Renlabs was founded in late 2024 as a hard-fork of Commune AI, with a more focused vision: Torus, a blockchain designed to operate as a living cyber organism. The idea is a network of AI agents forming swarms, coordinating autonomously, building on each other. The token launched on January 4, 2025.

Joining came shortly after the launch. There was no product yet, no internal structure, no documentation: just a technical foundation and a team that needed to move. The immediate work was the Torus website, contributing alongside a strong frontend lead and UI/UX designer to ship what the launch required.

From there: roadmaps, squad organization, hiring plans, and documentation written from scratch. A UX overhaul on the core bridging flow, designed and executed together with the frontend lead and designer, cut bridging time from ~20 min to ~3 min.

Torus is still early. The role spans product direction, technical program management, and direct code contribution.

Higher Order Company (HOC) was co-founded by Victor Taelin and Vitor Sipher in early 2023, built on a bet that Taelin had been developing for over 15 years: HVM, the Higher Order Virtual Machine, a parallel runtime based on Interaction Combinators that behaves fundamentally differently from anything before it. The starting question was practical: what could this technology actually do in the market? To explore that, the company raised $4M+, incorporated in the US and Cayman, and set out to find out.

The main exploration was Bend: a massively parallel programming language that feels like Python but scales across multi-core CPUs and GPUs without locks or mutexes. The community response was enormous, with coverage from Fireship and Primeagen. But the same question kept coming back: where does this actually make money? Entering the programming language market is brutal.

The conclusion reached was that traditional programming is probably not the right long-term bet. People will be vibecoding, not learning new languages. Part of the answer was rethinking Bend itself: Bend 2 shifted toward being an AI-first language, a way to interface with HVM where the intelligence comes built in and the underlying system stays invisible. You get the power without learning the machine. By then, a bigger picture had also come into focus: understanding computation at this level could be a path toward AGI. The thesis shifted toward symbolic AI: using HVM as a symbolic transformer rather than a general runtime, not LLMs, but a fundamentally different model where HVM's architecture gives a real, structural advantage. A final round was raised to fund the compute cluster needed to explore this.

As HOC moved into a phase of deep foundational research, the mandate shifted. The round was closed, the structure was in place, and what the company needed next was sustained technical experimentation, not commercial development. HOC is still building. I believe in what they are working on.

Kindelia Foundation was founded in early 2022 by Victor Taelin and Vitor Sipher as the first real attempt to put HVM to work. The runtime at that stage was more naive: the foundational ideas around Interaction Combinators were in place, but the implementation was early, still closer to a research prototype than a production system. The starting question was simple: could this actually hold up as the base for something real?

The concept was the Kindelia chain: a tokenless hub, a peer-to-peer functional computer for hosting decentralized applications that run forever. Where Ethereum had accumulated layers of complexity over years, Kindelia stripped everything back. No native coin. No proof-of-stake. Around 10,000 lines of Rust against 600,000+ in Go Ethereum. The design asked what the minimum viable decentralized computer actually looked like.

The work went beyond the chain design. The KGT (Kindelia Genesis Token) was written from scratch in Kind, Kindelia's own programming language and theorem prover: the code, the behavior, and the emission curve. A formal proof was also written that the transfer function cannot be violated: the total number of tokens in the tree is always conserved across any transfer between wallets. Not a test. A mathematical proof. Parts of Kind's standard library and additional proofs were contributed alongside this. A community of over 30 Kind students formed around the project, taught through regular sessions and guided to build game projects in Kind, which was intended to be the contract language for the chain.

The chain never launched publicly. Testing was internal: Fibulacci, a Tibia-like MMORPG running in real time on the chain. It worked. And the more the team explored what HVM could actually do, the bigger the questions became. Kindelia Foundation wound down in early 2023. Higher Order Company was founded the same month, with a more refined runtime and a much larger set of questions to answer.

UwU Games is where the story begins. The company was co-founded by Victor Taelin and Vitor Sipher in early 2021, after an accidental meeting: Taelin wanted to teach Sipher to code on Kind, his newborn programming language, and Sipher had a clear vision for a gaming company. The pieces fit. They decided to build it together.

The early phase was pure building. A team was assembled around the idea of a super star Brazilian Kind team developing the first game, an untitled project known internally as Grid Game. Large parts of it were coded from the ground up: connectivity systems, character building, frontend integration, and rendering. As the team grew, the business needed structure. In a team meeting, Sipher proposed to lead the company on the business side. The team agreed: bank accounts, accounting, operations from scratch. A first step toward everything that followed.

The turning point was Victor's progress on HVM. As the runtime proved what it could do, the question became unavoidable: keep building the game, or focus entirely on HVM? The team agreed that building Kindelia on top of HVM was the more important bet. UwU Games was put on hold. Kindelia Foundation opened the same period.

Years later, a 200k USD round was raised and UwU Games came back to life. The day-to-day passed to Ahalan "dasij" Bessa, who now drives the project forward. Sipher is now in an advisory role, covering cap table structure, token distributions, legals, and accountability alongside accountants and lawyers. A further 10k USD was raised to fund game development. The relationship today: available whenever dasij needs.