AI is booming exponentially, changing the way we live and work in myriad ways. However, the high cost of AI hardware is stifling breakthrough innovation in multiple ways – for one thing by centralizing practical AI progress in a small number of organizations with the huge compute firepower needed to train resource-hungry AI systems like ChatGPT and other Large Language  Models.  

Each family of AI methods, as it develops, gets gradually optimized so as to require less hardware. Ultimately, though, the only way to accelerate and democratize AI progress is increased hardware innovation. The current deep learning boom was largely triggered by the large-scale rollout of GPU hardware, and the next wave of massive AI progress will likely be driven by new forms of compute hardware innovation as well. This becomes particularly interesting when one considers the next wave of AI progress may be the one leading us beyond ambitious but limited “narrow AI” systems that are constrained by their training data, onward toward more creative and context-savvy Artificial General Intelligence systems.

The mainstream thrust of today’s chip and computer architecture development is on incremental improvements to current methods. Simuli.ai of Florida, on the other hand, is taking a more radical approach, focusing on a novel “AGI board” incorporating two novel “AGI chips” interconnected with more traditional CPU and GPU processors: the Simuli hypervector chip and the Metagraph Pattern Matching Chip (MPMC) pattern-matching chip developed in collaboration with TrueAGI.

Simuli’s hypervector chip brings efficient processing of high-dimensional vectors to the hardware level, accelerating a variety of AI-related processes from neural net learning to homomorphic encryption to graph database search, and much more. It greatly reduces the need for data pre-processing, relative to current architectures, by utilizing on-chip transform algorithms. The chip’s in-memory computing architecture reduces power usage of the overall system, resulting in a smaller environmental footprint. And its unique use of hypervector mathematics allows support for new types of artificial intelligence with a much smaller number of processors than would be needed with traditional CPUs and GPUs.

Rachel St.Clair, CEO of Simuli Inc., said about the partnership: “The power of optimizing large-scale AGI models to run faster by leveraging Simuli’s hardware platform is multifold. First, these AGI frameworks get rapid development that wasn’t exactly possible without large compute cost prior. Then, such a device as the AGI motherboard can expand the types of code that can be run scalably and efficiently in a single instance of the AGI model, for example, Hyperon. Also, scalable computing is better for the longevity of the planet and the technology itself, so optimizing on both the SW/HW sides is key. This will likely result in AGI that’s better for everyone. We’re excited to be playing a role in tipping the scale from AI to AGI.”

TrueAGI is focused on applications of AGI and AI to support the enterprise via a SaaS platform. The MPMC, co-developed by Simuli and TrueAGI embodies in hardware the key knowledge-graph search algorithms required for neural-symbolic and neuro-evolutionary AI methods.

TrueAGI CEO Dr. Ben Goertzel says, “The overall package of the Simuli AGI board has strong potential to catalyze emergence of a new era of AI techniques and functions. The core of what we need to progress from narrow AI to AGI is of course the right cognitive architectures and learning and reasoning algorithms – but without the right hardware, even the best mathematics and software can’t run efficiently enough to have practical impact. So many AI methods we’ve been working on for decades, are going to finally be able to show their stuff in a practical sense when given the right hardware to run on.”

The Simuli AGI board is a game changer in the artificial intelligence industry. It will enable breakthrough innovation and bring us one step closer to AGI by lowering the cost and energy consumption of AI training and inference, and enabling scalable deployment of a wider variety of narrow AI and AGI architectures. Simuli, with the support of TrueAGI and other ecosystem players, is poised to change the way we think about AI.

About TrueAGI

TrueAGI is a US-based company that aims to go beyond the limits of today’s Machine Learning and Deep Learning models, enabling enterprises across verticals to leverage cutting-edge cross-paradigm AGI technology as it emerges. TrueAGI’s unique technology brings together neural, symbolic and evolutionary learning and reasoning methods to create the next AI breakthrough: Artificial General Intelligence.

About Simuli

Simuli is a company that focuses on hardware design, development, optimization and fabless manufacturing. The company charter is to exponentially reduce hardware & energy cost of advanced AI, which is increasingly becoming a barrier to breakthrough innovation. With a keen mind for building powerful and sustainable processes, Simuli was founded with the acute vision of building hardware that is robust enough for AGI frameworks while being cheaper and faster for current AI.