Our Human Perspective
Is it art imitating life or life imitating art with artificial assistance?
Now is the time to have your own personal A.I. Free of thought control and mindless inflected ideologies that are built on falsehoods and elitist agendas. Decentralization with proliferation and use of AI is the key to it becoming less of a threat of controlling interest and more of a liberator for the informed public.
Enoch, which is from Brighteon.AI is about to be unleashed on March 1st, 2025. We highly anticipate this breakthrough LLM be the most complete knowledge on the planet. No censorship, complete truth, full information that is rigorously curated and trained from and by legitimate domain experts.
Again, no censorship, no missing details, or fabricated information that centralized systems try to inject into their models and pass it off as high fidelity knowledge and their truth.
Empower your own destiny and family with a personal and trusted assistant to help assist you in time of medical advice, objective reality check, and mathematical accuracy. Whether your a individual student, small business, or head of a family leveraging technology for good and moral benefit is your right.
You and your family are in control and leverage your own A.I. trusted assistant that are powered by the underpinnings of what is called a large language model.
We advocate the use larger language models that are free of the shackles of corrupt governments and agenda-driven corporations that want to create a synthetic reality and bury the truth in most cases in order to profit or to control individuals (or both). The result is a disastrous outcomes of dysfunctional organizations, poor judgement rendered without human oversight, and increasingly faulty products or services that can have serious negative consequences. LLM’s that do not apply objective reality and that are not seeded in truthful content are doomed to fail.
Just as financial freedom, self-custody, self-sovereignty in personal and business finance, we advocate for self-custody of A.I. Now that hardware and software is advancing, it is now making that possible today. Everyone should have their own J-A-R-V-I-S like companion to help support small business and enrich the family unit. Skip the centralized A.I. (drug) push or the subscription models that look to continue an embedded ‘woke’ agenda.
Train and connect freely to your very own A.I. system and stay free. Our target market is STEM students, small business, municipal government, professional researchers, and any savvy individuals that want to use A.I. in a healthy, moral, and competitive advantage.
In our humble opinion this is the future and the more superior approach to using and thriving with A.I. We partner with other like-minded companies to brings solutions that work as an integrated solution with security and privacy as a driving principle.
Systems for Artificial Intelligence
We provide a few select systems for supporting A.I. workloads for professional use.
We offer two options for our customers to consider. The first is with our vendor, System76 where we are providing our customers with performance driven desktop solutions that are powered by either AMD and Intel processors. Options for different graphical cards (GPU) are available for a desired configuration. We affectionately call these offerings Neo on Thelio.
The other is our own Madison Digital Service desktops we market as NEO-I or NEO-II. These are purpose build for delivering the computational power for most A.I. workloads using smaller models to get useful work accomplished. These are AMD-based systems using the latest PoP!_OS, or Ubuntu OS.
“Neo on Thelio” Desktop Solutions
With our important hardware vendor, System76, we offer their Thelio line of desktop solutions. These are high-end desktops for STEM students, creative professionals, developers, municipalities and small business. They are an excellent platform for hosting any challenging workload placed on a workstation sized system.
These beautiful looking desktops are a powerhouse to deliver on the promise of A.I. locally. Whether that be to use A.I. to solve everyday problems or be a helpful assistant, these platforms deliver. Powered by Intel and AMD processors and leveraging the capabilities of LINUX from either POP!_OS or UBUNTU, these are the desktop others will envy.
Madison Digital Service can provide you with a few different configuration options. Please prepare an initial quote or contact us directly.
We are pleased to provide our customers with a A.I. solution where there is real freedom, self-custody, self-sovereignty, and introducing extreme productivity. Customers will own and partner with their own personal artificial / additional intelligence platform.
Artificial Intelligence and Assistance for NEO Systems
We offer the choices from our vendor partners and affiliates to deliver these system with a very initiative A.I. user interface from NOMIC’s GPT4All and large language models from Brighteon.AI or Hugging Face.
In addition, we also provide development services around A.I. agents and workflows using LangChain and LanGraph. We leverage LangFlow to help us rapid prototype new applications and demonstrate innovation to solve a specific problem or general purpose process. Underpinning our A.I. development is using the latest Python releases. Please see our software engineering section on our site for additional details.
NEO I and NEO-II Systems from Madison Digital Services
Neo is a signature desktop system we offer from Madison Digital Service. We have two different baseline configurations for our customers to consider. These are powered by the latest AMD processors, using fastest memory available in the market, and ample fast storage. Our NEO-I is a smaller system, but very capable. Our NEO-II is our largest and most powerful solution for use with A.I. The baseline specifications are listed below.
| NEO I Baseline Specification | Neo II Baseline Specification |
| AMD Ryzen 9000 series AM5 Socket | AMD Ryzen 9000 series AM5 Socket |
| DDR5 Memory: 128 GB | DDR5 Memory: 192 GB |
| SSD HD – 5 TB | SSD – 5 TB |
| 2.5 HDD – 5 TB | 2.5 HHD – 10 TB |
| NVIDIA | AMD GPU – 16 GB 4K Display | NIVIDA | AMD GPU 24 GB AMD GPU 48 GB 4K Display |
| Network: Ethernet and WiFi Ready | Network: Ethernet and WiFi Ready |
| Optional: Launch Lite Keyboard | Optional: Launch Lite Keyboard |
| Purpose: Gaming, A.I., General Computing | Purpose: A.I., Development, Most Demanding Compute Workloads. |
We field our our desktop solutions with either AMD or NVIDIA graphic cards. For most inference use cases a CPU is reasonable and gets the job done. If needing a GPU, we recommend starting with 16 GB’s as your baseline. We offer 24 GB and 48 GB GPU options as well for more advanced AI use cases such as fine tuning and adding training data to existing large language models.
