Platform

AI Collaborative Platform

IT’S LIKE HUGGINGFACE, ONLY MUCH MORE DISTRIBUTED WITH AN ORGANICALLY GROWING LLM/AI MODEL POWERING ITS PLATFORM CHATBOT, FOCUSING ON FOSTERING OPEN LEARNING AND SHARING BETWEEN AI EXPERTS AND ANYONE WHO ARE ENTHUSIASTIC ABOUT ARTIFICIAL INTELLIGENCE.

How Things Works?

Note :
This architectural concept is currently patent pending, under patent act 1983 and patent regulation 1986. Our registration number : 2155PCT

  • THIS ARCHITECTURAL CONCEPT CAN INTERCEPT ANY TYPE OF AI MODELS THAT RECEIVES INPUT, PROCESS AND GENERATE OUTPUTS OR ACTIONS.
  • THE ‘SLMHUB’ IS WHERE THE DISTRIBUTED COLLABORATIVE AI PLATFORM WILL BE.
  • THIS REDUCES INFERENCE COST
  • SIMULTANEOUSLY IMPROVES OUTPUT QUALITY
  • ANY AI MODEL CAN BE SYNERGISTICALLY ADDED TO THIS ARCHITECTURE

App Store

Building AI App

Developer can build AI app using Wildnerve AI Model and Architecture. And then the app will be listed available in Wildnerve Platform. 

Hub

SLM Hub

About SLM Hub and connectivity on Wildnerve Platform. 

Business Model

Wildnerve can offer potentials for young enthusiasts to upskill their knowledge on AI through potential coding classes offered by senior AI/ML Engineers, Data Scientists, Analysts, Software Developers, AI/DevOps Engineers, etc. It encourages constantly growing AI repository, resulting in keeping communities up-to-date knowledge on cutting edge technologies and last but not least it also provide opportunities for cloud hosting providers to tap into these opportunities. All these are very promising revenue streams in coming years.

What is the ‘Pulling Factor’?

  • Not only a distributed community. It’s GLOBAL A.I. EDUCATIONAL NETWORK
  • Purpose driven platform, resulting an ever cyclical ‘loop of opportunities’
  • Potentially greater AI Model and cutting-edge technologies available
  • Greater freedom in investing money, effort and resources
  • Those who contributes computational resources, models, tools can earn a share of our subscription
  • Advertisers too can target their specific ads to specific people in the platform
  • Opportunities for AI enthusiasts resulting in greater adoption

How is the current ‘Market Scene’?

AI collaborative tools and platforms are so new, it does not really have many focused industry, but there are a handful worthy players, so the following competitive analysis will illustrate.

The ‘radar chart’ above illustrates based on five dimension, (1) Open-source vs Proprietary (2) Decentralized vs Centralized (3) Free vs Paid (4) NLP focused or Any A.I. (5) Has Collaborative Hub or not. Taking a zoom at four other worthy contenders – HuggingFace, Fast.ai, AI21, SingularityNet comparing it to our Wildnerve,

Collaborative Platform

We see that HuggingFace scores high at the ‘Collaborative Platform’ dimension because they are now very popular amongst AI communities, and frequently used as means to try open-sourced AI models uploaded by corporations and institutions that supports Open-sourcing AI, the likes of Meta Inc, Mistral, etc, technically savvy developers and data scientists will also upload, use the available datasets and Retrieval Augmented Generation (RAG) tools and other fine-tuning techniques to fine tune the open-source models, etc. The other

Open-source vs Proprietary

Although Fast.ai and SingularityNet scores relatively high on ‘Open-sourcing’, these players has different way to balance their business needs. HuggingFace on the other hand scores lesser in this dimension because ultimately their premium tier for computation to allow model contributors to run their powerful LLM models is not free.

Decentralized vs Centralized

In this dimension, currently only SingularityNet scores relatively higher because of their use of their Cryptocurrency token (AGIX) to allow investors to own a share into their movement, in which is to open-source AI and their ultimate goal of achieving AGI (Artificial General Intelligence). However it is not known if they incorporate a concept of decentralization in the aspect of AI models contribution, sharing, etc.

We, through Wildnerve aspire to push as far as the concept of decentralization allows before it becomes impractical, in which we will take a step back on this because based on our technical research, to truly employ a totally ‘Decentralized’ system to power an AI collaborative will be too complex and too costly to prevent the entire ecosystem to go chaotic in haywire. Which means you get malicious actors uploading AI models trained on illegal datasets, misusing the AI models and tools with dangerous effect and other mayhems, pushing the boundary of ethics.

Although the use of a monetary based consensus protocol like Proof-of-Stake can curb such incidents, but the big question is what IF it isn’t about money? What IF it is political? What IF the malicious perpetrators are filthy rich and hellbent to cause mayhem? Proof-of-Stake doesn’t work. Proof-of-Work is similar and with worst consequences as it adds on further impact on power consumption and impinging on the environment.

So we need a merit based consensus protocol. However, such a protocol if done fully automated will be very, very complex and likely you need to train another AI model to specifically focus on verify the incoming traffic who uploads illegal datasets or AI models pre-trained on these illegal datasets.

So we proposed a distributed Hybrid system of human and AI verification and rating system to ensure it is based on Merit instead of just computation or monetary stakes. Which is why we won’t consider Wildnerve is a fully decentralized system in an absolute sense. Anyway no cryptocurrencies or any AI platforms current is truly decentralized. To a certain extent even Bitcoin.

Free or Paid

The biggest challenge amongst all of those players in this artificial intelligence industry is probably the time and effort spent to design and curate a business model that strike a perfect balance between adopting open-source concept for the betterment of society versus meeting shareholders’ expectation and business needs.

Here we can say none of them are able to even meet such balance, in fact majority are just ‘scraping on the surface’ of Openness. After all, it is not cheap to operate an AI company, all of its software, the datasets and computational resources are enormously expensive. They have to make enough money to sustain, so often you get freemiums that doesn’t really meet your requirements like running your custom built LLM and dataset training needs.

This includes our Wildnerve, rather than making it free for a certain period of time or limiting users’ access after a certain number of tries and then announce you will be charging a monthly or yearly fee, which can cause some of the users to divert their attentions to another model.

Nowadays and going forward to the future, you will get increasingly sophisticated and powerful AI models available, like before Gemini Ultra, people praised how wonderful ChatGPT was, but then comes Anthropic’s Claude 3.0 Opus, then you have another open-source DBRX and OpenAI will be coming up with GPT-5.0 soon, More to come as we progress, one will be slightly better than the previous. So all these makes it difficult to strike a balance, surely charging is a must.

Charging a fee as per use would seem to be a more reasonable idea because by charging users, we are forcing the unserious and playful users to drop off. Resulting in better users targeting. However, people generally don’t like paying for things. So we need to find a way to make the act of paying as easy as possible and if possible, ‘under their radar of attention’. In other words coercing them to pay based on a basis similar to their impulse purchasing habits. Once it becomes habitual, then it will be as easy as buying a cone of ice-cream. You won’t even notice it.

NLP or Any A.I. ?

Here all the listed companies are either NLP libraries, tools or collaborative platform or marketplace. From this dimension, mostly scores high on the category of ‘Any A.I.’, except for HuggingFace because they are more focused onto Natural Language Processing (NLP) sector.

The following table gives a more detailed scoring based on the five different dimensions we mentioned above:

Milestone & Plan

The following is our milestone and plan from ideation stage, architectural design, workflow to initial coding are illustrated as follows: