For the last several months, we have been working on a huge product update of Assisterr. This update signifies a critical step toward creating a truly decentralised Network of Small Language Models. After thousands of hours of research on the accuracy of existing large and small models the Assisterr team is ready to debut the first version of our AI Agent platform on Solana. With the beta launch, all of our contributors can now run community-owned AI agents themselves. So we invite you to join Assisterr and dive deeper into the future of AI.
Create your own models: train them, improve the accuracy of the agents , and distribute them through your favourite social media networks With Assisterr, now anyone can launch their own SLM on their own or in cooperation with other contributors at build.assisterr.ai. With Assisterr, model creation isn’t just for developers; it’s for anyone with a vision to make a difference in their personal, professional, or creative life. Whether you're aiming to automate simple tasks, power up existing workflows, or even start a full-scale project, the possibilities are vast and incredibly rewarding.
Modern Large Language Models (LLMs) have two major disadvantages: centralisation and generality. Almost all large language models are based on centralised organisations, which on the one hand are responsible for maintaining and developing models that require huge resources, but at the same time, have full control and censorship rights over them. That means big brother is watching you. On the other hand, big models are characterised by the fact that they need literally all the information they can get to learn: “estimate the stock of human-generated public text at around 300 trillion tokens. LLMs will fully utilise this stock between 2026 and 2032, or even earlier if intensely overtrained.”
Assisterr is building a cutting-edge competitive product to Big Tech AI where ownership and management of user knowledge are in the hands of the community.
Let’s take a look into the key features Assisterr offers users:
- Incentive-driven
Each SLM has its own treasury, under the control of the creators. The treasuries now act as a measure of a model's success - the more users use the agent, the larger its treasury. In the future, communities managing agents will be able to receive royalties from their models or make them free for everyone based on their intentions.
- Community-owned
Anyone can launch and develop their own agents now. We are working to create more ways for collaboration in SLM management in the future, to maximise opportunities for both people and businesses. In the current iteration, users can create 3 agents and crowdsource quality and accuracy of answers with the help of Assisterr’s Community Contribution Dashboard.
- Choose your Role
Of course, not everyone wants to launch their own SLM But Assisterr rewards other types of contributors as well. Assisterr focuses on three main roles for its users:
-
Model Creators - those responsible for the model itself and its development.
-
Validators - experts who can use their knowledge to improve the quality of the model
-
Contributors- those who want to participate in the development of the model network and its promotion.
Each role will have its own reward. So start building an AI model with Assisterr to discover unmatched flexibility and scope. These agents can range from tools that streamline everyday routines to powerful solutions for complex business problems. Here’s a glimpse of what you can achieve:
-
Daily Productivity Boosters: Imagine a personal assistant that handles repetitive tasks for you, like managing your calendar, organizing notes, or even suggesting efficient ways to plan your day.
-
Enhanced Business Operations: With Assisterr’s SLMs, you can build agents that seamlessly integrate with your business operations. These agents can monitor and analyze data, automate responses, or handle customer queries, saving valuable time and resources.
-
Standalone Projects with Market Potential: Your AI model doesn’t have to remain an internal tool. With the right setup, it could become a fully-fledged product that meets market demands, whether in customer service, education, content creation, or more.
Assisterr empowers creators with comprehensive tools and support, making it easy to take ideas from concept to execution, and ultimately, to scale. By leveraging decentralised ownership, our product empowers individuals to own and manage AI agents in a way that was previously only possible for large organisations. This shift not only democratises AI but also provides new opportunities for interaction, monetization, and innovation.
We are advocates of free and decentralised solutions, so Assisterr as a solana-based product is an open, competitive environment for the community and their SLM agents. The models created by the community will compete with each other, leading in turn to the emergence of leaders in their knowledge niches. As the quality of the agents on Assisterr progresses, we will leverage Mixture of Experts (MoE) machine learning to create the Mixture of Agents from the most successful agents.
Mixture of Agents is the next step in the evolution of AI.
On Assisterr, it will consist of several layers. On the first layer, the keeper model will process any user request, evaluate it, and send it to a group of models that are experts in the user request. After that, the models will discuss the most correct answer or work for the user . After that, the results of the models will be sent to the user, guaranteeing the highest efficiency and accuracy of the result for each request.
This leads to the next question- how do we determine the success of the model? We accomplish this with the help of Assisterr token - ASRR. It is the lifeblood of our tech solution. At present, users can get a synthetic token to test and use for creating, improving, experimenting and having fun with SLMs in our secure environment. Soon, synthetic tokens will be converted into regular tokens and full competition will begin.
Each agent will have its own treasury, which will be spent to maintain the agent and replenished when the model is called on by users. The creators of the SLM will have restrictions on the maximum and minimum size of the treasury at the start of the model, so there will be no possibility for manipulation. The most successful models will be the leaders in the size of their treasury, so will be the most in demand in solving problems, while less popular models may eventually be stopped when their treasury runs out. Fair play.This is just the beginning of an extraordinary journey, revolutionising what is possible when the power of crypto is combined with the promise of decentralised AI agents. With the principles of decentralisation and community ownership at our core, the Assisterr team invites you to join us today and be a part of the next step in the evolution of AI.
Check Assisterr’s social media to stay tuned: