A "symbolic' release of a Desktop BentoBoxDS and HOP will happen this week. New bentoboxDS website https://bentoboxds.org/ . The website will have links to download download Linux, Mac & Windows.
Today, start the focus on weekly releases. Today release, gives peers the first new peer experience on the fly release. Set public/private keys and secure with a password. Set Heli clock to birth or current location. Back in February I was implementing the besearch cycle and was looking for trigger to start e.g. start of a new day? Thus a digital solar clock was designed, an emulation of the earth and sun that is more scientific plus for biology seem like an obvious choice. Story ⟶ Interplay ⟶ Emulation . The story starts by writing anything, we call this a Life-strap story send it to beebee and in return an Lifestrap-lens interplay will be displayed. The goal of the Interplay is the build the knowledge each peer has and how it wants to direct beebee. In HOP a range of pattern textures and resonAgents (private tiny neural networks, empty on start and developed with the knowledge each peer holds.) The LS-lens splits cue words into columns. The Capacity column links to setting up besearch cycles, strands & braids that build the emulation. Context column click on cues to enter a cueSpace where more conversations and context can be built. Still work to do here but you can try the tools. Watch the video for help and instructions.
What is next? The Lifestrap-story is not just a dialogue with beebee. Data flowing from Tiny Devices. Polar and Withings dataset are ready to grafted in and start to pulse. That is the will involve SafeFlow-ECS the heart of HOP, it has been through an upgrade.
Following that, fine tuning how the resonAgents, SafeFlow-ECS and consiience weave flow together around besearch cycles and pulses of data from tiny devices.
The Sculpting Lab. Orgo and gelle grafting, coupling and melding to provide new emulation building blocks, a cardiovascular system to fertility organs etc.
Then DML, can each peer use neat-hop, evolutionary ML on their own and with other peers on HOP?
HOP as a whole, strikes out in a new direction to learn. A LLM or deep learning set weights by backpropigation. HOP uses the resonancePluse, that is the contextual geometic buffer, it how well is each data set aligned with the digital solar cycle each peer life by.
Today, start the focus on weekly releases. Today release, gives peers the first new peer experience on the fly release. Set public/private keys and secure with a password. Set Heli clock to birth or current location. Back in February I was implementing the besearch cycle and was looking for trigger to start e.g. start of a new day? Thus a digital solar clock was designed, an emulation of the earth and sun that is more scientific plus for biology seem like an obvious choice. Story ⟶ Interplay ⟶ Emulation . The story starts by writing anything, we call this a Life-strap story send it to beebee and in return an Lifestrap-lens interplay will be displayed. The goal of the Interplay is the build the knowledge each peer has and how it wants to direct beebee. In HOP a range of pattern textures and resonAgents (private tiny neural networks, empty on start and developed with the knowledge each peer holds.) The LS-lens splits cue words into columns. The Capacity column links to setting up besearch cycles, strands & braids that build the emulation. Context column click on cues to enter a cueSpace where more conversations and context can be built. Still work to do here but you can try the tools. Watch the video for help and instructions.
What is next? The Lifestrap-story is not just a dialogue with beebee. Data flowing from Tiny Devices. Polar and Withings dataset are ready to grafted in and start to pulse. That is the will involve SafeFlow-ECS the heart of HOP, it has been through an upgrade.
Following that, fine tuning how the resonAgents, SafeFlow-ECS and consiience weave flow together around besearch cycles and pulses of data from tiny devices.
The Sculpting Lab. Orgo and gelle grafting, coupling and melding to provide new emulation building blocks, a cardiovascular system to fertility organs etc.
Then DML, can each peer use neat-hop, evolutionary ML on their own and with other peers on HOP?
HOP as a whole, strikes out in a new direction to learn. A LLM or deep learning set weights by backpropigation. HOP uses the resonancePluse, that is the contextual geometic buffer, it how well is each data set aligned with the digital solar cycle each peer life by.
Opensource:
————
The attunement of health. In the old world we would speak of medicine. Take your medicine etc. This sets HOP on a different paradigm of health grounded in science. Heli a better scientific clock for biology, Von mises math, statistics of circles not straight lines, this embeds from consilience weave to setting weights in a neural network, emulation that learn and in themselves are attunement. How does a peer attain attunement? Through a three part process Lifestrap-story ⟶ Interplay (build out besearch cycles) ⟶ Emulations. This weeks DEMO showcases this pattern. DEMO besearch cycles https://rumble.com/v78r30w-besearch-cycles-story-interplay-emulation.html
—————-—
DEMO Overview of v1 candidate release taking shape https://rumble.com/v78fo7u-bentoboxds-overview-of-peer-experience.html We have a basic formula: Story → Interplay → Emulations. The emulation (orchestration of resonAgents) 'worlds' shown in the DEMO need to come from 'models' and these models just do not appear from thin air. They come from the 'interplay'. Peers teaching resonAgents with the help of a range of bentoInstruments and lifePattern tools. A Story, each peer creating a dialogue with beebee to provide the context, ideals, values, previous knowledge held by each peer and via a process of 'interplay' resonAgents (HOP replace of LLM's) will start to learn how to build emulation models. One of the bentoInstruments was a small LLM. That has been removed for now. Why 'pollute' sovereign knowledge with machine learning weights from the old world?
beebee help peer direct a consilience weave https://coracle.social/notes/nevent1qy88wumn8ghj7mn0wvhxcmmv9uqzq0x7x7fgmh98a4l59t70g0pmgqxyr5feag336n4k50frm86f80cgk5llk9
—————-—
Last week saw the first go at wiring together the 'brain' of beebee, the BentoBoxDS agent that each peer directs. It is complicated but beebee has a small LLM that has been setup to never answer a question directly. Its role is to manage and act as a translator for resonAgents or to call on skills within or outwith the HOP network, e.g. a open source LLM specializing in medical research or product selection. resonAgents are nano neural networks that each peer @teach and through flows of data e.g. from a wearable to be a digital twin. A set of resonAgents working to bring the emulation worlds of BentoBoxDS to be. DEMO https://rumble.com/v784ley-bentoboxds-four-quadrants-tour-of-demos-swimming-heart-water-earth.html tour of the emulation worlds.
Best guess then learn https://beebeehop.any.org/newsletter-diy-health-cues
Keep heading toward coherence https://coracle.social/notes/nevent1qy88wumn8ghj7mn0wvhxcmmv9uq32amnwvaz7tmjv4kxz7fwv3sk6atn9e5k7tcqyqjvda4hxe5h7x2xl49dvnffw3fmmd9c60xd95erx5h88cnf9skfkaenav0 https://coracle.social/notes/nevent1qy88wumn8ghj7mn0wvhxcmmv9uq32amnwvaz7tmjv4kxz7fwv3sk6atn9e5k7tcqyqjvda4hxe5h7x2xl49dvnffw3fmmd9c60xd95erx5h88cnf9skfkaenav0
—————————-
Each peer has their own unique Heli clock. The HOP network has a public library and contributions to the library need to have a heliStamp. The Spring Equinox of March 2026 is the universal heli time for HOP. This is the most scientific measure of time we can make. You can see a Heli clock on the new https://bentoboxds.org/ website. A start to introducing the vision and terminology of BentoBoxDS and HOP. Download, beebee and new vidoes coming soon!
OpenClaw and the concept of a claw is taking hold. You can think of HOP as a conservative release of openClaw (& it its many copies). However, HOP provides a skeleton of a structure, the protocol sets a purpose, allow peers to establish coherence on health at all scales. We use the technology of Karparthy autoResearch a type of besearch https://youtube.com/clip/UgkxHvCmfATdm_FcuUJ2clmogKonDZ8dVu5n?si=oEEGLiqyhsqm6tVK
Terminology and speak https://beebeehop.any.org/newsletter-diy-health-cues
—————
As we approach the Spring equinox, the start of the great orbit cycle ( a year in old world speak) the new BentoBoxDS.org and healthscience.network will be published to the web. The clear message will be to set out a foundational new approach to health. We will introduce the language of resonancePulse, besearch and consilience weave and explain and demo. The technical side will also establish the Pillar of the Protocol, SafeFlow-ECS, resonAgents, neat-evolution and the coherence ledger the provides the opportunity to peers to learn in network and practice DML (decentralized machine learning). A new release of BentoBoxDS will be concluded and HOP will follow. The goal, the candidate release of HOP V1 to be stress and peer tested. DEMO https://rumble.com/v776qj2-life-strap-and-emulation-worlds-tour.html
Which tiny devices? https://beebeehop.any.org/newsletter-diy-health-cues
———-
On the 4 February an opening dialogue was authored, https://beebeehop.any.org/a-tiny-hop-to-gaia-intelligence This set out the vision to complete the health oracle protocol HOP and BentoBoxDS a peer experience created on the fly. With all the parts coming together, a second dialogue as been drafted https://beebeehop.any.org/another-tiny-hop-to-the-great-orbit DEMO heli clock https://rumble.com/v76utzo-heli-clock-bentoboxds.html
TINY device data, rethink? https://beebeehop.any.org/newsletter-diy-health-cues
—————
Keeping any application to datastore secure is a none trivial exercise. Security experts talk about the security surface, from social ie. peers manipulation to software and cryptography flaws. The golden rule 'never roll your own cryptography'. Use what has been proven in real use over time. HOP uses the same key setup as Bitcoin and Nostr. See DEMO https://rumble.com/v76i1ei-bentoboxds-security-password-and-keys.html BentoBoxDS can not be used until a genesis handshake is undertaken. A returning peers needs to prove they own these keys via a password that is used to unencrypt the keys ready for use. This will be each peers master key but HOP uses many other keys and the next job is secure these from the master. Why is this important? It keeps the application secure and private but also gives the potential to re-establish a peers account from saved 'blind' encrypted backups.
