image Starting into 2025 MOCA has dedicted themselves to move on from building web3 native SaaS products into the world of open source software development. From technical pov that means our new museum codebase can be deployed by any museum or enthusiasts whearas MOCA has to see themselves as a "customer" to their code base. Simultanously we guide the development of functionality we think is useful for digital art museums. Learnings over the last years But lets go back in time for a bit. We've built cool stuff over the last years. From the Multipass to multiplayer curation stacks like MOCA Show up to the metaverse exhibition product MOCA ROOMs we always tried to build tooling that lifts up artists and collectors by giving them tools to fuel their crypto art driven storytelling. Me back in 2022 talking about the stuff we build at MOCA Our team has learned a lot building all of this out and we'll make sure that our open source museum tech will inherit the most powerful facets. At the same time we're gonna lean heavily into agentic systems and other AI tech stacks which can be fueled by both cloud and locally hosted AI system which we enable via LiteLLM routed AI inference.

I joined the Museum of Crypto Art (MOCA) as CTO in early 2021. It was when Colborn Bell was seperating from his co-founder Pablo Rodriguez-fraile. The new mission was to truly decentralize the museum and I loved that spirit. Back in these days MOCA had few builds in Somnium Space and an incredible community of artists - but no tech stack. We teamed up with untitled, xyz who became the inhouse architect. He created the Agora which would become one of the very first interoperable exhibition spaces of the museum. <img src="https://hackmd.io/_uploads/r14Io4Bch.jpg"/> We started to envision how the future of the museum could look like and released the MOCA Manifesto as the incorporation of a non-profit foundation which would become the new vehicle allowing us to operate the museum on its path towards decentralization. The governance token was launched to be partly distributed via airdrop and to patrons who donated art into the museum. A year later, when we had 2 products live (Multipass + Show) and our flagship product MOCA ROOMs was in the works. Anyone could create their user profile, aggregate artworks into the MOCA Community Collection and create multi-player exhibits for free. At the time Fortune did a profile about MOCA which was cool but our flagship product that would push our vision to the next stage was not there yet. Lemme explain what this leap means for MOCA in my own words. Also check the MOCA Live Podcast by Max Cohen it's awesome. Get your weekly dose of cryptoart via Spotify and then we'll dig into the underlying technology.

In 2020 before NFTs went nuts, I decided to start an experiment and summoned the tokenized identity called Nature. I'd start to create modular, interoperable, web3 native IP that would aggregate value by growing provenance and influence across web3 platforms as well as traditional social networks. But what does all of this even mean? The Origins of Nature On the 7th of October in 2020 the avatar artist CornyNachos minted "Nature is on the TV" avatar (see photo below) on Cryptoavatars a platform that solves creator verifications for VRChat avatar artists which was (and still is) a problem over there. A week later I won the auction and obtained the vrm avatar which started a multi-year journey with lots of tinkering and applied research at the intersection of XR x Web3 x AI where I wrote various stories. fluctuo Watch early footage of the Fluctuo gallery from Somnium Space in 2020 Back then, when vtubing was slowly becoming a thing and more purely digital influencers came to social media, something was clear to me: The future of content production and entertainment will massively change, especially with the ability to spawn virtual selfs into all kinds of scenarios in cost efficient ways.

This summer I've been building a new piece of hardware which now is the heart of my homelab. The Minisforum MS-01 Work Station runs Unraid and is hosting various applications while leaving enough headroom for all kinds of experiments. The system has come a long way from its original form, increasing overall performance and capability. <img src="https://hackmd.io/_uploads/SyPEc-7hC.jpg"/> Considering its size, the specs of this small computer are pretty insane: Intel Core i9-13900H 96 GB Ram (2x48 GB Crucial) 2x 1TB NVMe (one used for Parity) Nvidia RTX 2000 Ada GPU (16 GB VRAM)

