technology behind the “perception suite” that’s used to make the car “see”, and some interesting scenarios for possible societal changes that might be caused by the introduction of autonomous cars.

That’s not how you Constitution, Mr. Obama

President Obama gave a speech today. The topic was the NSA; he was addressing the recommendations given by the panel of experts he appointed to review the actions taken by the spy agency.

I’m fairly disappointed in Obama’s response. He failed to address or paid only lip service to some pretty important issues.

Advocate for the defense in the FISA court

One issue of potential constitutional significance is that cases heard in the secret FISA court are completely one sided: the judges hear only the government’s side. The review panel recommended that an advocate be appointed to argue against the government (for the defendants who don’t even know that they’re being considered for surveillance).

In his speech today, President Obama said that he will create a “panel of advocates” to represent privacy concerns in significant cases.

From The New York Times:

A senior official said the advocates would be called on only in cases presenting novel and important privacy law issues. The panel would apparently not have the authority to monitor the court’s caseload and independently decide when a case warranted its presence.

So basically, the “advocate panel” doesn’t have any power whatsoever. How can the right to counsel be respected if no one has the power to argue against the government? Who decides if a case “[presents] novel and important privacy law issues”? If the FISA court or the prosecution (i.e. the government/NSA) is responsible for making such a determination, that’s hardly better than not having an advocate panel at all!

Secrecy of National Security Letters

The other disappointment for me was President Obama’s failure to implement the requirement that a judge sign off before a national security letter (NSL) can be served. He did change the policy on the mandatory gag orders that stop recipients of NSLs from discussing them or even mentioning that they’d received one. The gag orders will now have a fixed expiration date. It’s not perfect: he didn’t say how long that fixed expiration time period will be (if it’s longer than a few months, what’s the point?) and the weasel words, “in most cases,” make me uncomfortable.

There are a lot of other specific recommendations by the review panel that I think Mr. Obama stopped short of adequately addressing. I’d recommend reading the NYTimes piece – it does a great job of breaking down his response to each of the recommendations.