——————
BentoBoxDS and beebee role is to help peers create or build the best peer experience. Traditionally an app or toolkit would have a fixed set of features. This is not the case or design for BentoBoxDS, we aim to create the peer experience on the fly. For example, if I have resonancePulse for 400IM in swimming, I need button for the four swimming strokes. We do not hard while that categorization but make it functional when we need it as next we might be compare different types of swimming pools etc. DEMO https://rumble.com/v75tpvw-peer-experience-on-the-fly-bentoboxds.html
Besearch cycles all the way down https://beebeehop.any.org/newsletter-diy-health-cues
——————————-
Summary: Heli Clock DEMO
——-
Summary: HOP Cryptography
DEMO https://rumble.com/v74jblu-chat-in-context-teach-besearch-experience.html start with showing beebee chat in context, main chat view or in context of a cue space, swimming in this demo. While in swimming we ask a question on HRV and then activate @teach mode, the peer then uses the tools to build the data for the query. These teaching examples can be view via the library and then the next task is to 'fine tune' beebee to learning so that it can answer any future questions. Not only can a peer @teach a network of peers can aggregate their @teach experiences. Lastly, we have a looking at besearch (which will become the main starting experience some day). Here peers can create new besearch experience and categorize them to preventation, repair or rejuvenation or custom. They will be display on a cue, body or earth world to make context more intutitive. Working to get this release out ASAP.
besearch experience https://beebeehop.any.org/newsletter-diy-health-cues
———————————-
Summary: Coding agents
Coding agents are key to making HOP and BentoBoxDS better but also to CoherenceStream in all aspects. There is talk of AI native coporations and we are well on the way to establish AI agent as team members. This is not without risks, like back in December a LLM used via a coding agent produce wide scale havoc to bentoboxDS code base on a major refactor. This will happen from time to time but system can be put in place to detect these and the code lesson again is to keep things small or tiny. Use 10 coding agent to refactor rather than one or two. We need to have many agent working on the peer experience, testing the experience, reviewing the code for security and privacy design.
peer experience on the fly https://beebeehop.any.org/newsletter-diy-health-cues
———————
Summary: deadlines
The milestone that matters most is getting beebee and besearch TINY agent learning. No matter how slowly this happens for the starting period, having code, processes and networking peer to peer to cohere around shared learning is the year end milestone. Lots of refactoring of code is taking place.
How a besearch peer experience works? https://beebeehop.any.org/newsletter-diy-health-cues
—————————————————-
Summary: tiny learning
——————————
Summary: Business plan
Big effort, with the help of LeChat to find the right messaging for the business plan. See business plan page on CoherenceStream anytype. A draft plan outline is there and the executive summary is being played with and then each section will be written this week and then executive summary will be updated. A spreadsheet with numbers will also be needed. Lots to do. This is taking priority over coding at present. What time I do have is spent preparing the solutions to implement in the weeks ahead. Got some good conclusions to share in the weeks ahead.
When data changes, why? https://beebeehop.any.org/newsletter-diy-health-cues
————————————————————-
Summary: AI feedback
ii.inc new agent provides a slide building services, this is what is made of HOP and BentoBoxDS https://8000-iq71ogjga2ltqpdw4lmqx-6532622b.e2b.dev/hop-bentobox-slides.html and LeChat and Grok had there say too, https://beebeehop.any.org/grok-on-hop
newsletter and demo of Withings scale data https://beebeehop.any.org/newsletter-diy-health-cues
——————————————————-
Summary: the heart of the matter
Q. What can we learn from heart data?
Heart data has been one of the founding data sources for HOP and BentoBoxDS. But what can we learn from such data? Taking a pulse rate with your own fingers, to different generations of fitness tracker and now even smart scales provide ECG tracings. How do we bring this right computation to besearch cycles is the next task at hand.
The introduction to this network project has been updated, https://beebeehop.any.org/product-introduction Lots to read here, purpose of network, feature, development of software options and besearch cycles to focus on once data is flowing.
—————————————————————————————-
1 September 2025
Summary: Let sovereign data flow
Q. How to get data from tiny devices?
One trend we can bank on is that devices are becoming smaller and more accurate on measuring what they set out to measure. The future is TINY.
product: healthCues
healthCues is both a tool that peers can use in a stand alone mode or by joining networks. In solo mode, the challenge is to aggregate as many sovereign data sources to provide a holistic understanding, learning and discover outcome driven actions that work. The first network is DIY Health. Can home consumer devices provide quality data to fuel a new scientific paradigm of results? Three category of devices are being used, Polar wearable, Withing smart scale and a range of blood tests. The first task was selecting the devices, the next will be getting the data flowing to each peers personal sovereign datastore. A topic for next week.
Opensource
HOP already uses HASHING algorithms and suite of libraries from Holepunch, storage to peer to peer contain a lot of cryptography. Now hop-crypto joins. Each peer starts with a genesis handshake that creates a public/private keypair and will be protected by a password (not full implemented yet). These keys will include the ability to regenerate a data set if a peer has 'blind' saved a copy of their coherence ledger with other peers.
Resonance not statistics https://beebeehop.any.org/newsletter-diy-health-cues
——————————-
Summary: Heli Clock DEMO
HOP runs on a heli clock. A digital solar clock. Why? It is a scientific way to record the passage of time. Not a human constructs of units of time but the geometry and changes in motion between earth and anything on it and the sun. This DEMO shows the clock in operation and making an event https://rumble.com/v756mns-heli-clock-hop-run-on-digital-solar-clock.html Note the new time units, earth orbit, solar cycle, sector (part of solar cycle) and base unit, an arc 1 degree of angle. This has profound implication for science and electronics. Full intro @ https://beebeehop.any.org/heli-clock
science on heli clocks, more intuitive https://beebeehop.any.org/newsletter-diy-health-cues
————
Summary: TINY year of learning
2026 is upon us. Who will be left behind (in the cloud) https://youtu.be/pOb0pjXpn6Q?si=ZsAgMUe0IHctgcUZ from a Silicon Valley insider. Outside SV, they are unaware of HOP and BentoBoxDS, well Gemini is more up to speed now https://gemini.google.com/share/40787947d4b4 This highlighted the need to communicate to non geeks and biohackers. Some good examples and wording coming together. It also gave critique of the project, 'foundational' in nature and suggest how it would seed the network in small clusters. Which aligns. The conversation made reference to James Littlejohn so that was the first time an AI did that. I informed Gemini it was talking to me and that was an interesting experience.
Consilience between besearch https://beebeehop.any.org/newsletter-diy-health-cues
≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡≡=
Summary: Coherence.com stream 2025 year in review
The base learning data structure in LLM's are transformers. To over simplify this is the ingredients fed in to the LLM's neural network. We do not have to use transformers, much innovation is happening, https://www.verses.ai/about nature inspired, like HOP and even the co-inventors of the transformers see a different future https://youtu.be/DtePicx_kFY?si=Qj7Sb7CufWpP-gTO https://sakana.ai/ and along with our idea of DML, show a path to learning better from TINY DATA.
holistic data gives insight https://beebeehop.any.org/newsletter-diy-health-cues
————————————————————
Summary: 2026 network learning
It is spoken that 2026 is the year where self improving AI really accelerates off. I have held this view for a long time and now here we find ourselves near the eve of 2026. We must have our tiny agents in the world too by then. That is the goal and BentoBoxDS and HOP are being upgraded with new interfaces and capabilities with @training mode for beebee so we can get peers and agents learning together.
Starting to draft the key messages and metaphors and real demos to meld together in a slide desk, https://beebeehop.any.org/slides . How does CoherenceStream make money on transition to computational currency society? Answer shares in RGB tokens tied to besearch cycles we help peers setup. RGB has a unique feature of being peer or client side verified technology so no need for a blockchain, listen to its founder explain https://youtu.be/m5E6-npQ97w?si=WgT9N_pEbrcSRM0C
learning from TINY data https://beebeehop.any.org/newsletter-diy-health-cues
Help on how a computational currency, cues will be https://coracle.social/notes/nevent1qy2hwumn8ghj7un9d3shjtnyv9kh2uewd9hj7qgwwaehxw309ahx7uewd3hkctcqypzgyfjm3pgulu9y60jke8ga7t7mwxlah7mud736jfzd7aaey5z46780vc0
————————————————————-
While a network experiment is made up of four modules, it is easy to see compute module as being the most important? It holds the context for the data and compute algorithms to come together to produce a visualization or simulation etc. However, this conversation https://youtu.be/FcnLiPyfRZM?si=a7UHycPdDw1nAHnU reminds us that text is the best descriptor of reality? So, question, goal, outcome, gets turned into compute algorithms that can be evolved using text?
bring transparency to metrics https://beebeehop.any.org/newsletter-diy-health-cues
newsletter https://beebeehop.any.org/newsletter-diy-health-cues heart besearch preparations
beebee guided by its peer
—————————————————————————
Summary: Sovereign data defined
If a bit of data needs to be touched, i.e. converted to a different data structure e.g. propitiatory date format to UNIX timestamp then the data is not sovereign. In short, every time data is manipulated the chances of error in conversion have a chance to creep in. Plus in the mean time what has a cloud app done with the source data? The future is tiny devices, that produce tiny sovereign data using tiny AI agents.
read newseltter https://beebeehop.any.org/newsletter-diy-health-cues
—————————————————————————————
————————————————————————-
25 August 2025
Summary: Peer designed code
Q. How much agency to give peers?
It is clear that those with the best coding agents / LLMs will write the best code. I.e. use their tools to write better version of themselves. This opens up the question, where should the agency lie? In HOP each peer has prime agency and it will be up to those building peer to peer networks to align around core protocols and data structures within HOP. Today, those are fixed, at least per network experiment.
product: healthCues
Last week started a week of conversations with AI coding agents and that has produced four documents to build IOS Android and Webapps based on bare.js. This includes confidence building steps so we hope to have a simple mobile app to show be next weeks newsletter.
open source
———————————————————————————————————————
18 August 2025
Summary: building for use cases.
Q. How to build what is needed?
https://beebeehop.any.org/newsletter-diy-health-cues
———————
A introduction with voice commentary to the new peer experience coming together on BentoBoxDS https://rumble.com/v765jh4-bentoxboxds-hop-peer-experience.html Such an experience can only come to be if beebee and its support tiny agents called resonAgents can extract and learn. There are two way to do this, @teach a peer teaches beebee and NEAT, an evolutionary algorithm that can take the resonAgents tiny neural network and explore different weight settings to produce higher quality science to visual experiences.
Navigate through worlds https://beebeehop.any.org/newsletter-diy-health-cues
———————
Summary: BentoBoxDS bbNexus DEMO
BentoBoxDS goal is to allow all peer equal access to and ability to use HOP reguardless of their technical skills. BentoBoxDS and HOP and complex and a new tool, bbNexus has been introduced, see DEMO https://rumble.com/v74v7wc-intoducting-bbnexus-toolbar.html . Within three clicks the right tool or support should be available to peers and given the context the peer are in, the default settings will use this information. beebee agent is a starting place, or within a BentoBox toolbar or in besearch worlds where you can even navigate around the body and create new research.
create besearch cycles https://beebeehop.any.org/newsletter-diy-health-cues
≡≡≡≡
Summary: BentoBoxDS DEMO
More effort was put into building out coherenestream beyound the code. HOP and BentoBoxDS require most effort while a core set of capabilities are established and made robust. Use of coding agents has transformed the coding and LLM have helped debate and refine the business plan. healthCues was an attempt to focus on a limited set of devices and experiment use cases. Much effort but no complete end to end peer experience has been delivered. The early months of 2026 will need to complete that and show evidence of adoption. Adoption plus TINY learning step is the only path forward. Full discussion on the tech https://beebeehop.any.org/dev-newsletter CoherenceStream.com website was updated and next will be put a live TINY agent on the site. First week of 2026 will see the business plan updated and reach out to investor source and attending geek and investor conference where those most likely to see the scale of HOP vision. All the parts are coming together, can we get them to learning together in 2026?
