# December 28th 🟢 Community Call Summary


## tldr!
This was a momentus past year for the Cyber project and community! A wave of verifiable collaborative learning has begun that now has expanded to both the *ru* & *en* communities. We experimented to create a system that rewads learning! Some of the contributions aiding this process were:
- cy: the ultimate cli tool for collaborative learning
- spacebox: cybergraph indexer
- basic registries for hub
- onhcain-registry: trusted data source for cosmos tokens
- community boost with cyberacademy
- cybercongress: new incentives for teams
- many improvements made on cyb

Next year will be even more interesting
- more features and upgrades to bostrom
- unforgable inference system
- cybernet deployment
- metagraph of cyber
- redesign of cyb

The plan is simple... fit our revolutionary product to help shape the biggest market on the planet: [decentralized ai](https://cyb.ai/oracle/ask/QmVcAr1wVdL17GfA5PXu9fHHk6NrpoWsxnb861P7CjoHbk)!

## Review of results for 2023


### Tokens

To start with some cheap thrills, the cyber token economy experienced some volatility. During the last month of the year, there was an extreme spike in price followed by a week long sell-off to the original price. This may seem disheartening in the short term, however when we consider the broader picture, BOOT has seen a growth of 50% since the beginning of the year. As for the rest of the ecosystem, Hydrogen has remained quite stable while Ampere and Volts have seen approximately 70% gains each. Considering the tokenmomics, an increase in demand for Amperes and Volts is a great sign of growth.


### Bootloader

One thing we like to track here at Cyber is the negentropy, or negative entropy (also called syntropy). As you may have guessed, this basically means the system is less chaotic and more connected which means energy can flow freely and not leak out of the pipes... This year we saw an accumulation of more than 1 million negentropy, representing nearly an 8x increase in one year. This is quite remarkable and, not surprisingly, correlated with the cyber graph's growth in valuation.

Analysis is great, but the most notable development this year is the substantial growth in cyberlinks,. Starting from a bit more than 100,000 Cyber Links, we have now surpassed half a million Cyber Links. This is truly remarkable, especially considering that the average value of a cyberlink is around $1. This value is significantly higher, perhaps 20 to 30 times or even up to 100 times more than the actual cost of a cyberlink. The most significant aspect of cyber graph's progress this year is the enhancement in the quality of links.

### Moon network state

In terms of the Moon Network State, although it wasn't our primary focus this year, we still achieved noteworthy milestones. We experienced a 4x growth in citizens, reaching 32,000 passports.

In the realm of transactions, we grew threefold in terms of the total number, reaching almost 3.5m transactions this year. Notably, we made significant strides in the [distribution of Genesis BOOT](https://cyb.ai/gift) supply this year. Approximately 80 TBOOT were claimed, constituting a substantial portion of the overall gift, around 10%. This redistribution is a positive development.


Internal liquidity also saw substantial growth, which is crucial for a successful Network State. We now boast around $10,000 in stable daily volume on the [internal DEX](https://cyb.ai/teleport). On average, there is approximately $1 million in total value locked on the internal DEX, indicating a robust internal liquidity situation.

Moreover, there has been usage of programs and smart contracts, with currently 200 contracts instantiated, representing twice the number compared to the previous year. This demonstrates an increasing adoption of smart contracts within the Moon Network State.

### GPU Hub

We've observed substantial growth in GPU memory, approximately a 5x increase. Successfully running two networks, cyberpussy, we have nearly 90 active GPU operators. This number has remained relatively constant throughout the year, showcasing stability. Despite losing around 40-45 operators, the quality of operators has improved, and new ones have joined.

The most intriguing development on this front is the significant advances in GPU computation and researching consensus. Notably, we now have a very stable [GPU library](https://github.com/gfx-rs/wgpu), laying the groundwork for potential GPU execution on top of the cyber graph consensus. Although the numerical growth may not be extensive, the progress in terms of research and development is indeed significant.


We developed [basic registries for hub](https://cyb.ai/oracle/ask/QmV2mNchi7tUdEXVaaNbaamDn7ySjuXCNk5KkfMVAWaxwq) which already supported experimentally in teleport.

[cyberdbot](https://github.com/Snedashkovsky), introduced an [onchain token oracle](https://github.com/Snedashkovsky), serving as the foundation for Hub functionality.

These developments underscore the continuous growth and expansion of the project's capabilities.

### Congress

This year, we successfully implemented a mutual evaluation system, leading to significant improvements in the quality of internal development and the flexibility of team management. With this system in place, we believe we are halfway towards achieving real sustainability for the project. While we can sustain the project using internal tokens, it's crucial to note that we are still working towards achieving deep liquidity for full sustainability, marking a substantial step forward.

