(project founded and commenced on Friday January 19, 2024)
"A rose 🌹 by any other name..."
This project's GitHub repository gets its name, along with the initiative's working codename (Project FireSeed) and purpose, and is contextualised and informed by: 'Seed Vault' and associated and nested systems and technologies, along with general and specific prepping technologies, processes and systems. The project's repository and codename are informed by:
first, resilient, adaptive and archivable food-system seed vault processes, protocols and technologies; and secondly, by possibly the first human technology that ever developed, the adaptation to, and harnessing of: the boon/bane of fire. 🔥 Both 'seed' and 'fire' are potent and primal metaphors that are pervasive throughout the human condition, and fundamental to human culture and its continuity.
I am principally learning *NIX shell 🐚 scripting (and attendant website codification, through markdown, as website hosting is graciously available for GitHub repositories, by its patron and owner Microsoft) through roadtesting my shell scripting roadmap, that is 'A Fable of the Scriptkiddie Scriptoria', to then progress to Python 🐍 scripting and programming in order to develop this project, which is a free and open source, localised GPT and associated toolkit and auditing sustem, that functions as a knowledge 'seed vault' to kick-start humanity's regeneration, thriving and community development and capacity, in the eventuation of a cataclysm or apocalypse, possibly as a result of rogue AI technology, or somesuch. So essentially, by metaphorical extension from 'seed' and 'fire', it is a pragmatic, practical and functional set of protocols, tools and processes, to empower humans with associated technological automation, to realise the project's and programme's digital and non-digital objectives.
I envision this repository to house an actionable script (or suite of scripts), along with all the necessary information and processes and workflows, nested and embedded in an interactive and user-friendly, graphical and portable, user interface (GUI) environment, along with the markdown of the project's website, graciously hosted by GitHub and Microsoft. The website should seek to showcase all aspects of the project, its timeline and development, and seek to draw active skilled and/or interested and passionate parties to the project.
I hope that the prototyping and functionality of this project will subsequently be optimised for diverse systems and purposes, in manifold emergent and tangential projects and initiatives.
I feel that this is the most appropriate metaphorical section in our archetypal vertiginous and fecund project outline, to list the required skills (from a rudimentary or skill audit), to realise a working prototype:
Our SeedVault.L-GPT / Project FireSeed, is an ambitious and innovative initiative, that aims to create localised, portable and decentralised GPT, associated toolkit and auditing system, that functions as an interrogable and actionable knowledge 'seed vault' repository, with concomitant and associated technology and information, for humanity in case of a cataclysm or apocalypse. It is initially inspired by b9Joker108’s lived experience of the world and resultant worldview, as well as initially informed by a YouTube video (embedded herewith) that showcases the rudimentary process and framework, of how to capture, store and secure the fundamental aspects of human knowledge in an intelligent and interrogable system, that is localised, decentralised, portable and founded upon free and open-source, large language models (LLMs). I intuit that liaising with the resplendent Anna's Archive, to amass a targeted dataset for the training and tuning of the prospective LLMs, will be an imperative. As LLMs are developing and iterating with breakneck speed, it is vital that within the project's auspices, that a process is automated, as much as is pragmatic and save to implement, by hard-wiring and soft-coding systems, that seeks to ensure that the LLMs that are selected, remain relevant and appropriate given development, and changing technologies and circumstances. This should be auditable and for the most part, automatable, but must be checked, assessed and benchmarked by a centralised and de- centralised committee of informed and empassioned individuals.
Based on a rudimentary project analysis, the following skill areas have been identified in a generalised project skills' audit, to realise a working project prototype and bring it to fruition:
-
Shell scripting and Python programming: These are the core skills that you will need to create the scripts and the graphical user interface (GUI) for your project. You will also need to be familiar with the LLM framework that you choose to use, such as GPT-3, GPT-Neo, or GPT-J. You will need to know how to download, install, run, and update the LLM on your local drive, as well as how to interact with it using LM Studio or other tools. You will also need to know how to automate the processes and handle any errors or exceptions that may occur.
