About
    • AMMDI is an open-notebook hypertext writing experiment, authored by Mike Travers aka mtraven. It's a work in progress and some parts are more polished than others. Comments welcome! More.
Search
Page Contents
Full
Page Tree

Minsky/Brooklyn Project

11 Nov 2026 - 11 Nov 2026
Open in Logseq
    • An proposal to implement Society of Mind, https://jmail.world/api/attachments/c2d909feb8a37143c12faa88df5af375 , with Sussman and Abelson. Has the Epstein stank on it:
      • Ask Claude
        • Good — I now have a clear picture of what the Brooklyn Project was. The attachment you linked to appears to be the actual proposal document. Here's what's known:

        • The Brooklyn Project

        • This is a proposal for implementing Minsky's Society of Mind theory as a large-scale AGI system. Here's what's established from the available record:

        • The Document

        • Minsky sent an email attaching two copies of the proposal. The document is titled "The Brooklyn Project," and its listed authors are Hal Abelson, Marvin Minsky, and Gerald Sussman — all MIT professors.

        • When

        • On June 6, 2007, Minsky wrote to Jeffrey Epstein's assistant Lesley Groff, with the subject line "Brooklyn Project."
        • The email came to wider public attention when further Epstein files were released in February 2026.

        • What It Proposed

        • The proposal describes the Brooklyn Project as a six-year initiative, with the goal of developing a large-scale general intelligence based on the model Minsky had refined over decades.
        • The intellectual foundation was explicitly Minsky's Society of Mind theory.
        • The goal was originally the development of artificial general intelligence (AGI) that would reach the level of a toddler within four years and an adult within ten.

        • The Connection to Epstein

        • In the emails, the project is referred to as a new "Manhattan Project" — a large-scale, secret AI and biotechnology research initiative.
        • In one reply, Epstein proposes assembling an "all-star team" of the best researchers.
        • The network Epstein assembled around it consisted of AI researchers, neuroscientists, transhumanists, and researchers working on connectomics, consciousness modelling, and genetic immortality; Epstein's role was as a selector and organizer who brought the right minds together.

        • Where It Went

        • There's no public record of the Brooklyn Project being formally launched or funded.
        • The Society of Mind theory, notably, was never concretely implemented — unlike architectures such as Soar.
        • The proposal appears to have remained exactly that — a proposal circulated in Epstein's orbit around 2007, never reaching fruition as a research program.
        • What emerged instead was less a formal research program and more an informal network of emails, meetings, funding commitments, and quiet ambitions.
        • The attachment you have may be the actual six-page proposal document. It's notable that this is the same period Minsky was also developing The Emotion Machine (published 2006), which extended the Society of Mind framework — so the Brooklyn Project timing aligns with Minsky actively refining those ideas.
    • The Brooklyn Project is a six-year effort to build a large-scale general intelligence based on the model Marvin Minsky has refined over the past thirty years, and expounded in his two books "Society of Mind" and "The Emotion Machine". If successful, this project will be the foundation for a new generation of machines that exhibit general intelligence and common sense. Minsky's model proposes that the mind is composed of numerous agents each with limited expertise. Some agents directly affect behaviors, while others act as critics that modulate the applicability of other agents. Still others are critics on the critics. (Minsky proposes a six-layer architecture.) The Brooklyn Project will be an open software architecture that can assimilate many other programs and data sources that other projects have already developed to provide expertise in specialized domains. By using our layered structure of critics, our system will enable manyother existing resources to work together.
    • Self-Conscious Operating Systems. The largest change in how people use computers came with Apple's use in 1984 of the "desktops" invented by Alan Kay's group at Xerox PARC in the 1970s. Microsoft took 15 years to almost catch up with what Apple had done. Today, despite a great deal of hype, the only discernible movement is moving back to the Unix-like systems developed in the late 60s at the late Bell Telephone Laboratories, then further developed at UC Berkeley in the 70s, and adopted by Sun Computer in their Solaris software in the 1980s. These 30-year-old systems still know little about what their users want to use them for. But a supervisory program with common sense could intermediate between users and traditional dumb operating systems. Ultimately, it should take over most of the computer's functions, by replacing inflexible "operating systems" by processes that are more "conscious," in the sense that they know what they're doing, and why.