Latest Research
Our Approach to Ethical Investing

Our Approach to Ethical Investing

Invest in high-integrity founders making products people love and that our LPs will be proud to back.

We talk a lot about ethics at Loup Ventures. Debate might be a better term. The themes we invest in — AI, robotics, and neurotech, all of which sit on the frontier of innovation — have the potential to profoundly impact humanity, and that potential is not lost on us.

However, establishing a thoughtful ethical code for investing has proven difficult. The purpose of this writing is to detail some of the debate we’ve had and where we’ve landed with how we approach investments from an ethical standpoint. We hope it encourages others involved in frontier tech to share how they think about ethics, and we can all sharpen one another.

Things We Have to Accept (Axioms)

Before agreeing on a high-level code, there are several propositions that we believe need to be logically accepted as individuals and as investors in frontier technology. We consider these bolded statements to be axiomatic.

  • Judgements of “good” and “evil” are context dependent and subjective to individual and/or collective perspective. To the extent there are obvious cases of “evil,” for example, technologies designed with a specific intent of distributing child pornography, we obviously won’t invest. Fortunately, few companies are formed with such specific evil intent. I doubt those that are will have much success in finding venture funding. Companies developing frontier tech that triggers an ethical consideration generally fall in the subjective spectrum of grey between good and evil. It is those companies where rational debate about the tradeoffs of each side, reasonably expected use cases and outcomes, and the ethics of leadership must guide an informed investment decision.
  • No ethical code for the purposes of influencing investment decision will be effective at limiting the unpredictable. How technology will specifically impact the world long term, which we define as five years or more, is impossible to forecast with much accuracy. The most significant negative impacts from any technology will likely come from black swan-style events. It’s rarely the predictable events that cause the most damage because, by definition, you can prepare for those. It’s the unpredictable events that catch you off guard that cause the most harm. For example, no person could have reasonably predicted in 2010 that Facebook would be a tool for influencing international elections as an ethical reason not to invest.
  • Technology adopted at relative scale is a signal that a significant number of people derive positive impact from it; however, any technology adopted at scale will also inevitably impact some reasonably significant number of people in negative ways. Technology is meant to disrupt. Scale magnifies the disruptive effect of technology on humanity (see the Facebook example above). The change that comes from disruption will not be welcomed by all given that good and evil are subjective.

To consider a future example, robots that make factory work more efficient are still nascent today, but at scale, they may cause the elimination of millions of jobs. That will have a real, negative impact on those whose jobs are replaced. However, automated production should result in cheaper products and services, which many would see as a good thing for humankind. You can logically justify many transformative technologies as good for humankind, but you cannot prevent, and therefore must be comfortable with, some significant groups of people being negatively impacted despite the greater good.

  • An ethical code can tell us what not to invest in, but not what to invest in. It’s our fiduciary responsibility to act in the best interests of our limited partners (LPs). To that end, we absolutely want to avoid investing in ethically objectionable companies. We also don’t want to fool ourselves into investing in a company simply because they fit positively into some ethical framework but lack critical elements like team, product, or market. Thus, considered alone, any ethical framework for the purposes of making investment decisions can be a tool for negative selection only, not positive selection.
  • Any ethical code is only as effective as the belief in it of those who are supposed to follow it. In the case of building a venture firm that follows a defined ethical policy, this relates to hiring the right people that share the same values that shape our policy.
  • Establishing Key Firm Values

    Now that we’ve established some things we view as truths, we have to establish some necessities of the code for the purposes of our firm: it should be easy to follow and remember, scalable for a growing organization, and adhere to values we hold as important. These are all especially important given the final axiom that any ethical code is only as effective as the belief in it of those that are supposed to follow it.

    In addition to our firm’s core values, our most important values relevant to an ethical code include:

    • We respect the sanctity of individual choice. As a firm, we want to invest in things the world needs; however, we also respect that “what the world needs” depends on individual opinion. We do not have an interest in acting as moral police and judging those varying opinions based on our own subjective opinions.
    • That said, we are a conduit for the desires and values of our LPs. Our LPs have demonstrated trust in our judgment by investing in our fund, but it’s important to us that our actions reflect the desires and values of our LPs combined with our analysis of the opportunity. We can never forget that we’re investing their money and therefore should put ourselves in their shoes when making an investment. We work for our LPs and don’t want to lose their trust by investing in something they would find objectionable. We should always be prepared to explain to our LPs why we made an investment decision inclusive of both ethical and operational considerations.
    • Early-stage investments are investments in people first, second, and third, then product. The most important thing we have to assess in any investment opportunity is always the founding team. Fundamentally, the founding team and the culture they build shape how the organization interacts with the world. We spend a significant portion of our diligence time on understanding the motivations and desires of the founders and what that means for the entity they are creating. Because of the importance of the founders, it would be difficult for a code to be applied prior to interacting with the leadership team.

    Framing a Statement of Ethics

    To make our policy on ethics easy to recall, we wanted to establish a simple, high-order statement that encompassed the axioms and values described above. The first ethical statement we came up with was: Invest only in things with a reasonable chance to have a net positive effect on humanity. It’s a helpful guideline on the surface, but difficult to implement. All technologies exist on an ever-changing spectrum of useful and harmful as judged by both individuals and societies. Further, “net positive effect on humanity” is tremendously subjective and fails our desire not to be moral police.

    Does Fortnite reasonably have a net positive impact on humanity? I’m sure there’s a strong case to be made that it does not if you’re the parent of a teen that plays it all day (and doesn’t make money off of it). On the other hand, Fortnite brings joy to millions of people who play and watch it, including many of us at Loup. Considering our belief in the sanctity of choice, we would say Fortnite should exist even though you could question the net positive effect on humanity criteria.

    Do Facebook or Instagram have a net positive impact on humanity? There’s a strong case to be made that they don’t when we consider privacy, election tampering, bullying, and the general unhappiness that can arise from spending too much time on social media. On the other hand, both products can, in some ways, bring people closer together despite the negatives. They’ve enabled entrepreneurship. They can help distribute information about valuable social issues, even if in imperfect ways. Again, it would be hard to argue that those products have a definitive net positive effect on humanity.

    With the first code insufficient, we inverted on our second try: Avoid investing in companies that have a high probability of creating an extremely negative outcome on the world. Unfortunately, this code violates the idea of limiting the unpredictable, extremely negative outcome. It also suffers from the same subjectivity around defining what is a negative outcome. More importantly, both of these two initial attempts at a code fail to incorporate the founding team as part of what we’re investing in.

    For our third attempt, we tried to move away from the subjectivity of positive/negative impact by granting the market determination of positivity. We also added a consideration of the founding team: Invest in high-integrity founders making products people love.

    This attempt seems to better embrace a number of our axioms and address the problems we found in our first two attempts.

    First, by putting an emphasis on assessing the integrity of the founders, you create a sort of ethical margin of safety. The ethical margin of safety addresses our inability to predict how any given technology will impact the world long term. With an understanding of a founding team’s motivations and how they see the world, we can have a higher degree of confidence that a founding team will thoughtfully address unexpected ethical questions in the long term. That means founders are not only the margin of safety from a financial standpoint, but also an ethical standpoint, which meshes with our belief that early stage investments are investments in people first, second, and third.

    Second, considering consumer love for a product removes the subjectivity of judging good and bad and respects the sanctity of individual choice. For any given product, if a relatively large enough group loves the product, there must be some good in that product as determined by the group. Fortnite, social media, and, you could argue, things like alcohol or other vice products all pass the test of consumer love. Given the clause about consumer love, we accept that some products deemed worthy by the market can lead to excesses by a group of individual participants in the market. Therein lies the eternal tradeoff of freedom and safety inherent in respecting the sanctity of choice.

    Unfortunately, there’s still a key deficiency in this third attempt at an ethics statement about what we invest in: It doesn’t consider our LPs.

    That brings us to attempt number four, the ethical statement we landed on:

    Invest in high-integrity founders making products people love and that our LPs will be proud to back.

    Implementing the Code

    Now that we’ve established our highest-order ethical mandate, how do we implement it?

    Consider the code in three parts: assessing founder integrity, determining if the market does love or will love the product, and asking if our LPs would be proud to back the company.

    High-integrity people do the right thing consistently. They’re motivated by doing the right thing and derive joy out of doing the right thing. When the right thing isn’t clear, high-integrity people agonize over what the right thing is and act when they find reasonable certainty in the direction of truth. Note, the “right thing” is determined long term by natural law (i.e. truth), but short term by popular opinion, which can be distorted.

    High integrity is about going above and beyond to be trustworthy and truthful in all situations. We want founders that if they found a $5 bill in the grocery store would turn it into customer service in case someone came looking for it.

    High integrity teams developing technologies that could have a dramatic impact on the world may have developed their own ethical codes to guide how they implement their technology. They also respect existing codes established by the industries in which they work.

    When assessing integrity, it’s generally easier to invert and look for poor integrity. As humans, our antennae for poor integrity are finely tuned. Avoiding untrustworthy people is a survival mechanism. Poor integrity manifests itself in dishonesty, questionable motivations, and excessive self-interest.

    If our antennae go off even a little bit to question a founding team’s integrity, we shouldn’t invest.

    For market demand as a signal that a product is “good,” customer adoption is the best assessment tool; however, when investing in early-stage companies customer adoption may not exist. In this case, we have to rely on our intuition about the market for the product based on our ongoing research. We may follow a logical process using market estimates based on available data or conduct ad hoc surveys as part of that research.

    From an operational standpoint, considering the values of our LPs in an investment decision may be the most important consideration. Our LPs are explicitly trusting us to invest their money in good business opportunities, but we feel they’re also implicitly trusting us to invest their money in things they would support personally.

    Given these three components of our code, there are a few simple rules we can follow in order.

    1. There can never be any questions about a founding team’s integrity. If there is even a debate about a founding team’s integrity, the debate can end immediately because we should not invest.
    2. LPs should be the first consideration when determining the ethical viability of a product. Knowing what we know about an investment opportunity based on our diligence process, will our LPs be proud to back this product? Are we comfortable defending why we invested in this product with our LPs? If we think the answer will be no in either case, the debate can end immediately because we should not invest.
    3. Questions about the ethics of a product should be debated on the basis of knowable potential positives and negatives. If an investment opportunity passes the first two rules, it’s unlikely we should have any major ethical reservations about the opportunity: It’s a company run by high-integrity founders that our LPs would be proud to back. That said, there may be some cases where some final ethical considerations about a product are relevant. When that is the case, we should start by considering data about customer adoption — who already loves it? If we can establish this, we move forward. If not, we should consider the positive and negative potential of the product to the extent either can be known.

    Embracing the Code

    Invest in high-integrity founders making products people love and that our LPs will be proud to back.

    We undertook this exercise to establish an ethical framework and define what kinds of companies we want to invest in because we think is important. We’re proud of this code.

    We also accept that any code will be imperfect. Given enough time and investments, there’s a high probability that we’ll invest in some companies that wouldn’t fit our policy in retrospect. To that end, we are the last line of defense. When we make a mistake, we’ll work to fix it.

    We try to be high-integrity founders ourselves. We want to do the right thing all the time. Following our code may mean missing out on some great investment opportunities, but we view this as an acceptable tradeoff for doing the right thing.

    Disclaimer: We actively write about the themes in which we invest or may invest: virtual reality, augmented reality, artificial intelligence, and robotics. From time to time, we may write about companies that are in our portfolio. As managers of the portfolio, we may earn carried interest, management fees or other compensation from such portfolio. Content on this site including opinions on specific themes in technology, market estimates, and estimates and commentary regarding publicly traded or private companies is not intended for use in making any investment decisions and provided solely for informational purposes. We hold no obligation to update any of our projections and the content on this site should not be relied upon. We express no warranties about any estimates or opinions we make. 

    Philosophy
    10 min. read Show less
    Amazon and Walmart Are Taking Aim at Each Other
    Source: Investors Business Daily

    Amazon and Walmart Are Taking Aim at Each Other

    On June 16, 2017, at 9:01 am ET, Amazon announced its agreement to acquire Whole Foods Market for $13.7B. Four minutes later, Walmart announced its agreement to acquire online brand Bonobos for $310M.

    While the announcements were coincidental, the strategies are not. Amazon is growing its physical retail business to better compete with Walmart; Walmart is growing its online retail business to better compete with Amazon. We believe a retail duopoly is emerging.

    We’ve seen a steady cadence of news as Walmart has grown its online retail portfolio, which grew larger earlier today with the addition of Art.com. Similarly, it seems like nearly every day comes more news of Amazon extending its reach into physical retail.

    Amazon’s Physical Retail Presence

    For reference, here’s a select list of Amazon announcements related to physical retail in the last few years: 

    Amazon now operates 107 of its own physical retail stores (including Amazon Go, Amazon 4-star, Amazon Pop-Up, and Amazon Books), plus another ~500 Whole Foods Market stores. Just recently, Bloomberg reported that Amazon has plans to build up to 3,000 cashier-less stores by 2021. But even an ambitious growth plan leaves Amazon well short of Walmart when it comes to physical retail. While Amazon’s total footprint of ~600 retail stores may be surprising, Walmart dwarfs Amazon’s footprint with more than 11,200 stores globally.