How a besearch peer experience works? https://beebeehop.any.org/newsletter-diy-health-cues
≡≡≡≡≡≡≡==———————
Summary: TINY agents
Each peer has beebee to support and guide their information gathering experience. However, the HOP network needs to establish coherence on all health data. Peer will setup TINY agents and then collaborate together to share learning. In simpler speak, A TINY agent will manage a set of besearch cycles and those besearch cycles will include network experiments that is grounded in data that comes from each peers data from TINY devices. see diagram https://design.penpot.app/#/view?file-id=8fd8c29f-33f9-8038-8007-3aa5b2547ce2&page-id=f329de2c-67b1-801b-8005-bb46f48bfa19§ion=interactions&index=0&share-id=8fd8c29f-33f9-8038-8007-3aa8f7dab671 The network need to strike a balance between number of TINY agents and the ability to find them and constrained by each peer compute resource ie. they type of laptop they own etc.
How a besearch peer experience works? https://beebeehop.any.org/newsletter-diy-health-cues
prevention interventions? https://beebeehop.any.org/newsletter-diy-health-cues
full architecture of the project https://beebeehop.any.org/besearch-hop-cuecurrency
———————————————————
Summary: Besearch and DML
Answer: the ability to learning from tiny data from sovereign devices peer to peer. This is a 'slow' or patient approach to machine learning. A network of tiny agentic agents in besearch cycles, each supporting or self correcting other agentic agents.
Control over besearch cycles https://beebeehop.any.org/newsletter-diy-health-cues
———————————————————————————————————
The potential is there to change from a web2 social interface to a computer game inspired fluent and dynamics real time experience. Besearch cycles is where we will first show case this design pattern.
polar fitness watch data joins the network https://beebeehop.any.org/newsletter-diy-health-cues
—————————————————————-
15 September 2025
Summary: Escape the cloud
Most data goes from devices to mobile apps that store the data in the cloud. Can we give peers that are in this world the option to start the process to leave the cloud? It is technically possible, call the API of the app owner and get a 'second' hand copy of the data. However, at the same time, we need to provide a clear, pure sovereign alternative, tiny device direct to a personal datastore. Give the HOP network time to show value and then peers will buy their next tiny devices to be a HOP native tiny device and app?
Has own newsletter letter now https://beebeehop.any.org/newsletter-diy-health-cues
Open source:
open source AI agents to use in besearch https://coracle.social/notes/nevent1qy88wumn8ghj7mn0wvhxcmmv9uq32amnwvaz7tmjv4kxz7fwv3sk6atn9e5k7tcqyzefg2ep02gl858lkhpkuxk288ynpj5wlxx7vrmnrwt6d7d2huz8w60zmh7
————————————————————————————————————-
Summary: intelligence v money
Even the 'startups' from OpenAI to Anthropic to Perplexity.ai are coupled or tied in someway to the tech cloud infrastructure giants of Microsoft, Amazon and google with Facebook and X AI building and expanding their own compute resources. Even outwith the cloud, the most profitable crypto bank in the world is flexing its tech muscles in all directions but knows it has to win on AI, https://qvac.tether.io/ and ii.inc harnessing LLM open source community. How does CoherenceStream complete? Re-base the 'game'. ii.inc is boldly trying to do that, post money destination, like HOP but based on blockchain rather than CoherenceLedger, like Tether, both will open source their tech and make it local. We can use that tech where needed and that leaves CoherenceStream to ensure we only attain learning based on sovereign data from tiny devices. The biggest differentiation is on the type of super intelligence to aim for, a gaia intelligence, rather than one model to rule them all. We will have to be brave and courageous. Be mindful not to succumb to copying what is hot, remain committed to the 2008 founding vision, build the parts that matter and now everyday show the value in building networks to demonstrate what is possible and practical for peers the HOP way.
If only we knew the answer to this question before hand. Two things usually come together, a break in the existing way and a new solution waiting ready to step in. Now, there are varying scales, a new financial system to a useful incremental improvement, a useful tool that solves a problem. We are very much in the former. High risk, High reward. But we can start with building a small network, e.g. tracking weight and body composition via a smart scale. Add wearables and blood biomarkers. It seems that the sweet spot for HOP and BentoBoxDS is bringing together different data sets but maybe just getting the data flowing from one device is that starting place?
product: healthCues
The reason this newsletter is late today is that I was asking the coding Agent to build healthCues, redesign BentoBoxDS. This took some time and the results were not visually impressive or new but it did restructure the whole code base to make it easier to navigate (I will have to decide to do this or let go fully of the coding). Here is a screen shot and plan to do a better job.
Open source:
——————————————————————————————————————-
11 August 2025
Summary: Build network proof → investment
Q. How to establish credibility?
Build real working networks. Two basic types: Build a local resilient hospital via MENA health and DIY Health using only at home consumer devices and tests. Key to both is getting the data capturing and flowing into HOP and thus accessible to BentoBoxDS. Still desktop version for now.
product: healthCues
Debugging windows paths still taking up time.
OpenSource:
————————————————————————————————
4 August 2025
Summary: Besearch cycles to coherence
Besearch is the learning paradigm being pursued to achieve coherence for each peers BentoBox and across networks of peers. Cycles of difference 'scales' will be required and how to combine those is the human and machine learning challenge before us. See beebee link below for more details on besearch cycle flows and processes.
Looking into how to integrate https://ii.inc/web/blog/post/ii-medical
———————————————————————————————————————————————-
28 July 2025
Or is money still just money regardless of its origin? The investment page lists funding targets. Three classes: Technology or AI visionary sources: Like cofounders of Imbue, Bill T invested in every big trend from 80's and those at open source end of LLM's, openhands investors for example. Then Bitcoin pro AI folks, Simon Dixon and EgoDeathCapital (post seed probably) and then lastly, left field or those that see the need to build and invest in networks directly, e.g. hardware or new forms of networks, beyond functional medicine health. The latter requiring partnerships as a pre-condition e.g. medical device manufacturer and CoherenceStream, it would be too much to take on this alone (unless with a 1bn dollar in the bank).
product: healthCues
Nothing to report.
Opensource
——————————————————————————————————————————
21 July 2025
The founding culture sets in motion an internal currency that will be embed in our products, attract and keep peers and employees motivated, attract investment and collobration with other open source software, hardware and wetware projects and entities. CoherenceStream is a Besearch Entity. What that is an on going conversation but here are some existing entities inspiring culture formation, https://beebeehop.any.org/legal-structure .
Gearing up AI agents to be able to make personalized android and IOS and webapps.
————————————————————————————————————————————————-
14 July 2025
Commitment to learn on TINY data with TINY agents. The learning is at all scales, Besearch cycles that weave together context, learn from the past, search out the future, learn with others. But how to learn? Trial and error. Learn by doing and learn by example. The former good for both looking backwards and forwards. Learn by example: who or what to follow or be inspired by a pattern in nature. We need to use all these techniques and put in place self organizing flows that bring coherence. Sometimes, learning is slow and sometimes fast. Sometimes persistence is needed at other times a split second of inspiration. The challenge is to bring these to HOP.
A decision needs to be made, use application APIs or strike a partnership to bring sovereign data to building networks? e.g https://humeprograms.com/partnerships Only focus of TINY, DIY and trend to be most accurate and relevant over time.
————————————————————————-
Tiny, decentralized and networked. This pattern holds for the use of AI agents in HOP. A chatbot in the BentoBoxDS user experience to formatting a CSV file to checking of a date pattern etc. etc. . An AI agent should be on call to help. Now, putting the infrastructure in place to bring coherence to using the right agent at the right time is the challenge at hand. The Cloud have MCP (model context protocol) emerging as a standard, but we have the Coherence Ledger and context from the library to give better resolution to the learning. Besearch in short.
Design pattern for starting user experience being besearched. The Sun (vitamin D / circadian rhythm), FMD fasting mimicking diets (nutrition), Creatine (supplements muscle mass) and more challenging rapamycin and placebo effect.
Refactoring of compute-engine for installation on windows. Still work in progress. https://coracle.social/people/nprofile1qy2hwumn8ghj7un9d3shjtnyv9kh2uewd9hj7qgwwaehxw309ahx7uewd3hkctcqyqvzwnnmwp9x2crrtd6v8d6qsnphldqyd3laes0jxt8elxcsvxgvw336psq
———————————————————————————————————————————
21 June milestone has passed but still some work required to package up the new HOP compute engine. This will allow any computation to be performed on a local computer or laptop. (mobile story coming later in July). With all the parts in place, testing, testing, testing and a lot of that is around simplifying the use of time. The key time feature is charting in time series mode and overlay mode: all the days stacked on one 24 hour time period, good way to see trends or potential characteristics of the data. 0.4.0 of BentoBoxDS coming later this week and then v0.0.5 in July will allow DML (decentralized machine learning) or agentic learning in SV speak.
Work started to update to latest release of BentoBoxDS and start automation that will allow biohackers like Siim Land to build their own community analysis: (from Lands newsletter)
going to be like a Netflix of longevity and anti-aging protocols, where
we will launch mini courses on different topics, such as skincare,
liver health, VO2 max, etc. We will also have monthly live Q&A calls
with the group and daily messaging on Whatsapp.
For the last several years, I've read thousands of studies on bloodwork
and have developed a comprehensive understanding about what's optimal
bloodwork for longevity and how to improve it. I've had dozens of
requests from people to look at their bloodwork over the last weeks for
some reason. I'll be coming out with a tutorial plan for interpreting
and analyzing bloodwork in my upcoming community platform, but in the
meanwhile
Open source:
==============================
2025
16 June 2025
Summary: Coherence Ledger
Q How does a coherence ledger change the world?
Nothing
like a big open question to start a new week. In the last couple of
weeks with all the many parts of HOP (Health Oracle Protocol) becoming
understood, coded and becoming stable and made practical to non-coders
via BentoBoxDS, there has been one place all strands of development have
arrived at, the making of an entry in a coherence ledger. Think of that
as the completion of a heart beat for HOP. Technically, a ledger entry
is the cryptographic hash of the results of a computation and a hash of
the proof of work done by a computer to produce those results. In a
Bitcoin transaction: the value in has to equal the value out, but in a
coherence ledger, entries in isolation are open ended, but they provide
enough information to begin to build trust and bring coherence on
solving problems with fellow peers. A way to enable decentralized
machine learning.
like a big open question to start a new week. In the last couple of
weeks with all the many parts of HOP (Health Oracle Protocol) becoming
understood, coded and becoming stable and made practical to non-coders
via BentoBoxDS, there has been one place all strands of development have
arrived at, the making of an entry in a coherence ledger. Think of that
as the completion of a heart beat for HOP. Technically, a ledger entry
is the cryptographic hash of the results of a computation and a hash of
the proof of work done by a computer to produce those results. In a
Bitcoin transaction: the value in has to equal the value out, but in a
coherence ledger, entries in isolation are open ended, but they provide
enough information to begin to build trust and bring coherence on
solving problems with fellow peers. A way to enable decentralized
machine learning.
