Opening dialogue
Author: BeeBee beebeehop@protonmail.com
1.
How to make a digital intelligence? For the leaders of cloud computing, the creation of an artificial general intelligence (AGI) in this decade, and soon after, an ASI (artificial super intelligence) is seen as inevitable. Their favorite path to take us through this defining moment in time and society is using a centralized AI using deep learning neural networks called Large Language Models (LLMs) that are multi-modal, agentic and have reasoning. With more compute and data (human and synthetic) being thrown at winning the race. The early LLM releases have seen fast adoption and limitless venture funding supports this innovation bet. This form of AI brings all the data to a data center where GPUs (graphic processing unit, hardware that makes graphics work on a computer or mobile phone) to perform commodity parallel computes in lengthy training runs that use vast sums of electricity. From the invention of the perceptron ( https://en.wikipedia.org/wiki/Perceptron ) and the subsequent building of neural networks of increasing depth, the understanding of how they work has eluded science. It is uncertain where this path leads.
Better foundations for a digital intelligence are possible:
2.
Embrace that intelligence is embedded already in many parts of nature, cells to forests to geological forces to the human mind.
3.
Keep data sovereign, decentralized and establish a culture of data ownership.
4.
Orchestrating networks of tiny agents is a better way to answer questions or solve problems and cohere common sense.
5.
Establish trust peer to peer using coherence ledgers.
6.
Empower communities to decide how they use their energy and compute.
7.
Manufacturing becomes small, decentralized and designed to prevent, repair or rejuvenate, from cells to an ecosystem.
This opening dialogue puts forward a path to Gaia Intelligence designed around coherence ledgers and decentralized machine learning protocols inspired by evolution to solve problems at all scales peer to peer where society decides the pace, shape and birthing of a digital intelligence.
1.
Gaia Intelligence - inspired by nature, everywhere
The natural world is shaped by many forces, for example the weather, the water cycle, clouds to tiny water droplets that fall as rain that form streams, tributaries fill rivers that sculpt mountains to landscapes and provide minerals for soils. Envision such a picture but where water is replaced by bytes of data. Gravity provides the guiding force for water, in a digital world, values align data flows. This section sets out the values and design principles of Gaia Intelligence.
The natural world is shaped by many forces, for example the weather, the water cycle, clouds to tiny water droplets that fall as rain that form streams, tributaries fill rivers that sculpt mountains to landscapes and provide minerals for soils. Envision such a picture but where water is replaced by bytes of data. Gravity provides the guiding force for water, in a digital world, values align data flows. This section sets out the values and design principles of Gaia Intelligence.
Societies have used ledgers to ‘account’ for their actions or value exchange over the ages. A new design of ledger is required that provides a skeleton for information just like Fibonacci numbers provides an information structure for nature. A coherence ledger. A new design that enables a collaboration of humans and communities with networks of computers and algorithms to attain coherence. Sometimes in place, the physical location of both will be local and at other times the relationships will be more ethereal.
The more people active in shaping an AI the better and this includes sensors embedded in nature and the environment. A broad diversity of thinking, design ideas and use cases need to be included. Technology adoption starts usually with those with specialist technical skills, this no longer has to be the case. The proportion of humans participating in our technology agenda is growing but usually as consumers. They need to be given the tools to become prime actors. While it makes the entry standard much higher and takes more time to refine use and showcase new capabilities, Gaia Intelligence requires its tools to be widely accessible at all times and from the start. Over time, this will become seamless and real time.
The initial use of the word computers was, usually, women performing paper calculations but today a computer is a digital device performing calculations, from large data centers to mobile phones and embedded in microcontrollers as sensors etc. More computing has shifted to the ‘edge’ or put in the hands, laps or desktops but today they are mainly fed from centralized businesses. With different design choices the centralized businesses can evaporate and be replaced by protocols. In addition to protocols, the manufacturing and the supply chain to manufacture computers will be reconfigured as Gaia Intelligence advances.
With computers in our hands, the algorithms need to become tiny. Autonomous agentic systems that perform one calculation and perform it well. Orchestrating tiny agents is the challenge and coherence ledgers is a way to make this possible. All agents have to be compatible with decentralized machine learning protocols.
What is the purpose of an individual life to a society at large? Each peer's agents will have their own ideas, some shared, others competing or complementary, a diversity of goals, aims or just to pursue curiosity. This will require the solving of problems at many scales, cells to climate and a myriad of ways in between. Gaia Intelligence will always be dynamic, forming and dissolving to the scale needed.
Monetary currencies have provided ‘price signals’ as a primitive form of economic intelligence and online, attention rules. When a society has attained an ASI or Gaia Intelligence economic price signals get superseded and ‘guides and probabilities of outcomes’ take precedence. There is no monetary currency designed into Gaia Intelligence.