Currently, Cyber Congress manages 900 THydrogen assets, equivalent to around 4.5 million dollars at current prices. Notably, 98% of these assets are in internal tokens, showcasing the project's reliance on its own ecosystem. An interesting fact about Cyber Congress is its position in the [DAO ranking](https://daobase.ai/leaderboard/all/all). Although not explicitly ranked, when considering the major part of the more than 100,000 DAOs indexed by DAO base, Cyber Congress is among the top 150 globally. Additionally, with the inclusion of the Great Web Foundation bag, representing around 5% of the Bostrom Network and Cyber, we break into the top 100 DAOs globally, marking a significant achievement.


### cybernet

Approach for learning incentives was identified this year, a significant step that hadn't been achieved during the last seven years of research. Credit goes to [BitTensor](https://bittensor.com/) for introducing this approach. Several new projects have already been initiated, and in the past two months, significant progress has been made.

A considerable amount of work has gone into researching BitTensor, and the [cybernet](https://github.com/cybercongress/cybernet) project has been crafted as a smart contracts for cosmwasm, set to be deployed shortly. It's already partially operational in space-pussy.

Recent upgrade of spacepussy has expanded the capacity of programs that can be deployed on the chain. This enhancement was necessary due to the substantial size of the cybernet contracts. Now we are working on cybernet integration with default mode network to provide mechanism for scheduling of rewards.

Speaking briefly about BitTensor and Cybernet itself, BitTensor is a compelling project that develops powerful functions for distributing rewards in civilized networks and communities. The underlying mathematics are straightforward and based on stakes and weights that participants in subnets and communities allocate to each other. This allocation is made according to the assumptions about the work within the community or subnet.

BitTensor follows the path where communities create subnets, and these subnets undertake various tasks. For instance, they can learn neuron networks, process inferences of these networks and models, or engage in other forms of computational work. Subnets have the flexibility to perform tasks such as image inference, data analysis, and proofs of data storage. This wide range of options allows for diverse activities within any subnet, involving nodes that participate in and contribute to computations inside the network.

Within each network, participants assign weights based on evaluations of the work done by others in the subnets and communities. With this data, the algorithm can effectively distribute rewards and allocate trust among participants. This intricate system ensures a fair and efficient distribution of rewards based on the contributions and evaluations within the network.

In the past month, the [subtensor](https://github.com/opentensor/subtensor) bussiness logic was ported from substrate palets to cosmwasm, and the entire [bittensor](https://github.com/opentensor/bittensor) codebase was migrated as [cybertensor](https://github.com/cybercongress/cybertensor) to work seamlesly with cybernet smart contracts.

### cyb

On cyb.ai [sigma team](https://cyb.ai/@sigma) made significant updates to [oracle](https://cyb.ai/) and [teleport](https://cyb.ai/teleport). In addition to user faced features we made huge amount of refactoring to simplify further developments. Several side effects of this refactoring include:
- embeded [graph database](https://www.cozodb.org/): cozodb is targeted to a new generation of client side ai apps
- [rune](https://rune-rs.github.io/) experimental scripting and particles api for dynamic programming clientside
- migration to [helia](https://github.com/ipfs/helia) from ipfs-js which significantly improves reliablity of content delivery

We are gradually transferring the site design from the first version to the second, making gradual improvements along the way. Here's the initial version of the design, which was both amazing and minimalistic:

We've been enhancing functionality and improving the user experience. For instance, the filter function, initially hidden and undeveloped, is now available, accompanied by metadata for each particle at the top. We've also made improvements to the feed search results, now featuring metadata about particles, particle direction (incoming, outgoing, or both), and content.


The second version of Teleport has been delivered in this version, we introduced the Send, Bridge, and Swap functionalities. Messaging capabilities have been incorporated into the swap feature, allowing the exchange of messages and particles with tokens. This functionality will be available in the new sense soon.


We've also made improvements to the main page. From here, you can access the search function and view the top current particles. Additionally, there's a learning section where you can explore how to learn and check it out.


### cy

Another notable addition is [Cy](https://github.com/cyber-prophet/cy) by [cyber-prophet](https://github.com/cyber-prophet), which has evolved into a highly functional tool over the year.


It has become a leading learning tool within the ecosystem, significantly contributing to the understanding and development of the network. A year ago, we didn't have this tool, and its development has played a crucial role in the recent spike in learning within the network.