    Source: The Verge

    Walmart’s Online Retail Presence

    Meanwhile, Walmart has been just as aggressive in its effort to expand its online retail capabilities. Here’s a select list of Walmart announcements related to online retail in the last few years:

    Walmart’s announcement of the Art.com acquisition is instructive: Adding Art.com to our Palette. The company is aggressively building a quickly growing portfolio of online brands. And yet, Walmart’s online retail business was just $11.5B in 2017, expected to grow 40% to $16b in 2018. For comparison, Wall Street consensus calls for Amazon revenue of $235B in 2018.

    Source: SupplyChain Digital

    A few related insights

    Bottom line: Amazon is way behind Walmart in physical retail and needs to continue making big bets to catch up; Walmart is way behind Amazon in online retail and needs to continue to acquire large online retailers to catch up.

    Our prediction that Amazon would acquire Target in 2018 was wrong. We continue to believe the combination makes sense, however, because Amazon needs to acquire Target-like footprints (1,800 stores) in order to continue its growing retail presence.

    Amazon’s core retail competency lies in logistics (not merchandising). The Amazon Go initiative, for example, works so well because it leverages logistics in physical retail. Amazon Go is a convenient, self-service “warehouse” for items you know you need. But to build most other physical retail — stores where people love to discover new items — requires a completely different competency: merchandising, curation, details, space, and experience. Amazon isn’t great at product design (cf. Amazon Fire Phone, and even amazon.com or the Echo lineup), and retail stores are products. Perhaps this is why The New York Times review of the Amazon 4-star store was so critical: “It is grim. A permanent store with the harried, colorless mood of a hastily assembled clearance-sale pop-up. Lot-Less Closeouts stores have more vim and charm.”

    Disclaimer: We actively write about the themes in which we invest or may invest: virtual reality, augmented reality, artificial intelligence, and robotics. From time to time, we may write about companies that are in our portfolio. As managers of the portfolio, we may earn carried interest, management fees or other compensation from such portfolio. Content on this site including opinions on specific themes in technology, market estimates, and estimates and commentary regarding publicly traded or private companies is not intended for use in making any investment decisions and provided solely for informational purposes. We hold no obligation to update any of our projections and the content on this site should not be relied upon. We express no warranties about any estimates or opinions we make. 

    Amazon, Retail
    3 min. read Show less
    “Alexa Answers” Coupling Human and Artificial Intelligence

    “Alexa Answers” Coupling Human and Artificial Intelligence

    Amazon announced today that it is testing Alexa Answers, a program that crowdsources responses to questions that Alexa currently doesn’t have answers to. One of the principal issues with voice assistants is that they can do a lot, but not everything, creating user frustration. No matter how much data it is trained on, no AI assistant can answer every bizarre question we can think of – similar to the current challenge of self-driving cars. You can drive millions of miles on public roads without encountering every possible edge case.

    In a post on Amazon’s Day One Blog, Bill Barton writes, “Our vision has always been that Alexa will be able to answer all questions in all forms, from anywhere in the world.” Alexa Answers is a step toward making this a reality by leveraging human intelligence where artificial intelligence falls short.

    In the Loup Ventures Manifesto, we describe a spectrum of automation or different degrees to which the future of work will be automated. Of the four basic categories we outline, human-in-the-loop, human-as-AI, pure AI, and pure human, three of them include some form of human involvement. In other words, while we envision a highly-automated future, we recognize that humans will play an important role in the many steps in between.

    We may have spotted Alexa Answers in the wild. We are currently completing our annual smart speaker IQ test, in which we ask 800 questions to each of the leading voice assistants (Alexa, Google Assistant, Siri, Cortana). Not only have we noted an overall improvement in the Information category (where Alexa Answers would show up), but there have been several impressive answers to odd questions that have been incorrect in the past. For example, this round, Alexa correctly answered, “who did Thomas Jefferson have an affair with?” and “what is the circumference of a circle when its diameter is 21?” Going forward we anticipate more one-off questions will be met with helpful responses – an overall improvement to the utility of Alexa. Stay tuned for the full Smart Speaker IQ Test report to be released in two weeks.

    Disclaimer: We actively write about the themes in which we invest or may invest: virtual reality, augmented reality, artificial intelligence, and robotics. From time to time, we may write about companies that are in our portfolio. As managers of the portfolio, we may earn carried interest, management fees or other compensation from such portfolio. Content on this site including opinions on specific themes in technology, market estimates, and estimates and commentary regarding publicly traded or private companies is not intended for use in making any investment decisions and provided solely for informational purposes. We hold no obligation to update any of our projections and the content on this site should not be relied upon. We express no warranties about any estimates or opinions we make.

    Amazon, Artificial Intelligence
    2 min. read Show less