SILX AI - A frontier model at your pocekt

SILX AI – A frontier model at your pocekt

Back

Top

Overview

SILX AI is making a new frontier in AI by focusing on reasoning at the model’s core eliminating the need for CoT or complex prompting. Our early prototype, guided by our latest research paper, introduces groundbreaking mechanisms: the Token Temperature Mechanism (TTM) and Guided Sequence of Thought (GSoT). These innovations fundamentally enhance how models process tokens, leading to more structured and context-aware reasoning. SILX AI is redefining AI intelligence from the ground up.

Proposal Description

How Our Project Will Contribute To The Growth Of The Decentralized AI Platform

The creation of affordable AI models that can reason at their core, making them more scalable and efficient. By controlling the model’s own reasoning, we enhance predictability and mitigate safety risks, ensuring more reliable AI behavior. Additionally, this approach democratizes AI development, allowing small labs worldwide to build frontier models, fostering global innovation in AI research.

Our Team

Eyad Gomaa – CEO & Co-Founder

AI researcher
Ex-CEO of Triv AI, ranked among the top 42 transportation companies in Egypt
Co-author of "Guidance Is All You Need", a groundbreaking paper introducing new AI mechanisms

https://www.linkedin.com/in/eyad-gomaa-silx/

Mohamed Sharaf – CTO & Co-Founder

AI researcher and engineering expert
Ex-CTO of Triv AI, bringing extensive experience in AI-driven solutions

AI services (New or Existing)

Quasar-1

Type

New AI service

Purpose

Quasar-1 is a frontier AI model designed to achieve true reasoning at scale through its Token Temperature Mechanism (TTM) and Guided Sequence of Thought (GSoT), Quasar-1 dynamically refines token importance and structures logical progression, reducing computational overhead while improving accuracy. The goal is to create affordable yet powerful AI models that can perform complex reasoning tasks without excessive resource consumption.

AI inputs

Quasar-1 processes natural language inputs, structured data, and contextual information. It utilizes TTM to assign importance levels to different tokens and GSoT to structure its reasoning process in a more predictable and interpretable manner.

AI outputs

Quasar-1 generates highly structured, reasoning-driven responses with minimal compute overhead. Context-aware explanations that follow a logical reasoning path. frontier-level AI capabilities that remain affordable and scalable.

Company Name (if applicable)

SILX AI

The core problem we are aiming to solve

We are solving the challenge of making AI both affordable and frontier. Current AI models are either too expensive to scale or lack true reasoning capabilities. Our focus is on developing models that can reason at their core while remaining cost-efficient, making high-performance AI accessible to researchers and labs worldwide. By ensuring these models are predictable, controllable, and scalable, we push the boundaries of AI without requiring massive infrastructure. This approach democratizes AI development, allowing even small teams to build cutting-edge systems.

Our specific solution to this problem

Our solution, as demonstrated through our TTM + GSoT mechanisms, enables AI models to achieve reasoning at scale without requiring extra computation time. Unlike traditional models that rely on test-time compute to refine their outputs, our approach allows the model to reason in real time, eliminating unnecessary processing overhead. This results in lower compute usage while maintaining high performance, making AI both efficient and scalable without sacrificing reasoning capability.

Project details

Quasar-1 is a next-generation AI model designed to achieve true reasoning at scale through its novel Token Temperature Mechanism (TTM) and Guided Sequence of Thought (GSoT). Unlike traditional models that rely on brute-force computation and extensive test-time processing, Quasar-1 introduces a structured approach to token processing and reasoning.

At its core, TTM dynamically adjusts attention based on token importance, distinguishing between hot tokens (critical for reasoning) and cold tokens (less relevant). This mechanism ensures that the model focuses on key parts of the input, optimizing efficiency and reducing unnecessary computation.

GSoT further enhances the model’s reasoning ability by structuring token interactions in a way that mimics human-like logical progression. Instead of generating responses through trial and error, Quasar-1 follows a guided thought process, making it more predictable and reducing computational overhead.

By combining TTM and GSoT, Quasar-1 achieves frontier-level performance while remaining computationally efficient. This allows for real-time reasoning without excessive resource consumption, making AI more scalable, controllable, and accessible to researchers and organizations worldwide.

Existing resources

we have 3k worth of compute

Open Source Licensing

Apache License

Links and references

our paper :
https://arxiv.org/abs/2412.06822

Join the Discussion (3)

Sort by

3 Comments

0

Simon250
Mar 9, 2025 | 1:23 PM

Edit Comment

Under project details there are some weird symbols as shown below. 4o

Under project details there are some weird symbols as shown below. 4o
- 0
  
  Simon250
  Mar 9, 2025 | 1:24 PM
  
  Edit Comment
  
  I can't post the screeshots here. It's above Existing Resources, in the Project Details section.
  
