Not Jim’s Rho-nix Blog

The RChain Diaspora: Grand Theory, Broken Delivery, and the Holders Left Behind

There is a particular kind of crypto story that never quite resolves. It doesn't end with a single dramatic rug-pull or a perp walk. Instead it metastasizes: a grand theoretical promise raises real money, the working product never arrives in the form that was sold, the original vehicle stalls or collapses, and then the same people, the same ideas, and frequently the same source code reappear under a new name — raising again, promising again, leaving the previous round of believers behind.

RChain is the archetype. To understand it you have to follow not one project but a whole genealogy: a lineage that runs out of Synereo, through the RChain Cooperative and a Washington State regulatory order, and then branches into Casper, Pyrofex/Silvermint, and a chain of community forks named Rhonix, Gorki, and BigSur — with the founder himself now attached to an AI-flavored successor called F1R3FLY. The connective tissue is a body of elegant mathematics and a recurring cast who keep their credibility even as the products keep not shipping.

This piece traces that diaspora. Where there are court records, regulatory orders, or primary documents, this is reported as fact. Where the only sources are aggrieved community members trading accusations, it is reported as allegation — because that distinction is exactly what tends to get lost in stories like this one.

The theory that made it all possible

At the center of everything is one man and one idea. Lucius Gregory "Greg" Meredith is, by background, a genuinely accomplished computer scientist. Before crypto he was the principal architect of Microsoft's BizTalk Process Orchestration, and earlier worked on the Rosette reflective programming language at the MCC research consortium. His signature contribution is the rho-calculus — a "reflective higher-order" variant of the pi-calculus, a formal model of concurrent, communicating processes.

The pitch that flows from this is seductive to anyone who has watched blockchains choke. Most chains execute transactions in a single sequential line, so adding hardware mostly adds redundancy rather than throughput. Meredith's argument was that a smart-contract language built natively on a calculus of concurrency — Rholang, the "reflective higher-order language" — could let non-conflicting contracts run in parallel, scaling with hardware instead of bottlenecking. Layer on namespaces (a sharding mechanism), proof-of-stake consensus derived from the Casper line of research, and "correct-by-construction" formal verification, and you have the recurring promise: a blockchain that is fast, formally proven safe, and decentralized, capable of tens of thousands of transactions per second.

Two things need separating here, because conflating them is how the whole story works. The rho-calculus is real in the narrow sense that matters to mathematicians: it is published, cited in the concurrency literature, and internally coherent. What is not established — and what a decade of attempts has if anything cast doubt on — is the far larger systems claim layered on top: that this calculus could actually be engineered into a fast, formally safe, decentralized network at production scale. That does not follow from the calculus being valid; it is an empirical promise about software, and it is the promise that was sold. The realness of the underlying theory should therefore not be read as a point in the project’s favour. It is closer to the opposite: a genuine pedigree is exactly what allowed an undelivered product to keep raising money across years and rebrands. The trouble, as the rest of this history shows, is the persistent and widening gap between the theory on the page and anything working in the world — and the longer that gap persisted, the harder it becomes to credit the original claims as merely optimistic.

Synereo: the first crowd, the first split

Before RChain there was Synereo. Founded around 2014 by Dor Konforty, Anderson McCutcheon, Yuval Adam, and Greg Meredith, Synereo was a decentralized-social-network project built around the AMP token and the idea of an "attention economy" — a network nobody owned, where users rather than platforms captured the value of attention. Synereo raised on the order of five million dollars across two crowdsales, with the larger raise (roughly $4.7 million) running through the BnkToTheFuture platform in late 2016.

The split that followed set the template for everything after. According to a contemporaneous account by long-time supporter Jim Whitescarver — who bought into the first AMP crowdsale and became an active promoter — the two principals fell into a bitter fight over whether to pivot. Konforty pushed toward shipping a narrower decentralized-social product quickly; Meredith wanted to pursue the grander verified-contracting vision. Whitescarver wrote that he had been drawn in primarily by Meredith's vision, not Konforty's, and was "devastated" by the prospect of the founders parting. They did part. Meredith left to build the thing he actually wanted to build: a blockchain. AMP holders were, in effect, the first cohort to fund the vision and then watch it walk out the door into a new vehicle.