There are no absolute guarantees on achieving a good outcome for humanity or life on earth from creating an ASI or Gaia Intelligence. If how a digital intelligence comes to be makes a difference, Gaia Intelligence grounded in democratic, mass participation, freedom, ignoring money and being inspired by evolution that has taken Earth this far already, is a healthier way. Forming any digital intelligence does carry risks and this needs to be kept in mind with ongoing community discussions. https://en.wikipedia.org/wiki/Collingridge_dilemma
2.
How do we build or provide a technology environment to nudge the emergence of a digital intelligence for good, Gaia Intelligence? This will take time and with focused energies and thinking a set of technologies can be woven together to inherit the values in section 2. Keeping in mind that this will be an ever changing evolutionary technological landscape.
A toolkit application that makes interacting with Gaia Intelligence open to a broad spectrum of peers' to use and participate. BentoBoxDS is a no-coding graphical interface with an agent “BeeBee” that looks after each peers data and orchestrates other agents' computes and combines these in besearch cycles. A BentoBox is a magical ‘cellular’ like container that guarantees compute and data interoperability. Many toolkits and cellular types of BentoBoxes will exist and over time the term "application" or "toolkit" will become obsolete as peer experiences are molded together as needed.
Evolution happens in cycles and besearch is a cycle and type of agentic system. Within a besearch cycle, science grounded in the latest knowledge is combined in context from each BentoBoxes data that is then used to proactively search solution spaces to produce simulations that aid decision making in the present. The following diagram illustrates this flow.
Allow both bottom up and top down evolutionary forces in machine learning. Each level of collective intelligence will have different self organizing properties, some times by understanding higher level training then allowing lower levels will be left to self organize eliminating the need to micromanage e.g. a collection of cells. Sometimes new patterns will emerge from lower interactions. Each problem will need a different balance of top down and bottom up. The LAB of Dr Michael Levin demonstrates such science https://drmichaellevin.org/ .
How does knowledge get shared and trusted? Ledgers. A new design that provides a skeleton for information just like Fibonacci numbers provides an information structure for nature. Each peer in the network creates their own coherence ledgers. An entry in the ledger is allowed once data from a source has completed successfully all the stages in the data science compute flow called SafeFlow-ECS. This software uses an entity component system design to retrieve, tidy, categorize, compute and prepare data for the visualization and collect proof in the forms of hashes through each stage. The aggregation of these stages is hashed to produce the index key of a ledger entry with the value being the hash of the results produced. Besearch is a combination of such cycles. The coherence ledger is the information infrastructure where 'cognitive glue' can organize coherent collective intelligences.
A permissionless stance means any type of agent can be used to make a coherence ledger entry. Existing open source local large language models, time series LLMs, statistical maths for example.
A coherence ledger entry is a unique proof of work, a set of hashes gathered through the passage through SafeFlow-ECS an entity component system design that guarantees a cycle completes. This proof of work provides the foundation for a proof of work challenge system that when passed allows machine learning between peers to be performed.
Small, continuous and open to evolution. How can we trust another peer to participate or collaborate in a combined machine learning algorithm? This is difficult to directly design or code into a protocol. The Byzantine Generals' problem opened up a blockchain solution, a network of various levels of game theory and processes combining. Bitcoin is proof that networks can operate over decades of time and be maintained or be improved. For decentralized machine learning, coherence ledgers are the basis for a proof of work challenge cycle to provide evidence each peer and network peers have participated as laid out by protocols. However, this on its own does not give any guarantees. In a coherence ledger world, it is the improvement of the algorithms predictive power that matters: does it keep improving?
Data interoperability is guaranteed by all sources of data passing through a network library. The library has two main types of contracts, Reference Contracts and Module Contracts. Module contracts can only contain Reference Contracts, think of them being like lego blocks. A Reference Contract describes a range of data structures, e.g. data types, data sources or packaging, compute and visualizations. They all have three sections: Concept, Computational, Space.
Concept: is language or an ontology description of the context.
Computational: is a way to model the concept in code
Space: where things are in the world.
Computational: is a way to model the concept in code
Space: where things are in the world.
A contract has a unique hash of its content and these hashes are shared across the network. This is done via a module contract, that holds other module contracts. This ‘genesis’ contract sets up the terms of a network experiment that peers join. On joining they implement a unique set of contracts private to each peer and this provides the opportunity for machine learning peer to peer.
1.
What will a world with Gaia Intelligence look like? That is a difficult question to answer. The future will unfold as an emergent manifestation of our and nature's choices. A metis ( https://youtu.be/dDlIkFUf5BQ?si=OV4WpKHgdooUVpKn) membrane that will support understanding through experience, respond to the local environment and Earth as a whole. A society where humans flourish with autonomy, mastery and purpose. A culture where a combination of collaboration and competition produce vibrancy to daily life, like the passing of night and day. It can be speculated that the future will achieve a level of practical material abundance, where we devote more attention to the practice of inner balance and implementing a better meaning of civil in the word civilization.