-
Data science and machine learning: These are the skills that you will need to understand the underlying principles and mechanisms of the LLM, as well as how to fine-tune it for your specific needs and goals. You will need to know how to preprocess, analyse, and visualise the data that you feed to the LLM, as well as how to evaluate its performance and quality. You will also need to know how to generate various types of text formats using the LLM, such as summaries, instructions, how-to-do guides, teaching stories, inspiring poems and songs, etc., to seed cultural reflowering.
-
Information security and cryptography: These are the skills that you will need to ensure the safety and integrity of your data and your LLM. You will need to know how to encrypt, decrypt, and authenticate your data and your LLM, as well as how to protect them from unauthorized access, modification, or deletion. You will also need to know how to backup and restore your data and your LLM in case of emergencies, as well as how to detect and prevent any potential threats or attacks.
-
Project management, storytelling and communication: These are the skills that you will need to plan, execute, monitor, and control your project, as well as to communicate effectively with your stakeholders, collaborators, and users. You will need to know how to define the scope, objectives, deliverables, and milestones of your project, as well as how to allocate the resources, time, and budget. You will also need to know how to manage the risks, issues, and changes that may arise during the project lifecycle, as well as how to report the progress and outcomes of your project. You will also need to know how to document your project and your LLM, as well as how to provide user support and feedback.
-
Digital prepping: This is the skill of creating, storing, and accessing a digital library of useful information and resources that can be used offline or in a grid-down situation. You will need to know how to:
- Choose and download an open-source large language model (LLM) that can provide you with answers, instructions, and creative content, such as GPT-3, GPT-Neo, or GPT-J.
- Install and run the LLM on a local drive, such as an external hard drive or a solid state drive, using a graphical user interface (GUI) such as LM Studio.
- Power and maintain your computer and drive using a battery backup, a solar panel, or other alternative sources.
- Encrypt, backup, and restore your data and LLM, as well as protect them from unauthorized access, modification, or deletion.
- Download and store other sources of free knowledge, such as Wikipedia, Khan Academy, OpenStreetMap, E-Books, and others, using tools such as Internet-in-a-Box.
- Organize, search, and update your digital library, as well as share it with others in your network or community.
-
General prepping: This is the skill of planning, preparing, and surviving for various emergencies, disasters, and crises that may affect you, your family, or your community. You will need to know how to:
- Build a solid personal finance and health foundation, such as saving money, paying off debts, getting insurance, staying fit, and having a regular check-up.
- Get your home ready for two weeks of self-reliance, such as storing water, food, medicine, and other essentials, as well as having a fire extinguisher, a first aid kit, and a communication device.
- Be able to leave your home with only a moment's notice, such as having a bug out bag, a vehicle, a destination, and a route.
- Prepare for emergencies that happen away from home, such as having a get home bag, an everyday carry kit, and a survival mindset.
- Learn core skills and practice with your gear, such as fire making, water purification, shelter building, navigation, self-defense, and first aid.
- Share and recruit while continuing to learn and going beyond the basics, such as joining a prepper group, taking advanced courses, reading books, and watching videos.
Just as the project requires a skill audit of fundamental and necessary skills required to realise it, so too, all those who join the project, should similarly undergo the process of a skills audit. There should be some way to securely house and locally store and locate, this information and process it into a secure database, that is interrogable and reportable, to inform the membership of project working groups. Skills are only one aspect of generating efficient and inspired project working groups, its compliment is passion or interest,mwhich may or may not be so-skilled, but which should be encouraged, included, cherished and actively and demonstratively cultivated. Such is the lifeblood, the veritable sap, of any human project or activity.
I anticipate this project, to map, coordinate and programme, the digital and non-digital aspects, dubsuming the planning for contingencies and sundry eventualities, in order to realise the project, as well as keeping the localised project and initiative, audited and up-to-date, according to best practices. This is ambitious and currently, significantly beyond my propensity and skill level, but realising it will frame and inform my learning and practical endeavours.