We started working with Amber Island two years ago and have been building though a brutal bear market. Our crew at Qwellcode creates the tech side and teamed up with the amazing folks at GHARAGE Vision Hub which is subsidized by their parent company Gebr Heinemann, a family-owned travel retail business founded in 1879. They have massive experience and an incredible network in the physical realms which enabled us to bring Amber to the market. Collectors around the globe struggle to find secure ways to buy, store and trade rare whisy. Deals on forum boards and social media chat groups in Facebook for example invite scammers to sell faulty bottles (e.g. relabeled, wrong liquid, etc) to honest buyers who want to invest into these scarce assets. Amber Island provides both a community and platform that offers special whisky in direct cooperation with distilleries and brands. Physical bottles from Amber are securely stored in professional fashion while investors trade ownership certificates without friction. Owners can decide to redeem their bottle NFT for the actual physical whisky which flags the NFT onchain, changing their 3D appearance forever. Melting Digital and Physical Amber Island is all about melting the digital with the physical realm. Experiments like 3D printed bottles that were showcased at community events clearly demonstrate the love that the Amber team is putting into achieving this goal. We've been building through a tough bear market and believe that the high quality is going to convince both the web3 and the whisky space. <img src="https://hackmd.io/_uploads/BknSjvKnn.jpg">

In this guide I'll explain how you can train a LoRA of your VRM avatar that you can use to draw images of your 3D avatar in any Stable Diffusion Model that you can find. In this Tutorial we are going to train the LoRA for the tokenized identity Nature and I'm going to provide the file so you can play around with it in your own Stable Diffusion installation. The images below give you a sense of the prompt results. Find vtubing content on my Youtube Channel to get an impression about the actual 3D avatar. A LoRA captures and conserves a concept or an idea in a way that it can be aggregated into larger models to be part of its outputs. A LoRA can be anything but in this case it is a character. <img src="https://hackmd.io/_uploads/B1J2DEqk6.jpg"/> The SD1.5 LoRA that we've trained in this guide was released on CivitAI under cc0 licence. Play with it for free in your own stable diffusion instance, we added lots of example prompts. Add your creations on CivitAI and tag Nature on Twitter so that she can retweet your social media postings. <img src="https://hackmd.io/_uploads/BylRvotyp.jpg"/>

I recently upgraded my desktop with an AMD Ryzen 7950X3D and 64GB DDR5 RAM (6000 MHz) to push the performance for both productivity and gaming on par with the Nvidia 4090 FE that I got earlier this year. The system sits in a Sliger SV540 - I fell in love with this mini itx case a while ago and it wasn't an option to get a bigger case as I travel and use it in different locations. After installing all components on my new ASUS ROG STRIX X670E-I GAMING WIFI and doing the initial config (bios update + config, driver updates , fan curves, cpu+gpu undervolting, etc) to ensure everything runs smooth I realized unusually high chipset temperatures sometimes passing 100°C during stress tests. In one test this even affected the primary M2 drive (wd black sn850 4tb) as temps passed 90°C and the drive stopped working temporarily after throwing a blue screen. <img src="https://hackmd.io/_uploads/r1F4b6_4T.jpg"/> This post explains measures on software and hardware level (X670E thermal pad + copper shim mod) that helped me to solve these issues. The system now runs very reliable. Results may be way better for Micro-ATX and ATX builds especially due to additional case fans that you can install. Improving Temps of X670E Chipsets While digging deeper into the matter, I learned that the X670E chipset can have thermal problems depending on various factors. There are several ways to improve the situation and especially for a low volume case it requires a mix of various things to tackle this. Also I use two M2 drives which - combined with the hardware mentioned above - is quite challenging for the AM5 chipset as both are located directly above the chipset. But lets explore what can be done to reduce temps:

This year we've been building a very special project that launched this August. The project is called zeroone and allows artists to create and distribute art for free. As you create on daily basis, you are recharging your mint pass allowing you to collect up to 10 pieces on the platform. Users don't have to own any crypto and they don't have to use Metamask oder other web3 wallets as we provide novel account abstraction technology which enables login via email or phone. The platform focuses on three things: Create, Collect and Connect. Since we've originally started development it was our highest priority to facilitate these functionalities by removing any gas fees from the artists and collectors. This post explains how we've achieved exactly this. Zeroone is led by CEO Ludovica and was originally imagined by Colborn. I worked very close with both of them on though processes around functionality and tokenomics from the start. During beta testing which took place on Avalanche testnet we received lots of positive feedback. We did some further finetuning around the mvp before it was deployed to Avalanche mainnet. <img src="https://hackmd.io/_uploads/Syi2qrpp3.png"/> Dune Analytics (2023-08-30) https://dune.com/kleintonno/zeroone