  I can't post the screeshots here. It's above Existing Resources, in the Project Details section.
0

Sky Yap
Mar 9, 2025 | 12:59 PM

Edit Comment

I think SILX AI is a really innovative project that’s pushing the boundaries of how AI models can reason. By focusing on mechanisms like TTM and GSoT, it’s addressing a core challenge—enabling models to process tokens in a structured, context-aware way without relying on heavy prompting. This could make AI not only more efficient and scalable but also safer and more predictable. One thing to keep an eye on might be scaling compute resources as the model evolves, but overall, it’s a bold and exciting step forward in AI research.

I think SILX AI is a really innovative project that’s pushing the boundaries of how AI models can reason. By focusing on mechanisms like TTM and GSoT, it’s addressing a core challenge—enabling models to process tokens in a structured, context-aware way without relying on heavy prompting. This could make AI not only more efficient and scalable but also safer and more predictable. One thing to keep an eye on might be scaling compute resources as the model evolves, but overall, it’s a bold and exciting step forward in AI research.

Expert Ratings

Reviews & Ratings

No Reviews Avaliable

Check back later by refreshing the page.

The weighted average of the 4 perspectives Overall

0.0
To what extent will the project help the AI platform grow?Usefulness

0.0
Confidence level that the project is possible at allFeasibility

0.0
Market fit - Balancing needs and benefits against competitionDesirabilty

0.0

Review Headline

0 /50 chars

Review Summary

0 /5000 chars

Warning: Adding final group rating for this project will prevent expert users from adding new or editing existing reviews

Reviews and Ratings in Deep Funding are structured in 4 categories. This will ensure that the reviewer takes all these perspectives into account in their assessment and it will make it easier to compare different projects on their strengths and weaknesses.

Overall (Primary)
This is an average of the 4 perspectives. At the start of this new process, we are assigning an equal weight to all categories, but over time we might change this and make some categories more important than others in the overall score. (This may even be done retroactively).

Feasibility (secondary)

This represents the user's assessment of whether the proposed project is theoretically possible and if it is deemed feasible. E.g. A proposal for nuclear fission might be theoretically possible, but it doesn’t look very feasible in the context of Deep Funding.

Viability (secondary)

This category is somewhat similar to Feasibility, but it interprets the feasibility against factors such as the size and experience of the team, the budget requested, and the estimated timelines. We could frame this as: “What is your level of confidence that this team will be able to complete this project and its milestones in a reasonable time, and successfully deploy it?”
Examples:

A proposal that promises the development of a personal assistant that outperforms existing solutions might be feasible, but if there is no AI expertise in the team the viability rating might be low.
A proposal that promises a new Carbon Emission Compensation scheme might be technically feasible, but the viability could be estimated low due to challenges around market penetration and widespread adoption.

Desirability (secondary)

Even if the project team succeeds in creating a product, there is the question of market fit. Is this a project that fulfills an actual need? Is there a lot of competition already? Are the USPs of the project sufficient to make a difference?

Example:

Creating a translation service from, say Spanish to English might be possible, but it's questionable if such a service would be able to get a significant share of the market

Usefulness (secondary)

This is a crucial category that aligns with the main goal of the Deep Funding program. The question to be asked here is: “To what extent will this proposal help to grow the Decentralized AI Platform?”

For proposals that develop or utilize an AI service on the platform, the question could be “How many API calls do we expect it to generate” (and how important / high-valued are these calls?).
For a marketing proposal, the question could be “How large and well-aligned is the target audience?” Another question is related to how the budget is spent. Are the funds mainly used for value creation for the platform or on other things?
Examples:

A metaverse project that spends 95% of its budget on the development of the game and only 5 % on the development of an AI service for the platform might expect a low ‘usefulness’ rating here.

A marketing proposal that creates t-shirts for a local high school, would get a lower ‘usefulness’ rating than a marketing proposal that has a viable plan for targeting highly esteemed universities in a scaleable way.
An AI service that is fully dedicated to a single product, does not take advantage of the purpose of the platform. When the same service would be offered and useful for other parties, this should increase the ‘usefulness’ rating.

This represents the user\'s assessment of whether the proposed project is theoretically possible and if it is deemed feasible. E.g. A proposal for nuclear fission might be theoretically possible, but it doesn’t look very feasible in the context of Deep Funding.

Creating a translation service from, say Spanish to English might be possible, but it\'s questionable if such a service would be able to get a significant share of the market