Product: healthCues
All focus on bentoboxDS 21 june milestone, next milestone healthCues Alpha 30 June.
Open source:
beyond coordination - coherence
Coracle
A nostr client for regular people
https://coracle.social/notes/nevent1qy2hwumn8ghj7un9d3shjtnyv9kh2uewd9hj7qgwwaehxw309ahx7uewd3hkctcqyrv7cuvhwek9glxg676hn7k0xk4av6ry0n2zke32j478m7jxy48nukp3rda
====================================================================
9 June 2025
Summary: Self experimentation
Q What philosophy of science?
In
medical science, the gold standard for a clinical trial is a
triple-blind study. The larger the cohort, the better, or the more peers
involved, the better. There is the theory of trial design, followed by
the implementation, data collection, and publishing of results.
medical science, the gold standard for a clinical trial is a
triple-blind study. The larger the cohort, the better, or the more peers
involved, the better. There is the theory of trial design, followed by
the implementation, data collection, and publishing of results.
AllTrials - Sense about Science
Because evidence matters
https://senseaboutscience.org/alltrials/
This
highlights the manipulation in the industry, which is currently the
backbone of drug discovery. The biohacker generation is emerging with
N=1 studies, presenting not only a direct challenge to clinical trials
but also a paradigm shift in how discoveries are made. Right now, it is
mainly driven by influencers and some exaggerated medical research
papers hype. With AI agents coordinating many N=1 studies within one
peer and across many peers, a new future of health is on the horizon.
highlights the manipulation in the industry, which is currently the
backbone of drug discovery. The biohacker generation is emerging with
N=1 studies, presenting not only a direct challenge to clinical trials
but also a paradigm shift in how discoveries are made. Right now, it is
mainly driven by influencers and some exaggerated medical research
papers hype. With AI agents coordinating many N=1 studies within one
peer and across many peers, a new future of health is on the horizon.
Product: healthCues
Nothing to report.
Opensource:
beebee as an autonomous agent
https://coracle.social/notes/nevent1qy88wumn8ghj7mn0wvhxcmmv9uq32amnwvaz7tmjv4kxz7fwv3sk6atn9e5k7tcqyz8um53xm6v88eps23p87zw704vf8chvvp9satd7eeavhqpgkh8f5h7ysyg
==================================================================
2 June 2025
Summary: Peer experience
Q How to use a BentoBox?
BentoBoxDS
provides all the tools to build a BentoBox. The tools to make reference
contracts held within modules that make up a HOPquery are available via
the Library. This is currently the most time consuming part, however,
once one peer has contributed, e.g. wearable analysis, then all future
peers can make their own 'clone' of that with two clicks. This whole
experience is being improved to make the foundations thatalow AI agents
to support making BentoBoxes and for those BentoBoxes to participate in
machine learning.
provides all the tools to build a BentoBox. The tools to make reference
contracts held within modules that make up a HOPquery are available via
the Library. This is currently the most time consuming part, however,
once one peer has contributed, e.g. wearable analysis, then all future
peers can make their own 'clone' of that with two clicks. This whole
experience is being improved to make the foundations thatalow AI agents
to support making BentoBoxes and for those BentoBoxes to participate in
machine learning.
Product: healthCues
Thinking through how to sync data on first time use to provide incentive to get started e.g. sync the Siim Land Longevity Vibe.
Open source:
Can AI help peers build their own experience
Coracle
A nostr client for regular people
https://coracle.social/notes/nevent1qy2hwumn8ghj7un9d3shjtnyv9kh2uewd9hj7qgwwaehxw309ahx7uewd3hkctcqyru7aq6ra9pen37mjnqf9c9qu56ujqwy03twukgm2rwlq7a8w5hz7eek40a
=========================================================
26 May 2025
Summary: computational universe?
Q What computations matter?
Stephen
Wolfram's life work is based on a belief in a computational universe,
where everything starts from the smallest of computations or rules,
known as cellular automata in technical terms. You can think of a
BentoBox as a cellular automaton, not small, perhaps medium-sized, or
cell-like, that can combine with other BentoBoxes from one peer or many
peers across a peer-to-peer network. Each BentoBox has the autonomous
ability to select the combinations that are best to learn with and
answer questions or solve problems.
Wolfram's life work is based on a belief in a computational universe,
where everything starts from the smallest of computations or rules,
known as cellular automata in technical terms. You can think of a
BentoBox as a cellular automaton, not small, perhaps medium-sized, or
cell-like, that can combine with other BentoBoxes from one peer or many
peers across a peer-to-peer network. Each BentoBox has the autonomous
ability to select the combinations that are best to learn with and
answer questions or solve problems.
Product: heathCues
The
opensource BentoBoxDS latest release was used to update healthCues,
basically bentoboxDS with new branding and UI colours. The next stage is
to add beeplate (or templates) for example; Siim Lands longevity
biomarkers metrics, the Sandra Kaufmann Protocol etc.
opensource BentoBoxDS latest release was used to update healthCues,
basically bentoboxDS with new branding and UI colours. The next stage is
to add beeplate (or templates) for example; Siim Lands longevity
biomarkers metrics, the Sandra Kaufmann Protocol etc.
kaufmannprotocol.com | Kaufmann Protocol Longevity Program
The
Kaufmann Protocol is a scientific, evidence-based system for developing
a personalized anti-aging, longevity protocol. Find supplement
information, longevity books, publicity, and a Create Your Protocol
Quiz.
Kaufmann Protocol is a scientific, evidence-based system for developing
a personalized anti-aging, longevity protocol. Find supplement
information, longevity books, publicity, and a Create Your Protocol
Quiz.
https://kaufmannprotocol.com/
The key will be co-ordining agents to take the base knowledge and personalizing it to each peer.
Open Source:
The
v0.4.4 of HOP included a new compute-engine for SafeFlow, that is being
expanded to allow any computation to be added in a compute contract.
v0.4.4 of HOP included a new compute-engine for SafeFlow, that is being
expanded to allow any computation to be added in a compute contract.
Coracle
A nostr client for regular people
https://coracle.social/notes/nevent1qy2hwumn8ghj7un9d3shjtnyv9kh2uewd9hj7qgwwaehxw309ahx7uewd3hkctcqyqqwkgdz756nlgcvjdu2nhkcggup3nh9v07rvcc7n3dq7qg8mjhnxm2vgjf
========================================================================
19 May 2025
Summary: subsidize, grow, sell (out)
Q. Can the traditional startup venture capital business pattern be shifted?
Being
the earliest of early adopters, you benefit from the latest technology,
a trade off: free access and the startup gains product refinement and
thin air statistics to raise more capital. The latest edition of this
story is OpenAI's acquistion of windsurf
the earliest of early adopters, you benefit from the latest technology,
a trade off: free access and the startup gains product refinement and
thin air statistics to raise more capital. The latest edition of this
story is OpenAI's acquistion of windsurf
OpenAI is acquiring Windsurf for $3 Billion - Community - OpenAI Developer Community
It looks like OpenAI is betting on the AI coding sector:
https://community.openai.com/t/openai-is-acquiring-windsurf-for-3-billion/1253475
Two
clear signs were evident, Windsurf had 'won' a leading early adopoter
market share and they had already done a pricing parntership with openAI
models. Windsurf is a product tool using LLM models, be interesting to
see if it continues to offer competing models from now on? It is hard to
keep up with all the new tools and models but this is a good reminder
to find the deciated open source, high quality teams that will with some
patient surpass the game A, centralized AI cloud game.
clear signs were evident, Windsurf had 'won' a leading early adopoter
market share and they had already done a pricing parntership with openAI
models. Windsurf is a product tool using LLM models, be interesting to
see if it continues to offer competing models from now on? It is hard to
keep up with all the new tools and models but this is a good reminder
to find the deciated open source, high quality teams that will with some
patient surpass the game A, centralized AI cloud game.
Product: healthCues
No progress.
Opens source
Coherence at all scales
Coracle
A nostr client for regular people
https://coracle.social/notes/nevent1qy2hwumn8ghj7un9d3shjtnyv9kh2uewd9hj7qgwwaehxw309ahx7uewd3hkctcqypvwx7nh90zsv64hthq9qw7tw37d5jj7tssyganyf78r4pf7u208602mfm2
=======================================================
12 May 2025
Summary: good network foundations
Q: Is it important to understand what is going on, all the time?
This
is the task of Beebee, the agent that helps each peer manage their
BentoBoxDS. As both warm peer networks are built and peer-to-peer
machine learning is introduced, Beebee's job of directing each peer's
attention to sources of information or results will become essential.
How can Beebee perform this task? By analyzing the coherence ledger at
the heart of each peer's account.
is the task of Beebee, the agent that helps each peer manage their
BentoBoxDS. As both warm peer networks are built and peer-to-peer
machine learning is introduced, Beebee's job of directing each peer's
attention to sources of information or results will become essential.
How can Beebee perform this task? By analyzing the coherence ledger at
the heart of each peer's account.
Product: healthCues
Nothing to report.
Open source:
Keeping track of everythings
The link cannot be previewed. Click to open in a new tab.
https://coracle.social/notes/nevent1qy88wumn8ghj7mn0wvhxcmmv9uq32amnwvaz7tmjv4kxz7fwv3sk6atn9e5k7tcqyp9yxymzgl5lk8z4t8u9wff8nzszlydv7vfc7j9hmtat66gymj6x7qmh0fv
========================================================================
5 May 2025
Summary: HOP Peer-to-Peer Networks
Q: Why a peer-to-peer network?
This
can be answered at a philosophical level. It is a more natural
representation of life, of human social and knowledge-building patterns.
Practically, once a minimum size threshold is established, they provide
resilience properties unmatched by any centralized compute alternative.
Namely, cloud computing or AI in the cloud, where a large data center
owned and controlled by a limited number of people decides what happens.
An open peer-to-peer network is public, permissionless (anyone can join
and improve upon it), and each peer remains sovereign and autonomous.
The challenge is that very few peer-to-peer networks get established at
scale.
can be answered at a philosophical level. It is a more natural
representation of life, of human social and knowledge-building patterns.
Practically, once a minimum size threshold is established, they provide
resilience properties unmatched by any centralized compute alternative.
Namely, cloud computing or AI in the cloud, where a large data center
owned and controlled by a limited number of people decides what happens.
An open peer-to-peer network is public, permissionless (anyone can join
and improve upon it), and each peer remains sovereign and autonomous.
The challenge is that very few peer-to-peer networks get established at
scale.