RChain: thirty million dollars and a house in West Seattle

The RChain Cooperative was formed in Seattle on January 11, 2017. The cooperative structure was itself part of the pitch — one member, one vote, run for its members rather than outside venture capital. Members paid a $20 fee to join, and by January 2019 there were more than 1,700 of them.

What happened to the money is not a matter of rumor; it is documented in a consent order entered by the Washington State Department of Financial Institutions (DFI) Securities Division in February 2020. According to that order, between 2017 and 2019 RChain raised approximately $30 million through the sale of millions of RHOC tokens across multiple offerings that were never registered with state or federal securities regulators. The first sale, in August–September 2017, brought in about $9.9 million. Then, over just four days at the end of December 2017 — a window in which Meredith reportedly described demand as a "runaway train" — RChain raised roughly $18.7 million more. Additional sales in 2018 and 2019 added over a million dollars on top.

RHOC was sold as a placeholder: holders were promised they could convert RHOC to the network's eventual native token, REV, at a one-to-one ratio once the RChain mainnet launched. The token's market history tracks the arc of the hype. According to the DFI order, RHOC rose from around $0.20 in October 2017 to as high as roughly $3.00 in January 2018, fell to about $0.07 by late 2018, and by January 2020 traded for roughly one to two cents.

The DFI's findings are unflattering in their specifics. RChain, the regulator found, repeatedly engaged in general solicitation while calling its sales "private," kept using an outdated August 2017 whitepaper that described a "maximum" raise of $15 million even after raising nearly double that, and did not disclose to the Division that it had completed the second, larger offering until February 2019. The order also notes undisclosed items that any investor might have wanted to know: that advisors were paid a 3% performance commission on funds raised; that Meredith had filed for bankruptcy in 2007; and that roughly $1.5 million of proceeds went to buy a five-bedroom house in West Seattle as an office, which was later sold at a loss of about $275,000. It further found that Meredith was issued 2.4 million RHOC tokens that were supposed to be locked until 2019, but that he sold a significant portion in 2018 and transferred more than 250,000 to his daughter.

RChain Cooperative and Meredith neither admitted nor denied the findings. They agreed to cease and desist from the registration and anti-fraud violations, and to pay fines of $50,000 (the cooperative) and $25,000 (Meredith personally), plus investigative costs. The order closes with a standard but pointed line: willful violation of it is a criminal offense.

And the mainnet? It was originally supposed to be ready by December 31, 2018. As GeekWire reported in early 2019, that date slipped, and the network never arrived as the fully decentralized, high-throughput, formally verified system that RHOC holders had been sold. Meredith eventually resigned as president of the cooperative in August 2022. The RHOC-to-REV conversion that was the entire economic premise of the token never delivered the value holders had been promised.

The first branch: Casper, by way of Pyrofex

While RChain was raising and stalling, the talent and ideas around it were already dispersing — and the most commercially successful branch grew from that dispersal.

Pyrofex Corporation was founded in 2016 by two ex-Googlers, Nash Foster and Mike Stay. Stay holds a PhD in computer science with a category-theory specialization and had collaborated academically with Meredith; Foster came from a site-reliability and security background. By Foster's own account, Pyrofex worked with RChain from spring 2017 until about October 2018, largely recruiting and managing the Scala engineering team that built RChain's node software. One of Pyrofex's people on the RChain program was Medha Parlikar.

In 2018, Parlikar and financier Mrinal Manohar co-founded CasperLabs in Zug, Switzerland (other co-founders have been variously named in coverage, including Clifford Sarkin). As GeekWire reported in January 2019, CasperLabs used "some of the code base of RChain" — confirmed at the time by a managing partner of Reflective Ventures, the fund tied to RChain — and several people who had worked in the RChain orbit, including Parlikar, ended up there. CasperLabs went on to raise roughly $28 million and launched the Casper Network (CSPR) in 2021, pivoting hard toward enterprise blockchain and, later, AI-governance products.