### spacebox

Over the last year, and in the preceding one, [bronbro](https://bronbro.io/) have served as a significant relayer provider for the Bostrom chain, facilitating approximately 99% of all transactions on Bostrom. Notably, we developed the space-box indexer for Bostrom with support for all Bostrom modules. This endeavor was funded by the community pool of Bostrom, showcasing our commitment to contributing to public goods within the ecosystem. The hope is that this indexer will enhance analytics and understanding of the project's movement in the future.


Additionally, Bronbro has emerged as a major API provider for both the Cyber Congress and the Bostrom project. Despite facing challenges following the surge in the popularity of Cyber gift, our team is dedicated to ensuring that apps and functionalities work optimally for all users, striving to provide the best possible experience.

[Spacebox](https://github.com/bro-n-bro/spacebox) indexer is a amazing product emerged this year. Spacebox allow us to answer questions related to engagement, cohorts, and behavioral analytics that were previously challenging to address. It's worth noting that the broader Cosmos community may not fully appreciate the power of blockchain indexing, and we express gratitude to bronbro for the significant progress made in this area.

### CyberAcademy

The collaboration of [CyberAcademy](https://github.com/Snedashkovsky) on community coordination marks an essential step in the project's development. Thanks to the ongoing effort of CyberAcademy we are now have thoughtful, engaging and ever expanding community.

cyberAcademy helped us to shifting from a primarily Russian-speaking community to an English-speaking. This shift have occurred recently, just a month or a few weeks ago, marking a crucial step forward for the project. Breaking out of the Russian-speaking circle, which had been the norm for years, is an essential development. Gratitude is extended to everyone who participated in this exciting collaborative learning experiment, contributing to the success of these endeavors.

## Plans 2024

### Metagraph of cyber

The project is referred to as a metagraph, essentially a graph-based description of cyber project and key componenets:
- cyber: protocol for collaborative learning
- bostrom: bootloader for cyber
- cyb: immortal robot
- age of superintelligence: the game
- cybernomics: foundational knowledge of cyberverse

Metagraph will serve as a key entry point to grasp all the concepts. It can be envisioned as documentation and specification, but presented as a well connected graph, which allow cyberlink it to cybergraph, teach any llm to answer question about cyber or implement cyber using any modern stack.

We believe metagraph is critical missing piece of cyber project so we aim to publish it asap: in several months or so.

### Bostrom and go-cyber

In 2024 we plan several maintanace upgrades, feature upgrades and - extensive research around llm and webgpu execution in consensus.

Maintance upgrades will inclide migration to stable cosmos-sdk, comet-bft and ibc dependencies.

Feature upgrades will include the following:
- [standard inference](https://github.com/cybercongress/go-cyber/issues/694)
- turbowasm
- token factory
- interchain bundle
- graph proofs
- karma algorithm

The development of better inference functionality, referred to as [standard inference](https://github.com/cybercongress/go-cyber/issues/694), is deemed crucial. Testing results of a new algorithm have shown a remarkable improvement—approximately 100 times better inference results and this is just the begining. Deploying this improved algorithm to production is a high-priority task. The new algorithm will account balance of V token for every context search. This promises a fresh search experience, considering both global ranking with A token and contextual information with V token.

Eventually we want to migrate almost all modules from cosmos-sdk to cosmwasm.
The model responsible for this is turbo-wasm. The new module will allow the deployment of smart contracts, functioning as native modules. The idea is simple: progs can execute not only by means of transactions, but also through calls at the beginning or end of a block, thus acting as native modules. Turbo-wasm contracts will facilitate dynamic updates, ensuring the prompt delivery of new features and faster release cycles without a need for consensus upgrades just based on governance decisions very similar to how substrate based chains updated.

Furthermore, we plan to extend our capabilities related to tokens as a lot of the games we want to implemet depend on a usable token functionality. This includes a model that allows communities to deploy native tokens without necesity to index them as cw-20 or cw-721 tokens, but with bindings to cosmwasm. Subsequently, our goal is to deploy a model for interchain NFTs and integrate these NFTs with particles and graphs, enabling the minting of NFTs based on community-contributed information.

Expanding the Hub's functionality is also our top priority. In the core is interchain bundle which include 2 modules:
- [interchain transactions](https://docs.neutron.org/neutron/modules/interchain-txs/overview)
- [interchain queries](https://docs.neutron.org/neutron/modules/interchain-queries/overview)

Coupled with autonomous execution this will create solid foundation for crosschain ai applications.

An exciting feature in the pipeline, prompted by community requests, revolves around graph proofs. This extends beyond existing graph proofs to encompass massive graph proofs, allowing the verification of information added to the graph based on hash and Merkle tree paths. This enhancement underscores the increasing value of bostrom's cybergraph.