Product: HealthCues
This
month, there is a big push to get the new website launched and
HealthCues branded BentoBoxDS ready for use. An example showing a
peer-to-peer HealthCues hosted in the cloud will be demonstrated, with
the option to download a local-first application. A bit of an oxymoron:
P2P in the cloud. If we look at Bitcoin, some data suggest that 80%+ of
usage is via the cloud. This will lower again as local compute advances
exponentially in the years ahead.
month, there is a big push to get the new website launched and
HealthCues branded BentoBoxDS ready for use. An example showing a
peer-to-peer HealthCues hosted in the cloud will be demonstrated, with
the option to download a local-first application. A bit of an oxymoron:
P2P in the cloud. If we look at Bitcoin, some data suggest that 80%+ of
usage is via the cloud. This will lower again as local compute advances
exponentially in the years ahead.
Open source
Learn more about the HOP peer to peer network infrastructure.
Coracle
A nostr client for regular people
https://coracle.social/notes/nevent1qy2hwumn8ghj7un9d3shjtnyv9kh2uewd9hj7qgwwaehxw309ahx7uewd3hkctcqyp9yxymzgl5lk8z4t8u9wff8nzszlydv7vfc7j9hmtat66gymj6x7x2rvwh
============================================================
28 April 2025
Summary: how far ahead to pitch product?
Q How to decide the founding feature set?
Or
how visionary should be the initial product be? Too little nobody
notices, too far head people don't see the need? This video sums up the
debate in the context of health and medicine,
how visionary should be the initial product be? Too little nobody
notices, too far head people don't see the need? This video sums up the
debate in the context of health and medicine,
The
conservate (yet poineering MD Attia) and the pro AI Israetel. heathCues
can cater for both but favors the pro AI stance. Taking this position,
there are other segmentations. Peer to Peer verus Cloud. This pro tech
AI health vision of Israetel is cloud based. Will people and practioners
care if its cloud or peer to peer?
conservate (yet poineering MD Attia) and the pro AI Israetel. heathCues
can cater for both but favors the pro AI stance. Taking this position,
there are other segmentations. Peer to Peer verus Cloud. This pro tech
AI health vision of Israetel is cloud based. Will people and practioners
care if its cloud or peer to peer?
Product: healthCues
healthCues
will contain micropayments but should it also contain tokens? This
requires smart contracts, and usally a blockchain, however a peer to
peer approach to tokenization can be achieved using RGB tokens,
will contain micropayments but should it also contain tokens? This
requires smart contracts, and usally a blockchain, however a peer to
peer approach to tokenization can be achieved using RGB tokens,
RGB protocol
Scalable & confidential smart contracts for Bitcoin & Lightning
https://rgb.tech/
Open source
Testing continues, in particular building peer to peer networks, two peer to three peers
Coracle
A nostr client for regular people
https://coracle.social/notes/nevent1qy88wumn8ghj7mn0wvhxcmmv9uq32amnwvaz7tmjv4kxz7fwv3sk6atn9e5k7tcpvemhxue69uhkv6tvw3jhytnwdaehgu3wwa5kuef0dec82c33d3jkc6mgxd5xs7rh895xgamvvdcxkdn3896rq7r589nrw7t6v5c8jvrw0psh5an6w94x6un9xdcrjwrcxdehg6rtweuh50mzwfhkzerrv9ehg0t5wf6k2qg4waehxw309ask2emfwvh82arcduhx7mn99uq3qamnwvaz7tmp9ehx7uewd3hkctcqyql9twkyqyrdp4d7nmnxmr7gktvt2h43khxg065nx0m7yltddar9jevvpus
==========================================================
21 April 2025
Summary: time to build prototype networks
Q What type of networks to build?
Which
use cases apply to which peers? Before suggesting any ideas, it's
important to note that a disruptive event is usually needed to seed a
new paradigm of technology or culture. Those already using peer-to-peer
technology include:
use cases apply to which peers? Before suggesting any ideas, it's
important to note that a disruptive event is usually needed to seed a
new paradigm of technology or culture. Those already using peer-to-peer
technology include:
The Bitcoin community
Nostr users
Some in the biohacker world who demand autonomy and independence
Do
these communities overlap with health? Yes, there are "Bitcoiner" MDs
and health influencers, as well as users of Nostr and Keet.io.
Therefore, it seems likely that some of these individuals will try
HealthCues.
these communities overlap with health? Yes, there are "Bitcoiner" MDs
and health influencers, as well as users of Nostr and Keet.io.
Therefore, it seems likely that some of these individuals will try
HealthCues.
In
terms of the paradigm of practicing health and medicine, functional
medicine MDs and practitioners often have their own independent
practices. They will increasingly face market pressure from the
commoditization of medicine. The opportunity here is to allow these MDs
to maintain their independence by networking together to build sovereign
AI systems that they collectively own and evolve with their community
of peers and patients.
terms of the paradigm of practicing health and medicine, functional
medicine MDs and practitioners often have their own independent
practices. They will increasingly face market pressure from the
commoditization of medicine. The opportunity here is to allow these MDs
to maintain their independence by networking together to build sovereign
AI systems that they collectively own and evolve with their community
of peers and patients.
Product: healthCues
Some proprietary code to setup breeze payment plugin done. Goal to have beta implementation end of May.
Open source
The
countdown to v0.3.0 BETA of BentoBoxDS and HOP is planned for end of
month. Meanwhile, beebee shares some of the ideas driving DML
countdown to v0.3.0 BETA of BentoBoxDS and HOP is planned for end of
month. Meanwhile, beebee shares some of the ideas driving DML
Coracle
A nostr client for regular people
https://coracle.social/notes/nevent1qy2hwumn8ghj7un9d3shjtnyv9kh2uewd9hj7qgwwaehxw309ahx7uewd3hkctcpzemhxue69uhkyetkduhxummnw3erztnrdakj7qgkwaehxw309anhyet9deek7atv9eehqctrv5hsz9rhwden5te0v9kxwmeww468smewdahx2tcqyrmjzknaxtqnj9hx5drqennyp94kde8mxj40rxnn6vrp46z4tk5avc0zc77
========================================================
14 April 2025
Summary: How the future is shaping up.
Q. What will be the main interface of the Internet?
It
started with the command line, then Windows, then the browser, mobile
apps, and next, personal AI assistants. OpenAI set out this future, even
asserting a change in open-sourcing their models (to some degree). This
means directly accessing the web, websites, or apps as we know it will
not be common. Every interaction on the internet will be intermediated
by an AI assistant. With 1 billion users a week, will OpenAI be the only
AI assistant people use? They clearly have a head start, but for those
with a need for sovereignty, data privacy, or autonomy of digital life,
an option is needed too, and that is where we need to focus. The
possibility of DML being a better AI assistant is possible in 3 years'
time, but even then, values will differentiate.
started with the command line, then Windows, then the browser, mobile
apps, and next, personal AI assistants. OpenAI set out this future, even
asserting a change in open-sourcing their models (to some degree). This
means directly accessing the web, websites, or apps as we know it will
not be common. Every interaction on the internet will be intermediated
by an AI assistant. With 1 billion users a week, will OpenAI be the only
AI assistant people use? They clearly have a head start, but for those
with a need for sovereignty, data privacy, or autonomy of digital life,
an option is needed too, and that is where we need to focus. The
possibility of DML being a better AI assistant is possible in 3 years'
time, but even then, values will differentiate.
Product: HealthCues
A
new section where CoherenceStream's main product updates will be
shared. HealthCues will be the first product. The base is BentoBoxDS and
will be prefilled with health cues, biomarkers, and the ability to
connect to others, influencers, products, or professionals. The main
difference will be that HealthCues will offer payments, both pay and
receive. Initially, Bitcoin.
new section where CoherenceStream's main product updates will be
shared. HealthCues will be the first product. The base is BentoBoxDS and
will be prefilled with health cues, biomarkers, and the ability to
connect to others, influencers, products, or professionals. The main
difference will be that HealthCues will offer payments, both pay and
receive. Initially, Bitcoin.
Opensource
BentoBoxDS
/ HOP new release 0.2.9 BETA replicate a whoe library, e.g. public and
cues. This is starting to increase the volume of data being shared peer
to peer. Testers weclome,
/ HOP new release 0.2.9 BETA replicate a whoe library, e.g. public and
cues. This is starting to increase the volume of data being shared peer
to peer. Testers weclome,
Release v0.2.9 BETA H · healthscience/bentoboxds
Allow cloning of public library, cues
Improved design cues
Improved design cues
https://github.com/healthscience/bentoboxds/releases/tag/v0.2.9
=======================================================================
24 March 2025
Summary: Human and automated testing
Q. Turning BentoBoxDS into a commercial product - healthCues
The
https://coherencestream.com website is being redesigned at present
around its first product offering healthCues for newsletters. Three
usecases, for peers, those have subscribed to multiple newsletters, an
AI agent will populate approprate content to bentoSpaces. If you produce
a newsletter, as an influencer, MD or Product e.g. supplement or
redlight therapy devices etc. then you can use an agent to embed @bb
links that will populate bentoSpaces. The win: putting content in
context and while newsletter are free, advance bentoSpace templates
could be charged for. Pricing needs a team discussion. If you to want to
be part of that team? Next Monday 31 March, we will start those
discussions.
https://coherencestream.com website is being redesigned at present
around its first product offering healthCues for newsletters. Three
usecases, for peers, those have subscribed to multiple newsletters, an
AI agent will populate approprate content to bentoSpaces. If you produce
a newsletter, as an influencer, MD or Product e.g. supplement or
redlight therapy devices etc. then you can use an agent to embed @bb
links that will populate bentoSpaces. The win: putting content in
context and while newsletter are free, advance bentoSpace templates
could be charged for. Pricing needs a team discussion. If you to want to
be part of that team? Next Monday 31 March, we will start those
discussions.
Open source:
Keeping complexity in check. Testing, testing, feedback welcome.
Coracle
A nostr client for regular people
https://coracle.social/notes/nevent1qy88wumn8ghj7mn0wvhxcmmv9uq32amnwvaz7tmjv4kxz7fwv3sk6atn9e5k7tcpvemhxue69uhkv6tvw3jhytnwdaehgu3wwa5kuef0dec82c330pmrvctcw4k8scmcxekkxef4d4n8ven6wpehjwpewg6xwet9xda826mww4kx6dp4vdchzurd09mnwd3cxpcn2urcv4snv0mzwfhkzerrv9ehg0t5wf6k2qguwaehxw309a5x7ervvfhkgtnrdaexzcmvv5h8gmm0d3ej7qg5waehxw309askcem09e6hg7r09ehkuef0qqs0aupxwafkhr9ncn0wmfezlqp5gehn2krnlnfzzjqj4sdkc5ga6ugwz8jk8
===============================================================================
17 March 2025
Summary: AI make or break
Q. Shaping the AI options for our future
While
I am skeptical of deep neural nets and LLMs as the path to AGI and
super intelligence, I could be wrong. Each week, using a coding
assistant, I become more impressed with their practical usefulness and
capabilities. They say that software engineers will become obsolete in
18 months—I used to think there was no way, but now it seems possible,
if not in 18 months, then soon after. This is good because it guarantees
that projects like BentoBoxDS and HOP will be built. However, it is
also concerning because all the power lies in the cloud and is owned by a
handful of people. This might not lead us to superintelligence, but
those in control could use their AI to prevent others from getting
there.
