Previously on SneerClub, we looked at the maze of overlapping nonpofits which provide Bay Area rationalism / EA its infrastructure. A few months ago CFAR filed Form 900 for 2025. This confirms that in March 2025, they handed over $21 million of assets and $21 million of liabilities, mainly the compound built around the Rose Garden Hotel, to something called Lightcone Infrastructure, Inc.

EXCLUDED LIGHTCONE REVENUE FROM THE AUDIT REPORT SINCE LIGHTCONE BECAME A SEPARATE COMPANY RECENTLY. ON MARCH 20, 2025, THE ORGANIZATION AND LIGHTCONE INFRASTRUCTURE, INC. (“LCI”) ENTERED INTO A TERMINATION OF FISCAL SPONSORSHIP AND TRANSFER AGREEMENT PURSUANT TO WHICH THE ORGANIZATION TERMINATED ITS FISCAL SPONSORSHIP OF THE LIGHTCONE PROJECT AND TRANSFERRED TO LCI ALL ASSETS, LIABILITIES, AND RESPONSIBILITIES ASSOCIATED WITH THE PROJECT, INCLUDING THE ORGANIZATION’S MEMBERSHIP INTEREST IN LIGHTCONE ROSE GARDEN LLC. MANAGEMENT RECORDED THE TRANSFER FOR ACCOUNTING PURPOSES AS OF MARCH 31, 2025.

Lightcone Infrastructure, Inc. is also registered as a nonprofit.

CFAR still listed $2.41m of revenues and a $300k profit for 2025 which is a lot for an organization which had very few visible activities that year except for running the website and compound which they just renounced. Jaan Tallinn must be satisfied as long as someone keeps repaying his interest-free loan.

I have not looked into who they are still paying and how much and how the money flows. I have washed my hands of these particular cranks.

  • David Gerard@awful.systemsM
    link
    fedilink
    English
    arrow-up
    0
    ·
    6 hours ago

    Lightcone is Habryka’s thing. I thought Lightcone was also CFAR.

    (open question how much of the confusion is deliberate and how much is rationalists ratting)

    • CinnasVerses@awful.systemsOP
      link
      fedilink
      English
      arrow-up
      0
      ·
      1 hour ago

      My working hypothesis is that from 2021 to 2024 Habryka took over CFAR and used its bank accounts and 501(c)3 status for bigger operations in Berkeley while Salamon sat in a hole (or a former B&B with a seaview) and thought about all the thing she had done to give Yud child soldiers to fight Machine God and maybe provide other benefits after hours. In 2024 CFAR still listed itself as owning the Rose Garden compound and called Lightcome a “project brand.” In 2025 she got Habryka to move his projects and assets into their own organization but they do not seem to have announced this publicly even though they had fundraising campaigns at the end of 2025 and hold lots of events at the Rose Garden.

      If rats get their ideas of corporate governance from Bay Area VCs, its likely that she was dictator for life and they could not reorganize until they talked her in to signing something.

  • istewart@awful.systems
    link
    fedilink
    English
    arrow-up
    0
    ·
    11 hours ago

    I dunno, isn’t it usually a bad sign when these kinds of organizations consolidate down to one compound?

      • gerikson@awful.systems
        link
        fedilink
        English
        arrow-up
        0
        ·
        4 hours ago

        you jest but I have a high probability we’ll see something like Heaven’s Gate come out of this within a few years

        • flere-imsaho@awful.systems
          link
          fedilink
          English
          arrow-up
          0
          ·
          4 hours ago

          an attempt at black humour. because, yeah, some of them are probably cultish enough to not cope well when the autoconfabulator industry crashes.

          • scruiser@awful.systems
            link
            fedilink
            English
            arrow-up
            0
            ·
            19 minutes ago

            I think the biggest potential motivator for doomer suicide would actually be from the other direction: some AI apparently advances enough the rationalist become convinced it is going to secretly produce nanobot factories in preparation for a takeover (such as in this story or this classic piece of rationalist mythos building CW ponies) and they need to die before they can get “I Have No Mouth, and I Must Scream”'d (or uploaded into a pastel dystopia) by the malevolent AI. The current paradigm probably has a bubble bust before it can produce anything that gets them that scarred (even with how good they are at taking the hype at face value) but it still feels like a landmine… any news story that they can delude themselves into being an imminent indicator of unfriendly AI takeover might prompt such action.

            Agreed the boosters and e/accs might have some crashouts when the LLM industry crashes.

          • gerikson@awful.systems
            link
            fedilink
            English
            arrow-up
            0
            ·
            4 hours ago

            they’re already worried about “abliterated” models designing their own chips and self-upgrading recursively

            https://www.lesswrong.com/posts/yQmzPMMeBLmGuZokr/i-m-worried-about-llms-accelerating-hardware-for-cheap-asi

            You could probably make a superintelligent AI with some currently unknown new paradigm for <$10k if you had the right hardware and algorithms

            discovering this paradigm is left as an exercise to the reader

            • scruiser@awful.systems
              link
              fedilink
              English
              arrow-up
              0
              ·
              17 minutes ago

              Well obviously if LLMs can slop together a Lean proof, slopping out new AI paradigm(s) and major hardware advancements is just around the corner!