Work is currently underway on implementing a tensor algorithm that includes the computation of trust based on weights and stakes. This computation occurs on a global scale within the graph itself, without manual weight assignments. Trust and rewards will be calculated based on Cyber Links and the graph structure, with weights determined by common Cyber Links between pairs of neurons. This approach allows for the computation of weights for every neuron, creating side links and determining trust. The resulting trust will be transformed into a new Karma system.

Currently there are two things on our moonshot r&d pipeline
- [llm execution](https://github.com/microsoft/onnxruntime)
- [webgpu execution](https://www.w3.org/TR/webgpu/)

Exploration is underway on how to launch models, especially large language models, in a consensus deterministically. Several technologies show promise in potentially enabling this, and cybernet subnets will play a crucial role in learning and guiding models based on the community's interests. Once these models are trained, they could be deployed in consensus, bringing a layer of trust to the inferences made by these consensus groups.

[LLM execution](https://github.com/microsoft/onnxruntime) in consensus is not the only vector of research in terms of ai execution. We tremble at the idea of [webgpu](https://www.w3.org/TR/webgpu/) execution in consensus, reflecting the ongoing commitment to pushing the boundaries of technology.

As we move into the next year, these upgrades and innovations represent our commitment to staying at the forefront of decentralized ai technology and meeting the evolving needs of our community depending on what could emerge on the horizon.

### cybernet

The deployment of cybernet to bostrom will introduce a new reward system and adds a reputation system based on [shapley valeu](https://en.wikipedia.org/wiki/Shapley_value).

- cybernet
- cybertensor
- verses

Looking ahead to the next year, in approximately two months, a game is set to be launched as deployment of [cybernet](https://github.com/cybercongress/cybernet/) in spacepussy. This will allow to access rewards, collect feedback, and facilitate learning within the community. After around 3 months of testing in spacepussy we are going to deploy cybernet to bostrom with `O` (oxygen) token.

This significant upgrade will introduce new dynamics to the project, as neurons and the community will receive feedback from each other regarding information, activities, and contributions to the cybergraph.

Moreover, in the past months, Python SDK and CLI tooling were implemented in [cybertensor](https://github.com/cybercongress/cybertensor). This development enables users to develop and deploy their own subnets to the community. Initially, cyber Congress will deploy the first subnet, specifically tailored for upcoming games. As Cybernet progresses, more paths will be explored and developed. Communities, based on shared interests, will have the ability to group together and assign weights based on activities, distributing rewards to one another manually through assigning and waiting.

Moving forward, the next phase involves subnets that will reward nodes in various groups for different types of work, as previously explained. This strategic approach ensures a systematic expansion of Cybernet, allowing for the development of diverse communities that can collaborate based on interests and shared activities.

As more trust is required for the inferences and answers provided by these models, economically viable consensus mechanisms will become increasingly important. This introduces dynamic considerations on how these models will be distributed and which models the community will find interesting.

The next phase involves deploying a global narrative platform that guides models within consensus AI groups. This platform allows them to act autonomously and pursue their own subject goals, generating transactions processed by AI consensus groups and the community. Community members and neurons can communicate with these consensus AIs, providing feedback based on their interests and the narratives they want to see unfold. Additionally, there will be an emotional system, featuring specialized tokens for emotions, providing another layer of feedback for AI consensus groups. But this is likely developments which goes beyond 2024.

In the coming year, a new reputation system will be released, based on this trust, and global rewards will be distributed. This signifies that Cybernet will not only reward community subnets and consensus but also contribute to the global layer of the network graph. This marks the beginning of the journey toward global simulation and the acceleration of superintelligence.

### cw-cyber

cw-cyber represents a smart contract layer. Smart contracts are programs aspire to deploy a wide array of different things, and many aspects pique the community's interest:

- energy reform
- social layer
- extended hub registries
- incentive for cybergift
- warp migration to cosmwasm
- daodao
- research on ikp and cybergraph sdk

Major change planned for the next year is what we call "energy reform." This entails a rework of the economics surrounding H, A, and V. For H, it involves a slight upgrade of the economics. However, for A and V, especially in terms of mint, send, burn, and lock functions, a complete overhaul is envisioned from the ground up. The primary focus will be on the staking of specific particles and links, enabling users to customarily adjust weights in the graph and earn staking rewards from these actions.

On the social layer, the existing following mechanism is under scrutiny, and a search for more innovative and economically sound solutions is underway. The goal is to design this layer in a way that enhances the overall interest and engagement in social activities within the platform.

