• What this course is about: networked computers as tools for social self-organization.

    • Computers are powerful tools enabling people to communicate, cooperate, deliberate, make decisions, and innovate in a multitude of ways.  Some uses:

      • Communication: Alice sends Bob an E-mail or makes a conference call.

      • Association: a set of users form an online discussion group or forum.

      • Cooperation: users work together to write or design something together.

      • Curation: users finding, filtering, vetting, organizing relevant content.

      • Deliberation: users employ formal processes to debate issues or policy.

      • Decision: users employ social choice (e.g., voting) to decide collectively.

      • Organization: users form organizations/societies, possibly decentralized.

      • Governance: users employ processes to govern membership, behavior.

      • Incentives: digital payments, cryptocurrencies, finance and investment.

      • Evolution: collectively finding and introducing better ways to self-organize.

    • Promises of computing and networking technology as a “democratizing” force:

      • Inclusiveness:

        • Borderless communication with anyone anywhere on Earth

        • Access to information for “everyone”

        • Ability to collaborate productively with anyone globally

        • Ability to associate with like-minded people anywhere

      • Support for individual freedoms:

        • Speech: privacy, anonymity supports expression of true sentiment

        • Information: access to [good?] information from anywhere

        • Association: find, form groups with like-minded people anywhere

        • Press: anyone can publish blogs, social media, etc.

      • Resistance to control, censorship by entrenched authorities

        • Internet as a tool to “route around” failures, censorship

        • Cryptocurrencies as tools for bottom-up economic empowerment

    • But these tools sometimes succeed, but often fail to serve people in a multitude of ways, either due to flaws in the abstraction or flaws in the implementation.

      • Abstraction flaws and challenge examples:

        • Unscalable communication or human attention costs: e.g., broadcast communication, or “pure” direct democracy.

          • Naive cooperation designs O(n2), fail to scale.

        • Unreliable or misbehaving human participants: e.g., spam, trolling, sock puppetry, harassment, fake news, fake reviews.

          • The design of the platform and process greatly affects the behavior of its users - so how to design properly?

          • Who should “police” a community, and how?  Government, platform operator employees, community self-policing?

        • Unintended effects of algorithms: e.g., polarization, radicalization

          • Social recommendation/newsfeed systems, echo chamber

          • Interaction-maximizing and click-maximizing algorithms and anger-exploiting media (e.g., YouTube).

        • Bias, both real and perceived, from self-selecting online groups

          • Tribalism, mistrust of information from other ideologies

          • Information relativism: lack of objective notion of “truth”

        • Overcompression of noisy, unreliable, manipulable information.

          • Large-scale campaigns, “beauty contest” elections.

          • Marketing-driven vs “efficient” markets, lemon markets.

          • Accountability of large [government] bureaucracies to electorate via the bottleneck of representation.

        • The competence- and expertise-identification problem

          • Populace choosing experts: may be poor judges

          • Experts choosing experts: unaccountable, may gradually become insular and fragmentation-prone.

        • Funding/incentivizing the production, curation of information

          • Free information vs monetization, copyright, DRM

          • Social media undermining traditional business models

          • Who bears the cost of information review and policing?

            • Users?  Private platform operators?  Governments?

      • Implementation flaws and challenges:

        • Network security, availability, resilience; “routing around” failures.

          • The upsides (resilience) and downsides (spam)

        • Identity and personhood problem: weak or strong digital identities; security, privacy, and coercion, Sybil and sock-puppet attacks.

          • The tension between anonymity (for freedom) versus accountability (for civility).

          • Individual misbehavior versus Sybil amplification.

        • Software bugs, exploitable flaws.

          • Bitcoin wallets as a universal bug bounty.

          • Smart contract bugs: e.g., DAO, Fomo3D attacks.

        • Fairness challenges:

          • Investment-centric versus democratic stake models, the problem of “rich get richer” and power centralization

            • Corporations versus governments, foundations

            • Proof-of-work/stake/etc versus PoP

        • Inclusion challenges:

          • Digital divide: access to/expense of devices, connectivity.

          • Deliberate exclusion: disenfranchisement, guardianship.

          • Effective exclusion via trolling, harassment.

          • Identity attributes, personal data, coercion, black market.

    • Goals: What would we like to achieve?  What might our long-term end goals be?

      • Technology enabling and incentivizing civil bottom-up self-organization.

      • Ensure freedoms of expression and association but with accountability.

      • Enable attention-limited users to obtain the (good) information they need to participate fully as time and interests permit, avoiding polarization.

      • Give communities not just voting but effective and scalable deliberation, agenda-setting, and long-term governance evolution processes.

    • Particular social/technical tools and techniques this course will explore:

      • Basic fault analysis techniques applying to failures and compromises.

      • Network/graph algorithms for communication, and social/trust networks.

      • Content filtering, recommendation, curation, and peer review processes.

      • Social choice: election methods, sampling, direct and liquid democracy.

      • Deliberation processes: peer review, online deliberation, agenda-setting

      • Techniques to achieve privacy and anonymity with accountability: e.g., verifiable shuffles, anonymous credentials, zero-knowledge proofs, PoP.

    • Deployed technologies we will look at, exploring their promise and flaws:

      • USENET (historical) and other public discussion forums.

      • News media: both traditional and social media platforms.

      • Deliberation: applications like Loomio and Liquid Feedback

      • Peer review: Slashdot, Reddit, GitHub, HotCRP

      • Cryptocurrencies and smart contracts: Bitcoin, Ethereum.

      • Organizations and decentralized autonomous organizations (DAOs).

  • Course logistics

    • TAs introduction

    • Content: Readings, lectures, quizzes, occasional hands-on exercises

      • Some may involve light programming in various languages, but not programming-intensive.

    • Sessions: approx 5 hours per week total expected time commitment

      • Lecture: attendance normally expected and required

        • Will make provisions to ensure that missing one or two won’t seriously hurt grade, provided you catch up via colleagues

        • Discussion-oriented, normally blackboard-based, no slides

      • Exercises: need to attend only as directed in assigned exercises

        • TAs will present to help with assigned exercises

      • Practical work: purely-optional time/space for reading, discussion

        • Instructors will not be present

    • Grading

      • Approx 50% - quizzes and exercises during semester

      • Approx 50% - final exam


Post-lecture blackboard snapshot from 2019:



Modifié le: lundi, 5 octobre 2020, 13:23