Distroid Issue 32
Findings covering the growth of Y Combinator, how to define and measure amplification in recommender systems, pondering the need for tech criticism, and decentralizing DApps with tokenized rebates
Introduction
Welcome to this week’s edition of Distroid, a newsletter for curated findings, actionable knowledge, and noteworthy developments from the forefront of tech, governance, research, and technology (i.e., the frontier).
In this newsletter:
Digest
Y Combinator: The Institute of Innovation
Making Amplification Measurable
Does Tech Criticism Still Matter?
Centralized Dapps Through Tokenized Rebates: A Modest Proposal
Bringing data minimization to digital wallets at scale with general-purpose zero-knowledge proofs
Digest
Y Combinator: The Institute of Innovation
News
Mario Gabriele
The Generalist
2022-06-12
If you only have a couple of minutes to spare, here's what investors, operators, and founders should know about Y Combinator.
Full stack venture. YC is best-known for offering a three-month program for new entrepreneurs. That mental model is outmoded. The organization founded by Paul Graham now helps from inception through pre-IPO.
Network effects. YC is a venture capital firm with network effects. By batching its portfolio into cohorts and connecting them through internal tools like “Bookface,” it has created a structure that becomes stronger as it scales.
Pricing power. As venture capital has become more competitive, valuations have risen. Not so at YC. The accelerator manages to invest in startups at a significant discount that has increased relative to the market over the past decade.
Matters of diversity. Compared to industry benchmarks, YC invests in more women, black, and Latinx-founded businesses – but its numbers have stagnated in recent years. Many feel YC could do more to increase its share of underrepresented founders.
A global presence. Just 49% of startups in the latest YC batch came from the United States. As more pre-seed and seed funds have come to market in America, YC has found success attracting promising entrepreneurs from countries like India, Mexico, and Nigeria. One nation that YC hasn’t cracked? China.
Making Amplification Measurable
News
Luke Thorburn, Jonathan Stray, Priyanjana Bengani
Tech Policy Press
2023-04-28
The term “amplification” is ubiquitous in discussions of recommender systems and even legislation, but is not well-defined. We have previously argued that the term is too ambiguous to be used in law, and that the three most common interpretations — comparison with a counterfactual baseline algorithm, mere distribution, or a lack of user agency — are either difficult to justify, or difficult to measure consistently. That said, the way recommender systems are designed does influence how widely different types of content are distributed. Amplification is an evocative term to describe this reality, and remains a prominent subject of research. For these reasons, it would be useful to have a common understanding of what amplification is and how to measure it.
In this piece, we propose five properties that measures of amplification should have if they are to be useful in discussions of recommender policy. Specifically, they should: (1) define the content of interest, (2) define an appropriate baseline, (3) focus on impressions as the relevant outcome, (4) isolate the effect of the algorithm from that of human behavior, and (5) specify the time horizon over which they apply. The first three properties are relatively easy to satisfy, while the latter involve causal inference and systemic equilibria, and pose greater challenges. A measure that satisfies these properties can be said to measure relative algorithmic amplification.
relative algorithmic amplification: A change in the distribution under one algorithm as compared to an alternative algorithm, holding user behavior constant.
We think that this definition is both clarifying and useful, and that, as a research community, our goal should be to develop measures of relative algorithmic amplification which satisfy each of the above conditions. We discuss each in more detail below."
Does Tech Criticism Still Matter?
News
Jasmine Sun
Reboot
2023-04-08
As Logic Magazine cofounder Ben Tarnoff recently tweeted, “tech critiques themselves have begun to feel algorithmic.” The headlines repeat themselves: X is not a panacea. Silicon Valley is reinventing the Y. Tech bros are ruining Z. On one hand, it’s a good thing that tech criticism has graduated into the mainstream. On the other, these ChatGPT-level takes feel like a worrying signal of a dying discipline.
I consider it irresponsible to continue editing Reboot, a publication about technology, without engaging these questions. In the vein of organizational subtraction, it would be both wasteful and self-important to continue the project past the need for its existence—especially as we enter our fourth year.
So I ask honestly: Ought Reboot continue to exist? Are we adding any value? Are we teaching people something they don’t already know? Or, conversely, at what point are we simply “raising awareness” for the same old problems, gloating about our moral superiority, and doing nothing to change them?
Tech critic Evgeny Morozov would say no—we don’t matter. In “The Taming of Tech Criticism,” he mourns his field’s inefficacy:
Radical technology critics face an unenviable choice: they can either stick with the empirical project of documenting various sides of American decay (e.g., revealing the power of telecom lobbyists or the data addiction of the NSA) or they can show how the rosy rhetoric of Silicon Valley does not match up with reality (thus continuing to debunk the New Economy bubble). Much of this is helpful, but the practice quickly encounters diminishing returns. After all, the decay is well known, and Silicon Valley’s bullshit empire is impervious to critique.
What’s missing, argues Morozov, is vision. Most critics are preoccupied with “design problems, and their usually easy solutions,” or attribute technological harms to builders’ “false consciousness” rather than economic structure. He continues: “Changing public attitudes toward technology—at a time when radical political projects that technology could abet are missing—is pointless.”"
Centralized Dapps Through Tokenized Rebates: A Modest Proposal
News
Caleb Shough
Variant Fund
2023-05-08
Crypto founders face a challenge: Many want to give token ownership to their users from the beginning, but as soon as they do, fear of regulatory action forces them to decentralize and distance themselves from their product.
As our general partner Jesse Walden recently pointed out, user ownership and decentralization are not the same thing; we should be able to have the former without requiring the latter. This post explores an approach that would allow crypto startups to achieve user ownership while maintaining centralized control where necessary.
We can do this by using rebates.
A rebate is an old concept: You pay money now and receive some portion back later. Let’s turn rebates into tokens. These tokenized rebates are a potential pathway to giving users “ownership” at a company’s earliest stage.
Bringing data minimization to digital wallets at scale with general-purpose zero-knowledge proofs
Research
Matthias Babela, Johannes Sedlmeir
2023-01-02
Today, digital identity management for individuals is either inconvenient and error-prone or creates undesirable lock-in effects and violates privacy and security expectations. These shortcomings inhibit the digital transformation in general and seem particularly concerning in the context of novel applications such as access control for decentralized autonomous organizations and identification in the Metaverse. Decentralized or self-sovereign identity (SSI) aims to offer a solution to this dilemma by empowering individuals to manage their digital identity through machine-verifiable attestations stored in a “digital wallet” application on their edge devices. However, when presented to a relying party, these attestations typically reveal more attributes than required and allow tracking end users’ activities. Several academic works and practical solutions exist to reduce or avoid such excessive information disclosure, from simple selective disclosure to data-minimizing anonymous credentials based on zero-knowledge proofs (ZKPs). We first demonstrate that the SSI solutions that are currently built with anonymous credentials still lack essential features such as scalable revocation, certificate chaining, and integration with secure elements. We then argue that general-purpose ZKPs in the form of zk-SNARKs can appropriately address these pressing challenges. We describe our implementation and conduct performance tests on different edge devices to illustrate that the performance of zk-SNARK-based anonymous credentials is already practical. We also discuss further advantages that general-purpose ZKPs can easily provide for digital wallets, for instance, to create “designated verifier presentations” that facilitate new design options for digital identity infrastructures that previously were not accessible because of the threat of man-in-the-middle attacks