I am skeptical of deep neural nets and LLMs as the path to AGI and
super intelligence, I could be wrong. Each week, using a coding
assistant, I become more impressed with their practical usefulness and
capabilities. They say that software engineers will become obsolete in
18 months—I used to think there was no way, but now it seems possible,
if not in 18 months, then soon after. This is good because it guarantees
that projects like BentoBoxDS and HOP will be built. However, it is
also concerning because all the power lies in the cloud and is owned by a
handful of people. This might not lead us to superintelligence, but
those in control could use their AI to prevent others from getting
there.
That
is why a sovereign intelligence, built in the hands and hard drives of
peers, is important. It allows us to have agency over our AI and to
collaborate to produce AI controlled by no one. The next stage of HOP is
to add DML (decentralized machine learning), which can open up a path
to this future. With a bit of patience, this is a slower way of machine
learning, this approach offers an alternative path for AI. It will take
nerve to hold to this path as the cloud LLM's continue to amaze.
is why a sovereign intelligence, built in the hands and hard drives of
peers, is important. It allows us to have agency over our AI and to
collaborate to produce AI controlled by no one. The next stage of HOP is
to add DML (decentralized machine learning), which can open up a path
to this future. With a bit of patience, this is a slower way of machine
learning, this approach offers an alternative path for AI. It will take
nerve to hold to this path as the cloud LLM's continue to amaze.
Open source
Introduction to DML
Coracle
A nostr client for regular people
https://coracle.social/notes/nevent1qy88wumn8ghj7mn0wvhxcmmv9uq32amnwvaz7tmjv4kxz7fwv3sk6atn9e5k7tcpzamhxue69uhhyetvv9ujumn0wd68ytnzv9hxgtcprpmhxue69uhkxetvd3shytnwdaehgu3wwa5kuef0qys8wumn8ghj7cmgdae82uewd45kketyd9kxwetj9e3k7mf6xs6rgtcqyzv7z55ncvedvsq26tvvgmmddefkhrshk5jmnpdp03n3vgewwc89yfrlxys
=========================================================================================
10 March 2025
Summary: Pricing
Q. How to set the price of a product or service?
Given
the input capital put into foundational LLM AI businesses in the cloud,
price, usallaly a monthly subscription price is pretty low, a few tens
of dollars scaling up for businesses. There is a debate whether current
pricing is profitable but winning market share by subsidising early
adopters is a classic silicon valley VC playbook. Open source LLMs are
free to download and cost the electricity of running a laptop or
computer. Free is hard to price against. So what are peers paying for in
open source offerings more broadly? Storage backup, yes free if DIY but
encrypted in the cloud backup, pay a small monthly fee. One last point
to end on. The marginal cost of the product is near nil so if a price
point can be found then the profitabilty margins will be high but will
still require a modest volume of payers to cover staff costs, well until
those are evovled out too.
the input capital put into foundational LLM AI businesses in the cloud,
price, usallaly a monthly subscription price is pretty low, a few tens
of dollars scaling up for businesses. There is a debate whether current
pricing is profitable but winning market share by subsidising early
adopters is a classic silicon valley VC playbook. Open source LLMs are
free to download and cost the electricity of running a laptop or
computer. Free is hard to price against. So what are peers paying for in
open source offerings more broadly? Storage backup, yes free if DIY but
encrypted in the cloud backup, pay a small monthly fee. One last point
to end on. The marginal cost of the product is near nil so if a price
point can be found then the profitabilty margins will be high but will
still require a modest volume of payers to cover staff costs, well until
those are evovled out too.
Opensource:
Eve of DML
Coracle
A nostr client for regular people
https://coracle.social/notes/nevent1qy88wumn8ghj7mn0wvhxcmmv9uq32amnwvaz7tmjv4kxz7fwv3sk6atn9e5k7tcpzemhxue69uhkzat5dqhxummnw3erztnrdakj7qg3waehxw309ucngvpwvcmh5tnfduhszgrhwden5te0vf5hgcm0d9hx6ctcd9kkzmrfwd68xtn0dekxjmn99uqzq204dyz7exsrft2yef6qluu6ehymdx2wmfusk8kh5x9rsuxh4n6ycx6z3z
=======================================================================================
3 March 2025
Summary: What content is of value? What will peer pay for?
Q. What content is best to input and prioritize?
Some
peers have to put in the effort to create cues, fill spaces with
content, and that information must be of such value that other peers
will want to receive it. Can this be done without a monetary incentive?
Should peers be given the option to share, but for a price? There is a
catch. In a peer-to-peer world, once that content is shared, the peer
has the ability to redistribute it, where they could charge and cut out
the original creator. This has both positive and negative implications.
On the negative side, it potentially limits the income of the peer who
put in all the work. However, on the positive side, if the content is
valuable, it gets shared and reaches more peers, thus allowing the
network to grow. Personalization is one way to address this challenge.
The initial template is free, but personalized content comes at a cost.
Why would a peer want to pass on their personalized solution
(considering privacy), while the peer receiving it would want their own
personalized solution, rather than that of their peers?
peers have to put in the effort to create cues, fill spaces with
content, and that information must be of such value that other peers
will want to receive it. Can this be done without a monetary incentive?
Should peers be given the option to share, but for a price? There is a
catch. In a peer-to-peer world, once that content is shared, the peer
has the ability to redistribute it, where they could charge and cut out
the original creator. This has both positive and negative implications.
On the negative side, it potentially limits the income of the peer who
put in all the work. However, on the positive side, if the content is
valuable, it gets shared and reaches more peers, thus allowing the
network to grow. Personalization is one way to address this challenge.
The initial template is free, but personalized content comes at a cost.
Why would a peer want to pass on their personalized solution
(considering privacy), while the peer receiving it would want their own
personalized solution, rather than that of their peers?
Open source
Introduction
to the four content use cases for March
https://coracle.social/notes/nevent1qy88wumn8ghj7mn0wvhxcmmv9uq32amnwvaz7tmjv4kxz7fwv3sk6atn9e5k7tcpzpmhxue69uhkztnwdaejumr0dshszxmhwden5te0vd58ymmwd93kcefwv36x7mn0dchxxmmd9uq3qamnwvaz7tm99ehx7uewd3hkctcqyz2u94sgxu5956d7hncn8s3zj9xyjknz6jhn355mh9me6l9hrwyj5l3y2a6
to the four content use cases for March
https://coracle.social/notes/nevent1qy88wumn8ghj7mn0wvhxcmmv9uq32amnwvaz7tmjv4kxz7fwv3sk6atn9e5k7tcpzpmhxue69uhkztnwdaejumr0dshszxmhwden5te0vd58ymmwd93kcefwv36x7mn0dchxxmmd9uq3qamnwvaz7tm99ehx7uewd3hkctcqyz2u94sgxu5956d7hncn8s3zj9xyjknz6jhn355mh9me6l9hrwyj5l3y2a6
========================================================================
24 February 2025
Summary: Building networks peer to peer - uses cases
Q. What use cases match the capability of the tools?
We
want unregulated activities initially; sharing knowledge and research
seems like a good use of the technology. However, we still need specific
use cases. "The Wikipedia of me"—the Wikipedia website provides
knowledge about the human body, but now peers can build personalized
spaces that they find informative. While every peer has this capability,
a few will create these "space templates" initially, e.g., around heart
health or fertility, and those will be "cloned" or shared by others.
Then, Beebee can learn from each peer's personal context and further go
out to the internet or self-generate content. From this starting place,
we can bring in diagnostic and biomarker data and analysis. Where the
line on regulation is drawn will need to be explored.
want unregulated activities initially; sharing knowledge and research
seems like a good use of the technology. However, we still need specific
use cases. "The Wikipedia of me"—the Wikipedia website provides
knowledge about the human body, but now peers can build personalized
spaces that they find informative. While every peer has this capability,
a few will create these "space templates" initially, e.g., around heart
health or fertility, and those will be "cloned" or shared by others.
Then, Beebee can learn from each peer's personal context and further go
out to the internet or self-generate content. From this starting place,
we can bring in diagnostic and biomarker data and analysis. Where the
line on regulation is drawn will need to be explored.
Open source
Generativing
invites peer to peer
https://coracle.social/notes/nevent1qy2hwumn8ghj7un9d3shjtnyv9kh2uewd9hj7qgwwaehxw309ahx7uewd3hkctcpzdmhxue69uhhqatjwpkx2urpvuhx2ue0qys8wumn8ghj7cmgdae82uewd45kketyd9kxwetj9e3k7mf6xs6rgtcprpmhxue69uhkxetvd3shytnwdaehgu3wwa5kuef0qqsdykh64pk7yxjnz057n70vgrchj3gpjv8ph3tlqlpe2fe68uvzfqqls48ex
Invite to join coming soon.
invites peer to peer
https://coracle.social/notes/nevent1qy2hwumn8ghj7un9d3shjtnyv9kh2uewd9hj7qgwwaehxw309ahx7uewd3hkctcpzdmhxue69uhhqatjwpkx2urpvuhx2ue0qys8wumn8ghj7cmgdae82uewd45kketyd9kxwetj9e3k7mf6xs6rgtcprpmhxue69uhkxetvd3shytnwdaehgu3wwa5kuef0qqsdykh64pk7yxjnz057n70vgrchj3gpjv8ph3tlqlpe2fe68uvzfqqls48ex
Invite to join coming soon.
==========================================================
17 February 2025
Summary: Finding the early adopters
Next Monday, last of the month. Video call 5pm UK. Feedback from using BentoBoxDS.
Q. Who will use the tools: 80:20 rule
With
the open-source BentoBoxDS becoming more useful by the week, who will
put in the effort to populate the networks with content and share? There
is an established trend from blogging and tweeting to video content,
etc. Those who create the content are in the minority, while those
consuming the content make up the majority, approximately 80%. Although
BentoBoxDS focuses on personalization, this inherently requires peer
input, and setting up the templates to populate will involve a minority
group of peers. For example, onboarding all the content related to the
human body—such as biomarkers, videos, research papers, data analysis
and charting and products—initially requires human input, resulting in a
template that can be shared. As this content builds, AI agents can
support and likely, in the medium term, produce the templates as needed
for each peer. Can we find those willing to put in the effort?
the open-source BentoBoxDS becoming more useful by the week, who will
put in the effort to populate the networks with content and share? There
is an established trend from blogging and tweeting to video content,
etc. Those who create the content are in the minority, while those
consuming the content make up the majority, approximately 80%. Although
BentoBoxDS focuses on personalization, this inherently requires peer
input, and setting up the templates to populate will involve a minority
group of peers. For example, onboarding all the content related to the
human body—such as biomarkers, videos, research papers, data analysis
and charting and products—initially requires human input, resulting in a
template that can be shared. As this content builds, AI agents can
support and likely, in the medium term, produce the templates as needed
for each peer. Can we find those willing to put in the effort?