Casper also inherited a legal fight. The "Casper" name comes from the CBC Casper proof-of-stake research associated with Ethereum's Vlad Zamfir (work he traces back to 2015 with Vitalik Buterin). CasperLabs hired Zamfir as a consensus architect and then, in 2019, filed for and obtained a U.S. trademark on "Casper" in connection with blockchain. In March 2021, Zamfir sued CasperLabs in the Southern District of California, alleging false designation of origin under the Lanham Act and related fraud and unfair-competition claims, and arguing the company had commercialized a name that belonged to an open-source research lineage. CasperLabs disputed his account; the court dismissed some claims and let others proceed. Whatever its ultimate merits, the suit captured a theme that runs through this whole diaspora: a recurring tension between open-source communities and the more proprietary teams that commercialize their work.

Pyrofex itself, meanwhile, kept going. After parting from RChain, Foster and Stay publicized a claimed vulnerability in the Casper proof-of-stake design and promoted their own consensus protocol, "Casanova," and a platform called CDelta that was marketed at up to 50,000 transactions per second. During the pandemic this resurfaced, rebranded, as Silvermint — described as a "leaderless, scalable" proof-of-stake blockchain projected to handle more than 140,000 transactions per second. The throughput numbers escalate; the independently verifiable, widely used production network remains elusive.

The second branch: Rhonix, Gorki, and BigSur

The other branch grew not from people who left RChain early but from the community that stayed — and then inherited the wreckage.

When Meredith resigned in August 2022, the cooperative needed leadership, and Ting Yang was named interim president. This is where the cooperative's own governance record becomes the most reliable witness. The board minutes repository (rchain/board) documents the transition precisely: on 23 August 2022, both Ting Yang and Jim Whitescarver — the long-time supporter who had followed the project all the way from Synereo — were appointed to the board, rather than elected by the membership. The same roster captures the wider churn of directors over the project's life: Greg Meredith (elected 2020, resigned), and a 2021 cohort elected on 24 October that included Darryl Neudorf, Theo Hallenius, and Shixi Lin, the last of whom subsequently resigned. The roster is worth recording simply because it names who was steering at each stage; it carries no implication of wrongdoing by any director, several of whom appear to have stepped away precisely as the project foundered.

That transition is also where the documented record starts to give way to a community at war with itself. The appointment was contested by some long-standing members, and the period that followed produced exactly the kind of dispute that is impossible for an outsider to fully adjudicate: competing claims about who controlled the code, the funds, the domains, and the brand. (A widely circulated community video of the 23 August 2022 meeting is captioned to accuse Yang of misleading members; that is one faction's characterization, not an established fact, and is treated here as such.)

What can be stated from the public record is the pattern of forks. The RChain node software (RNode — CBC-Casper consensus plus Rholang execution) was forked and continued under a succession of names. Repositories appear under Rhonix Labs (the "rhonix" project) — including a fork under the GitHub handle of Theo Hallenius, "TheoXD," the same Theo who had sat on the RChain board — then under a "Gorki" project, then under "bigsur-network." The lineage is visible in the code itself: these repos describe the same "decentralized, economic, censorship-resistant, public compute infrastructure" using the same CBC-Casper-plus-Rholang design, and several still carry the same blunt disclaimer the RChain code always did — that it "has not yet completed a security review" and should not be used "to transfer items of material value." The same names recur on the engineering side: Tomislav Grospić, a wallet and tooling developer documented building RChain community software (he demonstrated a hardware-wallet integration in a 2020 community call), also surfaces in the later payment records — an illustration of how small the actual developer bench was, and how the same handful of people carried the code from one banner to the next.

The rebranding is also documented. The Gorki site now redirects to BigSur Network, accompanied by a notice that the community has changed its name and — tellingly — that it is "in no way affiliated with, endorsed by, or related to RHOCALCULUS INNOVATIONS LTD or any of its products or services." That single sentence is the public fingerprint of a falling-out: a community fork explicitly distancing itself from the corporate entity that had become associated with the project's stewardship.