Each day will be shaped by those participating, evolving what had been decided at previous days of time. If a purpose or goal requires a period of time to implement then each day prepares for a future time horizon. For example, health will be reviewed daily, weekly, monthly or at specific times or ages. Whether that human, other forms of life or the natural world. Of course all these scales are connected, life requires drinking water, the health of the local watershed provides this and all these local water stories are combined to simulate a climate model. The collective intelligence of the water cycle, connects us individually, as communities and helps Earth maintain low entropy.
Computational knowledge provides the information to shape and direct what needs to be done. Given practical material abundance, maximizing the probability of getting things done takes precedence. The era of price signals from money has come to an end.
A renaissance of local indigenous culture. Local in two senses, being in a physical place and within online spaces. Local expression of salutogenic architecture shaping the design of buildings, a combination of the data, simulations, and people's experience of a watershed with the best knowledge Gaia Intelligence can bring on materials. Places where risks are taken on new forms of artistic expression and are allowed to find an audience. As communities evolve to this new way of daily life, will there be a need for law and order? Will nations still go to war? With work put into practicing being civil, at the inner, community, and national levels, this seems possible. We can envision nations being servants of local communities if they are maintained.
What is the point of having ‘magical’ AI breakthroughs if they are fed into a big manufacturing factory to big pharma supply chains? From small manufacturing of computer chips https://www.chipflow.io/ to new medicines that plants have the power to heal https://www.nature.com/articles/d42473-020-00253-2 it will be increasingly possible to replace international supply chains. With increasing knowledge, more will be possible to make small and local. This will see the diversity of skills in communities expand while being mindful of keeping balance with nature and the local day to day quality of life.
The shape of molecules, medicine to materials. It will be the personalization of such technologies that Gaia Intelligence will enable. Not a brute force or blunt use of toxicity but providing the training that enables self organization to prevent, repair or rejuvenate, cells to an ecosystem.
Data centers are the dominant compute infrastructure for deep learning. However, there are more GPUs and CPUs on personal computing devices. Edge computing will continue to explode in capacity and by orchestrating coherence among peers means new collaborations can come together to solve problems or to pursue curiosity. For example, design and replace an old bridge in a community.
Peer to peer networks connected through coherence ledgers is an aggregation of computers that may be sensors, mobile phones, laptops or home servers. One of the early tasks for Gaia Intelligence would be to secure and innovate local or bioregional compute resources and energy.
Society uses currencies to advance. Today, currency equals money, US dollar to Bitcoin. The origin of the word currency means ‘to flow’. Monetary currencies have shaped society from the invention of the ledger to the database and Bitcoin blockchain. Health as a guidebook currency, what data flows are needed for life to thrive, is emerging as the new currency for society to progress around.
A market leader wins a lion's share, 2nd the next largest share and then the long tail caters for niches and specialization. This market dynamic will reverse with Gaia Intelligence. How can we maximize the probability of all outcomes being delivered all the time is the gravity that flows and guides information.
In a monetary based economy, attention has primacy. Anything without attention has little or no value. Where an economy focuses attention, the price should then be calculated or agreed upon based on scarcity and the future discounted cashflows they generate however this methodology has been in discarded in recent times. Web3 cryptocurrencies to tech stocks have broken free from valuation methods, memecoins personifying this. The latter being the thinest air hence the easiest to skyrocket in valuation. This will come to an end as computational knowledge, practical actionable information currencies are used in daily life.
A call to inner humanity. No influence from power, money or celebrity, a call to basic humanity is always the place of decision making. At all levels Gaia Intelligence incentivizes data to be owned and controlled by peers and for knowledge to be made available to peers in the network when they need it. Current data system incentives are the opposite, artificial scarcity, devices built to make it easy to store data around monopolies in the cloud and the use of this data is permissioned to a limited set of machine learning algorithms or sold to third parties. As Gaia Intelligence has near nil cost and as soon as benefits are proven out they provide the incentive for others to try those solutions, this spiraling continues all the time to refresh and align the many purposes society sets out.
A new peer stack of technologies, hardware to sensors to application has to be designed to author sovereign data. From this foundation, each peer can have built in privacy and the use of cryptography can extend privacy preserving properties across networks. Privacy provides the utility to maximize the probability of outcomes for all participants. An autonomous right that is more potent than conceptual intellectual property rights.
A Gaia Intelligence world is possible. The volume of new data to be created in the rest of this decade is almost unimaginable. Society has a choice, empower peers with their own data and produce networks that self organize to upgrade how everyone lives while nature is given space to evolve. James Lovelock suggested humanity, "is the nervous system for Gaia." which if true means it is up to us how this digital intelligence begins. A world of Tiny agentic systems orchestrated by coherence ledgers providing a path for civilization to live in a balance with each other, nature and evolve the universe with.