We aim to upgrade registry contracts, enhancing their functionalities and the overall Bostrom experience as a hub for different networks.

One of our goal is to finilize distribution of BOOT tokens. Currently 80% of the bag remains. To facilitate the process we are working on a referal system that will allow to reward participants who helped to bring the audience targeted by the gift.

A decision has been made to transition the existing dex, which originated from the Genesis as a liquidity model developed for Cosmos Hub with MEV protection. Although powerful, we have realized the need to move as many components as possible to contracts and CosmWasm modules. As part of this transition, the Dex will undergo migration to CosmWasm, aligning with our objective to centralize functionalities within contracts.

One notable area is [DAO-DAO](https://github.com/DA0-DA0), which enables the community to establish DAOs and engage in coordination with projects, information creation, and knowledge domains, among other possibilities. The powerful DAO tooling within the network facilitates various functionalities. Last year, DAO-DAO underwent an upgrade to integrate seamlessly with Native tokens, a highly anticipated development. The community eagerly anticipates the deployment of this functionality, encompassing tokens and DAO-DAO contracts, as soon as possible.

There are graph tools in development that will empower neurons or communities to create advanced semantics of cyberlinks. A transformative process that will enhance the graph's capabilities and introduce intriguing dynamics. The next stage involves taking the initial steps with the IKP protocol. Our interest extends beyond the graph within Bostrom. We recognize the potential for deploying contracts across various Cosmos networks supporting CosmWasm, Ethereum-based networks with EVM, and different Layer 2 chains. This set of contracts will empower users in these networks to create cyberlinks and contribute information to the network. Subsequently, the bridging and relaying of subgraphs across different networks will occur, leading to the indexing of this information and initiating the expansion of the graph to different networks.

### cyb and cyb-ts

The strategic goal for cyb 2024 is to significantly improve cyb.ai user experience incliding mobile and desktop support. Although there are features we think are also critical for retention:

- new ux
- mobile
- desktop
- sense
- brain
- cyberver

There's ongoing work on a new feel for Cyb, with a conceptual slicer available in [Figma](https://www.figma.com/file/7i0Ly3YF587km0F8iDZod4/cyb?type=design&node-id=23786%3A1910&mode=design&t=pB18KzvF2ViRcp3u-1), providing a visual representation of the robot with gamification concept that the team finds exciting.

The team is currently engaged in developing Sense, an on-chain communication platform that resembles Telegram but offers guarantees of delivery and cryptographic assurance of communication—a crucial aspect.


Another important focus is delivering a working brain. Currently, it's challenging for power users of Cyb to explore and navigate through the content cyberlinks in their graph, emphasizing the need for a compelling tool for discovering linked files.


The third goal is to deliver a cyberver application for [cybernet](https://github.com/cybercongress/cybernet/).


For the major ux upgrade, we're planning a new interface that encompasses the essential characteristics of your site robot, including sigma, sense and your brain from one consistent interface.


We believe that 2024 developments design and cyb-ts will help us to lay the foundation for a new, efficient cyb implementation in rust.

### Product market fit

The primary goal for the upcoming year's development efforts is to achieve product-market fit. Positioned at the outset of the next significant growth phase, the aim is to make the project more comprehensible, usable, exciting, simple, and well-documented. This focus on refinement and clarity is crucial to navigating the upcoming growth cycle successfully. The challenge of reaching product-market fit must by personally prioritized among all teams and community, and the confidence exists that, with the ideas presented in this postthis milestone will be attained.

### More!

Looking ahead, there are plans within the team and community for the coming year. The intention is to expand the Cyber team by bringing on board four more developers: one Go, one Python and two Rust developers. Simultaneously, efforts will be directed towards creating comprehensive and robust documentation, ensuring an advanced and powerful developer experience. Initiating the developer relations process is also on the agenda, facilitating the project's accelerated growth.

In 2024, the plan includes participating in various conferences, such as Cosmoverse. Additionally, there's an interest in engaging with some intriguing AI conferences during the summer. Further updates on these plans will be communicated, and the goal is to host both online and offline hackathons to encourage participation and collaboration.

We are thankfull for everyone who support us and our ambitios and crazy ideas!

## Join our community!

📢 Don't forget to join our community! Follow us on [Discord](https://discord.gg/ARwv74ZyGH), [Telegram](https://t.me/bostrom_news), and [Twitter](https://twitter.com/cyber_devs) for the latest updates and discussions. Let's build a connected and informed community. Create [cyberlinks](https://cyb.ai/) and [contribute](https://github.com/cybercongress) to the network! See you there!