Over the past few weeks our team at Qwellcode has designed and executed the technical side of the MOCA Fundraiser. In this massive event the Museum of Crypto Art facilitates the largest Cryptoart collaboration to date with over 700 artists involved and our crew ensured that the technical sophistication is on par with the overall collective effort around this fundraiser. What follows is an explainer on how we tokenized an entire collection of 713 individual art pieces into a single NFT to distribute that into the hands of anyone willing to support MOCA. In August 2023 the Museum of Crypto Art mints their fundraiser open edition. It's a mosaic that consists of over 700 artworks (1 submission per artist, each 500x500px) and while this sounds like Everydays: the First 5000 Days by Beeple we decided to go the extra mile to ensure that: each of the over 700 artists is properly credited onchain all individual art pieces can be explored stand-alone on any device the artwork dynamically re-arranges in randomized fashion every time it loads all submitted art pieces are highly interoperable across the web3 ecosystem

Introduction I've been researching and tinkering a lot with locally hosted LLMs recently. There are several great tutorials out there which explain how to run LLaMa or Alpaca locally. After digged through a couple of them I decided to write a step-by-step on how to run Alpaca 13B 4-bit via KoboldAI and have chat conversation with different characters through TavernAI - entirely on your local machine. The performance of the quantized model loaded on the GPU is incredible and shows the potential for on-prem LLM systems. I'm aware of CPU based solutions like alpaca.cpp and played around with them. However in this guide I'll dig how to install the Alpaca that I personally like most on gaming hardware. It runs incredibly well on my RTX 4080 - see the video below. What makes this stack special? The ability to run this setup locally on gaming hardware is pretty neat. It amazed me for the same reasons Stable Diffusion amazes me. The modularity is another reason. You can configure the language model interface in KoboldAI and plug that API into other frontends: Instead of TavernAI you could embed it into Hyperfy, Webaverse or other web3xr platforms. We already saw ChatGPT integrations in Hyperfy and the Webaverse Character Studio already showed very powerful AI integrations. I hope this guide helps you to understand the modularity aspect. I'm currently exploring the Langchain framework which is going to allow the creation of more sophisticated LLM systems that are open and can be hosted on premise.

In this document I'm going to explain how you can setup an extremely sophisticated VTubing Rig that costs you almost nothing but brings incredible results to the table. The setup consists of two main components. The final result can be recorded or streamed into Meet, Zoom, Teams, etc. ::: info While XR Animator receives and computes tracking from your webcam feed to stream it via the VMC protocol, VROOM visualizes your avatar with that data in beautifully inside all sorts of 3D environments. ::: Prepare your Windows PC You don't need much to make this happen. The software is available for free but I highly recommend you to support the developers especially the indie devs @butz_yung and @ojousa_ma_yo who are the creators of the blocks that this tutorials focuses on. Get yourself a regular webcam

Inspired by M3: How to git gud I'm writing this document to reflect on my existing git gud journey and to learn more about HackMD as a platform for recording and sharing my experiences, knowledge aswell as future progression in an open and structured way. It's the beginning of my journal. 1. Mindset The past few years were quite exhausting but very fruitful at the same time. I've joined the web3 space in late 2017 when I found Decentraland and visited my first Ethereum conference (EthCC in Paris) in February 2018. One year later at ETHDenver 2019 I joined Patricio, helped him to build the first version of POAP (frontend code, design) which we deployed to Ethereum Mainnet. Over the past 5 years it was always about building while connecting with like-minded people in the space. It was about finding the right tribes while positioning the devshop Qwellcode that I co-founded 10 years ago inside the web3 space with long-term ambitions on our mind. We were among the first NFT builders on Polygon which allowed us to gain an understanding of how-to build accessible multi-chain EVM dapps early on. Our greatest achievement so far is the MOCA product ecosystem which I as CTO designed and our devs at Qwellcode have implemented. In parallel I soaked up any Metaverse related stuff that I could find while expanding hands-on research around various technologies with a focus on avatars and identity. I did read Ready Player One back in 2017 months before the movie came out and finally started to read Snowcrash during an Ayurvedic retreat in India earlier this year which was great timing.