I have been contemplating this project or its likes, and the need for it, since the emergence of the generative artificial intelligence GPT boom that began globally in 2023. This boom is both a boon and bane, as is AI technology. All my human lived experience informs me, that using AI technology, to counter the possible negative impacts of AI technology is fundamental, necessary and primary. This project is initially informed by:
This YouTube video provides a tutorial on how to store and secure the entirety of human knowledge in your own home. The speaker suggests downloading an open-source large language model (LLM) and storing it on a local drive. Here are the steps outlined in the video on how to achieve this:
-
Choose an LLM: The speaker recommends choosing the latest and greatest model that your computer can run. He suggests downloading an uncensored model, especially for emergency situations where you might need to access information that is typically restricted.
-
Download and store the LLM: You can download the LLM directly from LM Studio, which is a graphical user interface that simplifies the process. LM Studio also tells you which models are compatible with your computer. Store the downloaded model on an external drive for redundancy and portability.
-
Run the LLM: LM Studio provides a chat interface where you can interact with the downloaded model. You can ask it questions, get instructions, and even generate creative text formats.
-
Hardware and power: To run the LLM, you will need a computer, an external drive to store the model, and a power source. The speaker suggests using a battery backup or a solar panel for power in case of emergencies.
-
Maintenance: Keep your software and models up to date. Regularly check the battery health of your backup power source.
The following are some additional points mentioned in the video:
- The speaker acknowledges that storing an LLM locally is not a foolproof solution, but it can be a valuable resource in case of internet outages, government restrictions, or other unforeseen circumstances.
- They suggest storing the LLM in a fireproof safe and protecting it from electromagnetic interference using a Faraday bag.
- The video mentions alternative storage options like solid-state drives, tapes, and Blu-ray discs, each with its own advantages and disadvantages. I intuit that none of these storage solutions are omptimal for this project, and this will take considerable, informed research and inquiry to determine for probity.
Overall, this video provides a comprehensive overview of how to store and access a vast amount of information locally using an LLM. While it's important to remember that this is not a guaranteed solution for every scenario, it can be a valuable tool for preparedness and knowledge preservation.
https://www.kaggle.com/ (accessed: Saturday January 20, 2024)
https://www.kaggle.com/learn (accessed: Saturday January 20, 2024)
Imagine a vibrant arena where data scientists, machine learning experts, and coding enthusiasts square off in friendly challenges, wielding algorithms like magical spells. That's Kaggle in a nutshell! Here's the lowdown in three quick bites:
1. The Competition Colosseum:
Kaggle's heart beats with data science competitions. Companies and organizations pose real- world problems, from predicting medical outcomes to identifying fraudulent transactions. Aspiring data heroes compete to build the best predictive models, earning bragging rights, prizes, and valuable experience. Think epic hackathons, but fueled by terabytes of data!
2. The Collaborators' Guild:
Kaggle isn't just about one-upmanship. It's a vibrant community where knowledge flows freely. Members share datasets, discuss code, and mentor each other. Beginners find a welcoming space to learn, while seasoned veterans hone their skills and network with industry leaders. Think of it as a data science guild, bustling with collaboration and camaraderie.
3. The Learning Laboratory:
Kaggle is more than just competitions and community. It's a learning platform. Free courses, tutorials, and notebooks guide your data science journey. Explore cutting-edge research, experiment with new tools, and push your analytical boundaries. Think of it as a vast laboratory, brimming with knowledge and the potential for groundbreaking discoveries.
So, whether you're a data novice or a seasoned pro, Kaggle has something for you. Dive into the competitions, join the discussions, and embark on your data science adventure!
I fortuitously happenned upon Kaggle, by the grace and benevolence of Goddess Serendipity. I almost didn't click on the linkage due to being somewhat bemused by its fluffy nomenclature. I dont know what repaired my reservations in the name exactly, for I did go forth, through the linkage to the site, and was instantaneously pleased, appreciative, and perceived its value to my emergent project.
I am not confident, that this is the appropriate section to be populated with the following information, but I am going to install and park it here, for the interim.
While "visual data analysis" might not be the core element of SeedVault.L-GPT, it does have significant potential to support various project stages:
- Dataset Curation and Exploration:
- Visually analyze the topic distribution, diversity, and potential biases in your bespoke dataset curated from Anna's Archive.