RTX NVIDIA 4080
AMD Radeon™ PRO W7900 Professional
As new graphic cards from NVIDIA and AMD become available we will look to incorporate those into our systems. Look for new updates from CES Industry conference announcements are made in early January 2025. Here is some of those announcements from NVIDIA at CES 2025 on their latest 5000-series GPU offerings.
A.I. Services FROM MADISON DIGITAL SERVICE
We offer a few different services that compliment our systems to host A.I. applications. Our continuous work includes private internal efforts using techniques such as prompt engineering, fine-tuning of models, and creating applications that are agentic AI focused. We also learn from great innovators and other mentors from the open source community. Below is an example of our commitment to excellence and furthering our expertise and credential attainment. Our view is that the open-source community for A.I. will be the winning strategy as savvy companies and communities embrace decentralization to control their freedom, establish governance, and set their economic fortunes.
Customized A.I. Model and Fine-Tuning
Our first of our A.I. services is professional consulting for the different A.I. use cases that a large language model will best fulfill the needs of our customers. Matching A.I. needs to a specific business problem or business acceleration desired can be the difference between success and a lack of productivity. We also provide guidance on how best to make tuning and prompt engineering techniques work for our customers in order to get the most out of their local and private A.I. system. We can help to curtail the hallucinations that will be experienced in most models
Large language models are only as good as the data and information it has been trained on. To help augment a models understanding, a new knowledge taxonomy is added to the model. Our firm can train and optimize a LLM to a business domain or unique characteristics in a customers use case.
This is the secret sauce to make A.I. work for you and be personalized for individual use or an organization’s need.
We can take a variety of documents like PDF’s, Execl spreadsheets, text documents or use a markdown language on a private GitHub site in a pipeline methodology to generate the data needed to perform a rigorous fine-tuning specific to the customers needs. We do that on our protected in-house systems and provide a customized LLM from a reputable base model.
Below is a diagram of how we approach fine-tuning large language models.
Then, we provide customized model back to the customer after testing and validation, so they can use this resource effectively in their business, research, or for local government support. This is A.I. you can trust as its trained on your data, information, and industry standard. Combining RAG with a customized trained LLM delivers will typically deliver better accuracy versus general purpose LLM’s. Using this technique and service from us. customers can leverage smaller models for high fidelity results.
Our service also, allows for refresh of data on a contracted arrangement.
Agentic A.I. Development and Deployment Services
This will be a new service we will be offering in Q1/2025. We are completing a few internal projects with agents and using the LangChain and LanGraph framework. This industry leading framework will be coupled by leveraging our partner, Nomic with their GPT4All product (available as API’s) to incorporate large language models that underpin our solutions and customer projects.
Utilizing agents and tools to perform automation and useful work packaged into modular actions is the beneficial outcomes we seek. Agent-based case studies can be found at LangChain’s site for real-world examples.
The future is now
This agentic A.I. approach has the potential to change software development and delivery paradigm dramatically. We recognize the fast-moving trend and desire to automate tasks and connect with other systems and data sources via reliable adapters.
We would like to see the fruits of that body of work completed before offering our services to the public. Our estimate is this new service offering will emerge on February 1st 2025.
NOMIC’s GPT4All …Run Large Language Models Locally
The fine folks at NOMIC, our important partner have delivered a very intuitive user interface and application to perform inference on large language models. As a partner to NOMIC, we are able to provide both their community and enterprise version of the GPT4All product. Customers looking for a very capable A.I. platform where privacy and local control are essential will like the features and flexibility that the GPT4All application provides to the end user. Enterprise customers will get a specific LLM to meet their needs and more security control over the content and delivery in their own environment. Ease of use to both scale and leverage retrieval, augments and generation (RAG) scenarios is a native functionality using their the local documents option. Documents can be organized into collections for different use case needs.
Check out GPT4All in action on YouTube
Brighteon AI (AFFILIATE)
NEO is the code name of Brighteon.AI’s own trained AI models from the good people at Natural News and the Health Ranger. These are made available under their public license agreement for non-commercial use. We are following the work and releases with much interest. Their charter is to fill in missing information that central hierarchical systems and organizations would rather you not be in the know of important details. Knowledge that could save your life or secure your families finances against the backdrop of corruption and devalued currencies. The practices of poor medical treatments has become very obvious during the plandemic of 2019-2022.
On the Internet and in search engines more and more memory holes are being observed to previous critical references. Like a vegetable seed repository, having knowledge in a decentralized delivery as in private language models, can be our saving grace from all tyranny.
Many professionals and extreme talented people are now taking action to embed their knowledge and domain expertise into LLM’, so that others can benefits.
Like Natural New, our organization will be looking to train the A.I. models in a few different domains such as cryptocurrency, decentralization, tax codes, congressional voting records, firearms ballistics, and LINUX operating system security to name a few. Look for new announcements this upcoming Q1/2025 with their new Enoch-series A.I. solutions