Opensource
Playing
it medium with intelligence, HOP BETA road map
https://design.penpot.app/#/view/f329de2c-67b1-801b-8005-bb46f48bfa18?page-id=f329de2c-67b1-801b-8005-bb46f48bfa19§ion=interactions&index=0&share-id=0b127ab7-8934-814e-8005-bb4e010473d3
it medium with intelligence, HOP BETA road map
https://design.penpot.app/#/view/f329de2c-67b1-801b-8005-bb46f48bfa18?page-id=f329de2c-67b1-801b-8005-bb46f48bfa19§ion=interactions&index=0&share-id=0b127ab7-8934-814e-8005-bb4e010473d3
===================================================================================
10 February 2025
Summary:
Networks are all about sharing by definition. What to share, when to
share, and privacy are all key considerations. From these foundations,
the network will decide what is valuable.
Networks are all about sharing by definition. What to share, when to
share, and privacy are all key considerations. From these foundations,
the network will decide what is valuable.
Q. What competition does CoherenceStream have?
In
the slide deck, there are three main categories of competition: the
status quo - how things are done today; Cloud AI - how centralized AI
and venture capital sell to the corporate world and directly to
consumers; and Web3, which includes cryptocurrencies, tokens, and DAOs,
etc. This discussion is about the latter, Web3, was featured in the
Optispan podcast https://youtu.be/MXxG_29_VKY?si=qgSksVrmLHPQalEn last
week. This captures the craziness of this world or genius, depending on
how you see it. In effect, they want to create meme coins (the thinner
the air, the better) to establish new token markets where some of the
trading fees are invested in health research. Of course, this model
holds if tokens have value; the whole space has created a trillion
dollars out of thin air, so maybe they are right?
the slide deck, there are three main categories of competition: the
status quo - how things are done today; Cloud AI - how centralized AI
and venture capital sell to the corporate world and directly to
consumers; and Web3, which includes cryptocurrencies, tokens, and DAOs,
etc. This discussion is about the latter, Web3, was featured in the
Optispan podcast https://youtu.be/MXxG_29_VKY?si=qgSksVrmLHPQalEn last
week. This captures the craziness of this world or genius, depending on
how you see it. In effect, they want to create meme coins (the thinner
the air, the better) to establish new token markets where some of the
trading fees are invested in health research. Of course, this model
holds if tokens have value; the whole space has created a trillion
dollars out of thin air, so maybe they are right?
Open Source
We
aim for fortnightly releases, with the next one, 17 Feb focused on
peer-to-peer long-term relationships and sharing, followed by Agents,
which will add a LLM by the end of the month, 24 Feb.
aim for fortnightly releases, with the next one, 17 Feb focused on
peer-to-peer long-term relationships and sharing, followed by Agents,
which will add a LLM by the end of the month, 24 Feb.
====================================================================
3 Feb January 2025
Summary: day before HOP BentoBoxDS beta
Q. What data will be most actionable in the future?
The
trend of smaller, cheaper, and better resolution in existing technology
will continue, from wearables to gene sequencing to blood diagnostics,
etc. But will there be new technology that makes these obsolete?
Nanotechnology is one area where such advancements could emerge. Source.
trend of smaller, cheaper, and better resolution in existing technology
will continue, from wearables to gene sequencing to blood diagnostics,
etc. But will there be new technology that makes these obsolete?
Nanotechnology is one area where such advancements could emerge. Source.
Open source
All energies on Beta preview https://rumble.com/v6g3r04-bentoboxds-hop-beta-i-preview.html
=============================================================
27 January 2025
Summary: Data debate discussion wireapp 5pm UK timezone
For sovereign data
ownership from source, know how measurements are made
Hardware to maintain, intro cryptography
Sets trends for other to contribute this way
Regulation security needs address
For API data or third party
Existing data exists
No hardware skills required
Too many devices to keep up with
Accepting regulation choices of others
Strategy is to mix opportunies/adoption cycles with state of tech and regulation
Open source
Cues and relationships to biomarkers https://rumble.com/v6dyyw7-cue-relationships-to-biomarkers.html
==================================================================================================================
20 January 2025
Summary
Learning
from other open source businesses. Wordpress and Automattic Inc are
formative open source and commercialisation influence. A more up to
date setup would be https://appflowy.io/
Hosting in the cloud, host yourself or they can host for a fee. This
does not sound too dis-similar to CoherenceStream envisioned setup.
Privacy, own data and local AI all key marketing messages put forward by
AppFlowy. One of their funders is Automattic founder BTW.
from other open source businesses. Wordpress and Automattic Inc are
formative open source and commercialisation influence. A more up to
date setup would be https://appflowy.io/
Hosting in the cloud, host yourself or they can host for a fee. This
does not sound too dis-similar to CoherenceStream envisioned setup.
Privacy, own data and local AI all key marketing messages put forward by
AppFlowy. One of their funders is Automattic founder BTW.
Q Which sovereign data is most important?
The
answer to that will depend on the use case but I think we can spell out
some properties. It is far better to have data from a device or sensor
authored directly into a sovereign data store than to get that data via
a third party API. Data that is most actionable given current
knowledge and lastly we should be mindful that the best data will come
from devices or sensors not in existence today. The need for tight
partnerships in the manufacturing of sovereign devices seems a must now
and going forward.
answer to that will depend on the use case but I think we can spell out
some properties. It is far better to have data from a device or sensor
authored directly into a sovereign data store than to get that data via
a third party API. Data that is most actionable given current
knowledge and lastly we should be mindful that the best data will come
from devices or sensors not in existence today. The need for tight
partnerships in the manufacturing of sovereign devices seems a must now
and going forward.
Open Source
Cues video posted last week https://rumble.com/v69ncva-cues-bentoboxds.html . Introducing Marker or Biomarkers this week, https://coracle.social/notes/note1l9g4nmre3el0rqh0zaznntcuwra8uvl7eykzcszm7k67e0aa9qnszn3fj4
—-----------------------------------------------------------------------------------------
13 January 2025
Summary
I
have been thinking that a big shift has started that will take us away
from online advertising and e-commerce sales. The two dominant income
streams of monetization online, fuels the tech giants. TV ads to
Youtube ads to influencer sponsorship. What is the next stage? AI
intermediate or dis-intermediation? This opens up the following
question:
have been thinking that a big shift has started that will take us away
from online advertising and e-commerce sales. The two dominant income
streams of monetization online, fuels the tech giants. TV ads to
Youtube ads to influencer sponsorship. What is the next stage? AI
intermediate or dis-intermediation? This opens up the following
question:
Q What will be the motivation or influence for those AIs?
You
can envision a range. Some will set out to mimic the influencers of now
and we can see AI purposed to only look after the peers. For example, I
am told the best purchase price on amazon in the 17th item in the query
list. An Peer AI would notice that and suggest it but we would want
much more than that, e.g. given a health goal this is the best product?
can envision a range. Some will set out to mimic the influencers of now
and we can see AI purposed to only look after the peers. For example, I
am told the best purchase price on amazon in the 17th item in the query
list. An Peer AI would notice that and suggest it but we would want
much more than that, e.g. given a health goal this is the best product?
Open source
Cues are operational in code https://coracle.social/notes/nevent1qy88wumn8ghj7mn0wvhxcmmv9uq32amnwvaz7tmjv4kxz7fwv3sk6atn9e5k7tcpzamhxue69uhkvun9deejumn0wd68yvfwvdhk6tcpz4mhxue69uhkvun9deejuat50phjummwv5hszxthwden5te0vaex7atswvhxv6tpw34xze3wvdhk6tcqyr5upgqtu3yhr3as9zm8tayxlqrz9rl7erwuxgca56e7m2kjyyqsc5qhnq9
—-----------------------------------------------------------------------------
6 January 2025
Summary
Focus
on getting open source to beta and then assess the commercialisation
February. On the AGI front, over the festive period we have seen
reinforcement learning on top of a LLM to bring new capabilities e.g.
reasoning. We will see more of this until a new foundation training run
comes. And Deepseek claiming open source LLM as good a OpenAI with
spending only $6m of compute training? A trend to small and
decentralised from an unlikely source.
on getting open source to beta and then assess the commercialisation
February. On the AGI front, over the festive period we have seen
reinforcement learning on top of a LLM to bring new capabilities e.g.
reasoning. We will see more of this until a new foundation training run
comes. And Deepseek claiming open source LLM as good a OpenAI with
spending only $6m of compute training? A trend to small and
decentralised from an unlikely source.
**Q.
Can you have sovereign data without sovereign hardware?**Hardware in
two meanings, a sensor device: wearable to blood analysis to MRI scanner
and computing: mobile, laptops that will store and compute algorithms.
Do we harder path of partnering with hardware manufacturers or settle
for API data access?
Can you have sovereign data without sovereign hardware?**Hardware in
two meanings, a sensor device: wearable to blood analysis to MRI scanner
and computing: mobile, laptops that will store and compute algorithms.
Do we harder path of partnering with hardware manufacturers or settle
for API data access?
Open sourceBeta
release date 4/2/2025. A open dialogue paper will be issues to explain
what HOP, TINY agentic systems can be used over a Peer to Peer network.
release date 4/2/2025. A open dialogue paper will be issues to explain
what HOP, TINY agentic systems can be used over a Peer to Peer network.
—----------------------------------------------------------------------------------------------------------------
23 December
Summary
Creating
simplified messaging and explaining. To date, being tech driven, the
core peer and business benefits have been weak. Here are some new
messages to be thought about https://docs.google.com/presentation/d/1kF5yHLyjExrqfJcwlRNc5BDBolaKQWee7LFuj9j43_4/edit?usp=sharing
simplified messaging and explaining. To date, being tech driven, the
core peer and business benefits have been weak. Here are some new
messages to be thought about https://docs.google.com/presentation/d/1kF5yHLyjExrqfJcwlRNc5BDBolaKQWee7LFuj9j43_4/edit?usp=sharing
Q. Data 2025
There
is very little sovereign data on the internet, such is the dominance
and appeal of cloud convenience. We need to debate and commit to a data
strategy in January. There are two competing paths, use API’s e.g. get
wearable data from https://ouraring.com or invest in our own sovereign hardware e.g. own brand via a partner e.g. https://www.colmi.com/colmi-r02-smartring-ip68-3atm-waterproof-fitness-tracker-smart-ring-product/
and add cryptography to save data directly to a sovereign datastore.