- Identify gaps and redundancies to guide targeted dataset expansion or refinement.
- Explore connections and relationships between different knowledge domains within the dataset, informing LLM architecture and training strategies.
- LLM Training and Monitoring:
- Track and visualize loss functions, accuracy metrics, and model drift during LLM training to identify potential issues and optimize hyperparameters.
- Implement attention visualization techniques to understand how the LLM processes and generates text, revealing potential biases or blind spots.
- Compare and contrast different LLM training runs visually to select the most promising models for further development.
- Auditing and Benchmarking:
- Visually present the results of auditing your LLMs for biases and factual errors, enhancing transparency and trust.
- Compare the performance of different LLM versions on benchmark tasks, ensuring continued relevance and appropriateness.
- Develop interactive dashboards for the centralized and decentralized committees to easily monitor and assess the LLMs' performance and evolution.
Additional Points:
- Automating visual analysis through scripts and tools can save time and resources while ensuring consistency.
- Involving the community in designing and interpreting visualizations can promote inclusivity and trust.
- Remember, visual analysis is a tool, not a solution. Combine it with other methods for comprehensive project evaluation.
Conclusion:
While not the central focus, visual data analysis can be a valuable asset in SeedVault.L-GPT by providing insights into your data, optimizing LLMs, and ensuring their continued relevance and reliability. Embrace its potential and utilize it strategically to guide your ambitious and innovative project towards success.
In the metaphorical structure of this project and programme's outline, I feel and intuit, that this section is apt for enshrining the developmental community aspects and skill requirements to realise a working prototype of this initiative. It is my experience, that projects and initiatives engender their own lifeforce, humunculi of sorts, that bring about the mechanisms and processes for their own realisation, and completion. I appreciate that this is rather whimsical and fancifil, but this is my experience and it is informed and contextualised by my experience and worldview. To this end, I envision that this project will draw to it, people with the necessary skill-sets, to bring a prototype to fruition.
...
...
Through engineering and intelligent design, this project aims to have an in-built, automated and auditable process, to ensure the workable prototype remains optimal, optimised and up-to-date
Planning, designing and engineering a robust and adaptive, learning and developmental system and development community
...
...
In a nutshell, the preamble and executive summary to the project 'Code of Conduct' is to nurture, safeguard and stow the seed, all seeds, be they literal, figurative or metaphorical, for Winter is acoming...
...
...
Shell scripting
...
Python programming
...
Project Management
...
Engineering and Design of Intelligent, Learning, Adaptive and Resilient Systems
Auditing
...
Machine Learning, Deep Learning, LLM, AI, generative AI, GPTs, etc.
...
Sources that are yet to be vetted, curated, sorted and annotated
(1) Free Survival PDFs, Manuals, and Downloads, 'Prepper Library 2024...' https://www.trueprepper.com/survival-pdfs-downloads/.
(2) Digital prepping, a huge collection of knowledge for off grid use. https://www.reddit.com/r/preppers/comments/o44dmc/digital_prepping_a_huge_collection_of_knowledge/.
(3) 8 ways to get started with digital preparedness (Digital Prepping ....) https://www.youtube.com/watch?v=xnUtbfqHu_M.
(4) Emergency preparedness checklist: Prepping for beginners. https://theprepared.com/prepping-basics/guides/emergency-preparedness-checklist-prepping-beginners/.
(5) A.P.S.L Australian Preppers Survival League – Prepping and Surviving .... https://www.apsl.org.au/.
(6) 100 Best Prepper and Survivalist Youtube Channels - Urban Survival Site. https://urbansurvivalsite.com/100-best-prepper-survivalist-youtube-channels/.
(7) http://internet-in-a-box.org/.
(1) How to Conduct a Skills Audit at Your Organization - AIHR. https://www.aihr.com/blog/skills-audit/.
(2) How to conduct a skills audit with your employees - Robert Half. https://www.roberthalf.com/gb/en/insights/management-tips/how-conduct-skills-audit-your-employees.
(3) How to Conduct a Skills Audit at Your Organization - AIHR. https://bing.com/search?q=human+skills+audit.
...