Beyond that, the picture becomes a thicket of accusation. Aggrieved participants — most visibly Jim Whitescarver (the same long-time supporter who funded Synereo, posting as "jimscarver") and Patrick Mockridge, who writes the "Technology Truth" Substack — have publicly alleged misconduct in how the post-RChain effort was run, including misused funds and contested invoices.

Those tensions did harden into a formal dispute, and here the trail runs into actual documents rather than forum posts. By late 2023, Alan Tominey — who handled payments for Rhocalculus Innovations Ltd, the company that had become associated with the project's stewardship — emailed Ting Yang asking him to account for specific invoice line items, among them medical-benefit and contract payments and travel expenses that had been submitted without receipts. On 7 February 2024, solicitors acting for Rhocalculus sent a pre-action "letter of claim" to a developer associated with the project, Tobias Hoenig, alleging breach of contract, copyright and trade-mark infringement, unfair competition, and misuse of confidential information in connection with the Gorki Network, and asserting that he had acted "in concert with or as directed by" Ting Yang to damage the company's interests and mislead potential investors while it was raising further funding. A related matter was reported to Police Scotland.

It is essential to be precise about what that does and does not mean. A letter of claim is the opening move in a civil dispute — one party's case, drafted by its own lawyers — not a judgment. A report to the police is not a charge, still less a conviction. A queried invoice is not by itself a fraudulent one; Tominey's own message was a request to "clarify," and even allowed that he might simply be missing receipts on the payment platform. None of these allegations has been tested or proven before a court or regulator, and the available record does not include any response from Yang or Hoenig, nor any indication of how — or whether — the matter was ultimately resolved. What can fairly be said is narrow but real: a company at the center of the project pursued formal claims and a police referral against people connected to it, over money and intellectual property, while it was trying to raise more. That is the documented part. Everything past it remains contested.

What is not merely alleged is the structural outcome. RHOC and REV holders — the people who put up that roughly $30 million — did not end up with the network they were promised. The intellectual core they funded, the rho-calculus and Rholang, lived on and kept being forked into new projects; the holders did not necessarily come along for the value.

A parallel banner: Alux Network

Running alongside the Rhonix/Gorki/BigSur thread is Alux Network, another project pitched on the very same idea. Its website markets Alux as “the concurrency control layer for Web3” — a leaderless, EVM-compatible, sharded blockchain whose performance, it says, scales linearly as hardware is added, in deliberate contrast to “state machine” chains that slow down as nodes join. The explanatory metaphors (a “Tuple Space” in which each resource is a LEGO-like block that only one actor can hold at a time, so transactions need not queue) are a recognizable restatement of the rho-calculus concurrency thesis that has anchored every project in this lineage since Synereo. Whether Alux is best understood as a sibling fork, a rebrand, or an adjacent effort by overlapping people is not something the public record settles cleanly; what is clear is that the same core promise — concurrency, sharding, ACID transactions, near-unlimited scaling — is once again the headline, and once again the burden of proof is a working, independently validated, widely used network rather than a website describing one.

The latest branch: F1R3FLY and the AGI pivot

The most recent chapter folds the founder back into the story. Meredith's current vehicle is F1R3FLY.io, which continues the Rholang/rho-calculus line directly. In March 2023, F1R3FLY announced a partnership with Ben Goertzel's SingularityNET and TrueAGI to bring Rholang's concurrent execution to OpenCog Hyperon and its MeTTa language — repositioning the decade-old concurrency thesis as critical infrastructure for artificial general intelligence.

The marketing language has updated for the moment ("the AI-native execution layer for verifiable Web3 and agentic systems"), but the underlying claims are continuous with 2017: a validator-based engine running Rholang contracts in parallel against shared state, with blockchain-grade trust and concurrency that scales with hardware. The GitHub organization is active, and there are signs of overlap with earlier players (a fork sits under Nash Foster's handle, "nashef"). The public node code still carries the same warning that it has not completed a security review and should not be used in production. After nearly a decade and tens of millions of dollars across predecessor projects, the technology is, by its own repositories' admission, still pre-production.