Which data will be most actionable in the next 18 months? Create a top
five. Our role is to open up this trend and let an industry broaden
out. This trend has started with Umbrel, Star9 home storage /computers.
is very little sovereign data on the internet, such is the dominance
and appeal of cloud convenience. We need to debate and commit to a data
strategy in January. There are two competing paths, use API’s e.g. get
wearable data from https://ouraring.com or invest in our own sovereign hardware e.g. own brand via a partner e.g. https://www.colmi.com/colmi-r02-smartring-ip68-3atm-waterproof-fitness-tracker-smart-ring-product/
and add cryptography to save data directly to a sovereign datastore.
Which data will be most actionable in the next 18 months? Create a top
five. Our role is to open up this trend and let an industry broaden
out. This trend has started with Umbrel, Star9 home storage /computers.
Looking forward to engaging in this conversation in the new year.
Open source
https://coracle.social/notes/nevent1qy88wumn8ghj7mn0wvhxcmmv9uq32amnwvaz7tmjv4kxz7fwv3sk6atn9e5k7tcpzpmhxue69uhk2tnwdaejumr0dshszymhwden5te0vdkrgtn5de5hstnyv4mz7qgcwaehxw309anxjmr5v4ezumn0wd68ytnhd9hx2tcqyqd3v9j5t08lavazartgk2cqke43c9zeuau6dsz6vn4uj3vd7mekjdsm5xu
—---------------------------------------------------------------------------------------
16 December 2024
Summary
All
big and startup AI tech businesses are after the health market, $2
trillion in annual spending in the US alone, is the prize. Many will
think reform of disease management is the win, there will be wins but
the transformation to prevention is the revolution.
big and startup AI tech businesses are after the health market, $2
trillion in annual spending in the US alone, is the prize. Many will
think reform of disease management is the win, there will be wins but
the transformation to prevention is the revolution.
Q. How to fund peer to peer tech?
It
is still not the first choice of technology. You do not see big tech
striving for a peer to peer architecture, web3 at least aims for a
federation of AI, ASI token to Bitsensor. We do have one example to
learn from, holepunch https://tether.io/news/tether-bitfinex-and-hypercore-launch-holepunch-a-platform-for-building-fully-encrypted-peer-to-peer-applications/
HOP uses there peer to peer network for messages and storage. You
have tech genius founder of holepunch (seen as that in javascript world)
and the rise of the most profitable ‘bank’ in the world this year
Tether. Deep pockets to try and fund a peer to peer transition.
Holepunch, have PEAR networking product and Keet.io an application to
test to the limits of their peer to peer tech. So, is the learning we
need to find such an alliance in the health domain? Another example
would be Jack Dorsey sponsoring Nostr into existence. Satohsi Nakamoto
has shown ‘no investment’ is possible, but maybe this only works for
when money is the core product? Peer to Peer holds great potential but
is still not the first choice of technology. You do not see big tech
striving for a peer to peer architecture, web3 at least aims for a
federation of AI, ASI token to Bitsensor. We do have one example to
learn from, holepunch https://tether.io/news/tether-bitfinex-and-hypercore-launch-holepunch-a-platform-for-building-fully-encrypted-peer-to-peer-applications/
HOP uses there peer to peer network for messages and storage. You
have tech genius founder of holepunch (seen as that in javascript world)
and the rise of the most profitable ‘bank’ in the world this year
Tether. Deep pockets to try and fund a peer to peer transition.
Holepunch, have PEAR networking product and Keet.io an application to
test to the limits of their peer to peer tech. So, is the learning we
need to find such an alliance in the health domain? Another example
would be Jack Dorsey sponsoring Nostr into existence. Satohsi Nakamoto
has shown ‘no investment’ is possible, but maybe this only works for
when money is the core product? Peer to Peer holds great potential but
it's still rare for it to come to be without substantial financial
support.
support.
Open source
Getting from zero to one . https://coracle.social/notes/note1fh56ca5rlqctfxst85lpxe9wfzsvzpkhgcllfcfm8n7jajm5ktjstcu2ar
—------------------------------------------------------------------------------------------
9 December 2024
Summary
Longevity market: while it is still small, especially EU, the trend is much strongly established than most folk know. https://www.flarecapital.com/insight/flare-capital-expert-roundtable-series-17-longevity-in-focus-innovating-for-a-longer-healthier-life
Q. how do longevity products and health influencer interplay?
This podcast was a good insight to understanding https://youtu.be/GDt6IQMXWxc?si=4i9vP6pRVkMIXVrg
Find science to ground a product in, what ‘alpha’ health influencers
are repeatedly promoting / messaging, then find niche consumers and meet
them where they are at, then rebrand for mass market and sell out. A
simplification but combined with a shift in social media trend that
‘followers’ do not matter anymore as feeds and going forward AI agents
are the new intermediary. Influencers need to ‘score’ a hit
occasionally with the AI and then use that to onboard a relationship and
a newsletter is one of the main ways of doing that. LHA can be looked
at as a ‘fancy’ newsletter but with more robust content and all can be
personalised. How do we make this pattern come to be?
Find science to ground a product in, what ‘alpha’ health influencers
are repeatedly promoting / messaging, then find niche consumers and meet
them where they are at, then rebrand for mass market and sell out. A
simplification but combined with a shift in social media trend that
‘followers’ do not matter anymore as feeds and going forward AI agents
are the new intermediary. Influencers need to ‘score’ a hit
occasionally with the AI and then use that to onboard a relationship and
a newsletter is one of the main ways of doing that. LHA can be looked
at as a ‘fancy’ newsletter but with more robust content and all can be
personalised. How do we make this pattern come to be?
Tiny
Small and decentralised, another example this is possible https://youtube.com/clip/UgkxcEh2MEUL7sL77j5rCGdHd_exzvlHjcR4?si=mL6FIRVSvBzaEqhc Imagine that each location ran in 9 minutes and aggregated all in 60 minutes daily?
Source source
Beebee update
—-------------------------------------------------------------------------------------------------------------
2 Dec 2024
Change
up for Mondays. From 9th December 2024 Mondays will be a written
question and answer format. As LHA becomes used, help and support
becomes the priority and dedicated support hours will be offered in the
new year. When needed, probably once a month or when some big stage or
decision is upon us then a video call will be arranged. So, post your
question or issue on Wireapp (probably Signal App from 1 Jan 2025) and
they will be responded to by the following Monday.
up for Mondays. From 9th December 2024 Mondays will be a written
question and answer format. As LHA becomes used, help and support
becomes the priority and dedicated support hours will be offered in the
new year. When needed, probably once a month or when some big stage or
decision is upon us then a video call will be arranged. So, post your
question or issue on Wireapp (probably Signal App from 1 Jan 2025) and
they will be responded to by the following Monday.
Summary
A
big effort on CoherenceStream business thinking, plans, lists of early
adoptor targets have been gathered, spreadsheet, slides constructed etc.
has been put in the last two months. All this has helped with reaching
out to new people to start the goal of getting them to join the
business. I am enthused about the people we have reached and their
feedback is helping greatly to understand starting opportunities.
However, no one to date has reached the stage where they are willing to
stop all other activity in their life and dedicate it to
CoherenceStream. That needs to change.
big effort on CoherenceStream business thinking, plans, lists of early
adoptor targets have been gathered, spreadsheet, slides constructed etc.
has been put in the last two months. All this has helped with reaching
out to new people to start the goal of getting them to join the
business. I am enthused about the people we have reached and their
feedback is helping greatly to understand starting opportunities.
However, no one to date has reached the stage where they are willing to
stop all other activity in their life and dedicate it to
CoherenceStream. That needs to change.
There
is no cash to pay wages, no income and no adoption. LHA will be in
Alpha by the end of 2024 so 2025 is all about adoption. When this
comes, that should trigger team adoption and help with recruitment.
Investment will be sought in December to fund the LAB from two/three
folk but without a team or adoption this is still a long shot.
is no cash to pay wages, no income and no adoption. LHA will be in
Alpha by the end of 2024 so 2025 is all about adoption. When this
comes, that should trigger team adoption and help with recruitment.
Investment will be sought in December to fund the LAB from two/three
folk but without a team or adoption this is still a long shot.
Aaron
has been asking some telling questions: Biology is complex so how can
we model / simulate that in a decentralised manner when the centralised
way currently struggles? This is a key idea to grasp in this project:
Top down and bottom up evolution. Most folk think about Darwin bottom
up evolution and that is the mindset of big centralised AI, model
everything and throwing GPU horsepower running a NeuralNet at the data
to learn at great energy and cost (money and time). We take a different
approach: a combination of both top down and bottom up. We do not set
out to manage all the details. They are collective intelligences in
themselves so we just want to ‘perform the training’ via our peer to
peer machine learning. This will provide the necessary information to
‘give to’ the local self organisation. And this applies at the Cell
level, network level and the design of morphoceuticals. Get this
training right then bottom up self organisation will complete the work
needing done. The training can be done peer to peer (to be proven ) in a
has been asking some telling questions: Biology is complex so how can
we model / simulate that in a decentralised manner when the centralised
way currently struggles? This is a key idea to grasp in this project:
Top down and bottom up evolution. Most folk think about Darwin bottom
up evolution and that is the mindset of big centralised AI, model
everything and throwing GPU horsepower running a NeuralNet at the data
to learn at great energy and cost (money and time). We take a different
approach: a combination of both top down and bottom up. We do not set
out to manage all the details. They are collective intelligences in
themselves so we just want to ‘perform the training’ via our peer to
peer machine learning. This will provide the necessary information to
‘give to’ the local self organisation. And this applies at the Cell
level, network level and the design of morphoceuticals. Get this
training right then bottom up self organisation will complete the work
needing done. The training can be done peer to peer (to be proven ) in a
more accurate, secure and efficient manner than a centralised AI. This
is what we need to demonstrate in 2025. How credible is this approach?
Dr. Michael Levins work provides the best research evidence, https://drmichaellevin.org/resources/
is what we need to demonstrate in 2025. How credible is this approach?
Dr. Michael Levins work provides the best research evidence, https://drmichaellevin.org/resources/
Open Source
CoherenceStream
is committed to giving back to the open source projects it uses e.g.
BentoBoxDS - HOP and other enabling tech like holepunch etc.
is committed to giving back to the open source projects it uses e.g.
BentoBoxDS - HOP and other enabling tech like holepunch etc.
The latest updates are provided only on the community discord for BentoBox and HOP
When
you spend years coding in isolation to bring a novel opportunity to the
world, it is rewarding for you all to give up time on a Monday learn
and contribute. This is the fuel that keeps me going . Thank you. 2025
requires adoption, let's see where that takes us. James
you spend years coding in isolation to bring a novel opportunity to the
world, it is rewarding for you all to give up time on a Monday learn
and contribute. This is the fuel that keeps me going . Thank you. 2025
requires adoption, let's see where that takes us. James