What the genealogy actually shows

Step back from the individual disputes and a shape emerges that is more important than any single accusation.

First, the credentials are real, but they are not the same thing as either competence at the job actually promised or good faith in promising it — and the two should not be allowed to borrow each other’s shine. Meredith’s rho-calculus is genuinely published mathematics, and Mike Stay is a genuinely credentialled category theorist. But a published calculus and a Microsoft résumé are credentials in formal computer science; they are not evidence that the holder can ship a secure, decentralized, high-throughput network, and across roughly a decade the people in this story did not do so. That is the uncomfortable question the genealogy forces: when accomplished people repeatedly promise a specific working system, raise tens of millions on the promise, and repeatedly fail to deliver it while moving on to the next vehicle, at what point does “brilliant theorists who kept hitting hard engineering problems” stop being the most economical explanation? The record does not let an outsider answer that with certainty. What it does do is make the charitable reading harder to sustain with each repetition — and it shows that the real expertise functioned less as a guarantee of delivery than as the reusable authority that let each new vehicle raise from a fresh set of believers.

Second, the assets that travel are the ideas and the code — not the obligations to holders. Rho-calculus and Rholang move cleanly from Synereo to RChain to Rhonix/Gorki/BigSur, sideways to Alux, and on to F1R3FLY. The RNode codebase forks forward intact. But AMP holders, then RHOC/REV holders, then the later community do not inherit anything comparable. Each cohort funds a stage of the journey and is then, in practical terms, left at that station.

Third, non-delivery keeps getting reframed as a fresh start. A blockchain that didn't ship becomes a faster blockchain that will ship, becomes a leaderless blockchain, becomes an AI execution layer. The throughput numbers rise — 7,000, then 50,000, then 140,000 transactions per second — even as the independently verifiable, widely used, value-bearing network stays just over the horizon.

Fourth, the doubt this raises does not fall evenly, and it is worth being precise about where it lands. For Meredith, it is not merely a matter of missed deadlines: a state securities regulator made findings — neither admitted nor denied — that material facts were withheld from investors, that an out-of-date whitepaper understating the raise stayed in circulation, that sales solicited at scale were described as "private," and that tokens supposed to be locked were sold anyway. Those are findings about candor, not just competence, and they make the most generous reading of his later ventures correspondingly harder to extend on trust alone. For the technical principals — Stay among them — the legitimate doubt is narrower but real: a recurring habit of headline performance numbers (tens of thousands, then six figures, of transactions per second) attached to systems that, on the public record, never appeared as independently validated, widely used networks. Big claims, thin demonstrated delivery. And for the boards that governed these entities, the fair question is one of oversight: bodies that presided over tens of millions of dollars and, at RChain, a regulator's order, that at the pivotal 2022 transition filled seats by appointment rather than member vote, and that included members without evident relevant fiduciary or technical expertise. None of that, on its own, proves bad faith by any individual director or developer — several resigned, and resignation is at least as consistent with disillusionment as with anything worse — but it is a record that earns scrutiny rather than the benefit of the doubt.

So no, this does not require anyone to be a cartoon villain, and an honest telling will not pretend the contested civil and criminal allegations are settled when they are not. But the repeatedly charitable reading — brilliant people, unlucky execution, again and again — has to do more work each time it is invoked, and at some point the simpler account deserves a hearing: that grand, credentialled theory served, whatever the intentions behind it, as a renewable instrument for raising money that the underlying delivery never justified. Whether that reflects misplaced optimism, overreach, or something less innocent is not something the public record fully resolves, and this article does not claim to. What the record does show is the outcome, and it is the same every time: the theory survives, the founders move on to the next banner, and the holders are left behind. After a decade, the consistency of who pays and who walks is itself a kind of evidence.

A note on sourcing and allegations

This article distinguishes deliberately between three tiers of evidence:

Selected sources