The Evolution of GamerHash: Embracing AI for the Future of the Web3 Landscape

In this interview with Artur Pszczółkowski , Vice President of GamerHash AI , we explore the transformative path of GamerHash—a platform that emerged seven years ago, designed to capitalize on the untapped computational power of gamers’ GPUs for cryptocurrency mining. In response to the evolving tech landscape, GamerHash has pivoted towards AI, altering its mission and establishing itself as a vital member of the Web3 community. Artur elaborates on this strategic pivot, the changing dynamics of GPUs, and the anticipated role of AI in Web3.

Could you please provide a brief overview of GamerHash and its operational model? 

GamerHash was founded seven years ago and has attracted nearly 800,000 gamers to date. Initially, the goal was to create incentives for gamers to utilize their GPUs effectively. By pooling the computational power of gamers, we managed to engage in cryptocurrency mining, allowing us to maximize the utility of GPUs for tangible benefits.

Gamers would connect to our platform, mining one cryptocurrency using their GPUs while another one was mined via their CPUs. This process was dynamic, adapting to various real-time metrics we were constantly monitoring. Ultimately, gamers didn’t need to have extensive crypto knowledge as the system operated seamlessly—just plug it in and go.

Currently, we’re thrilled to be participating in the KBW event in South Korea. Our token has been trading for over four years now, with Bithumb Korea serving as our primary exchange. Approximately 80% of our trading volume originates from South Korea, making it an exciting opportunity for us to engage with our community, partners, marketing firms, and influencers. This is undoubtedly a crucial step for our journey ahead.

AI wasn't always part of the picture. What prompted the transition from mining?

The transition we undertook is quite fascinating. Last year, we realized that mining was starting to decline. With the Ethereum merge, everyone anticipated a migration of miners to other proof-of-work protocols, but surprisingly, many devices were simply turned off. This made mining notably inefficient across the board.

Most of our gamers hailed from Europe and the U.S., where energy prices surged, partly due to the ongoing war in Ukraine. Selling our product had become a challenge since it seemed less advantageous. This led us to consider what the next big opportunity would be for our network of gamers, and that's when we turned our focus to AI.

In just three weeks, we completely transformed our operations, from our websites and communications to our workflows and incentive systems—everything underwent a significant overhaul. This strategic pivot turned out to be beneficial; by January, our token value increased dramatically even while the overall market faced downturns. We view this as a sound decision.

Two years ago, AI was relatively unknown, and now it’s a prevalent topic in crypto. How did we reach this point?

Reflecting on two years ago, when OpenAI released ChatGPT, we couldn’t foresee the implications. Many were concerned about job losses and AI overtaking human roles, but we now recognize that’s not the reality. Instead, AI has often proven to be a tool that enhances our productivity rather than replaces it.

AI became the fastest technology to amass a billion users, marking an 'aha' moment for us as we realized how demanding AI is in terms of GPU resources. Unlike mining, which faced ecological criticism—after all, the Bitcoin protocol consumes more energy than some countries—AI's narrative is quite different. People are more accepting of the resource demands for AI because they perceive it as having a larger positive impact.

This shift was fortuitous for us. Our network of gamers, equipped with consumer-grade GPUs, became highly sought after. The timing was right as we recognized that our GPU power would be valuable to the AI segment in the Web3 arena. This aligned perfectly with the emerging trend of DePIN (Decentralized Physical Infrastructure Networks), which aids AI applications, placing us in a fortunate position thanks to our decentralized model.

At this moment, there’s fierce competition among leading DePINs vying for enough GPU capabilities to train various models. We’re not attempting to compete in that arena, as we lack high-end GPUs, but we see ample opportunity for ourselves.

What’s your perspective on AI in Web3? Do you find it compatible with the technology?

Presently, we’re facing a global shortage of GPUs. Major companies like NVIDIA, Tencent, and Meta are buying up every GPU they manufacture. This shortage exists while the hype around AI continues to grow, with every CEO striving to demonstrate their commitment to engaging with AI technology.

The challenge is that many companies eager to explore AI initiatives are hampered by the lack of available GPUs. IT teams scramble to find GPU options in the cloud, but platforms like Google Cloud or Amazon AWS have virtually nothing left to offer.

This means that whether you run a small, medium, or large enterprise wishing to incorporate AI—like building your own Large Language Model (LLM)—you find yourself at an impasse.

Can you elaborate on inference and its definition?

An intense battle is unfolding among the leading DePIN projects. These major players possess top-tier GPUs sourced from data centers and are competing aggressively in the market for training LLMs. If you operate a small to medium-sized business looking to train an AI model, you’ll need vast amounts of data and the most powerful GPUs available, which typically takes around two weeks.

Unfortunately, our consumer-grade GPUs prevent us from entering that market, and frankly, we’re not interested in pursuing it either. A new opportunity has arisen that didn’t exist a year prior, known as inference. To visualize this, think of model training that requires extensive GPU resources to create a model from a provided dataset, whereas inference is concerned with running that model once it’s been established.

Our retail-level GPUs are well-suited for inference tasks. As the major players wrestle for training capabilities, we remain poised to facilitate hosting once those models are in place. Utilizing sharding, we will distribute these models across our decentralized network, making inference an exciting narrative for the latter half of this year.

Which projects or players do you envision shaping the future of the AI Web3 ecosystem?

In the realm of AI within Web3, various project types exist. At the foundational level, we're seeing protocols or blockchains designed or adapted to meet AI requirements. This encompasses not just GPU provision but also hosting services and support tailored to AI needs.

We can differentiate the GPU market into two segments: supply and demand. GamerHash finds its place in the supply segment; we have GPUs contributed by our gaming community, and we actively manage this growing group. Our numbers are increasing, but we can't reach out to individual clients with offers like, 'You can host your LLM model on our platform.' That’s not feasible for us, nor is it our primary focus.

On the demand side, there are marketplaces that facilitate AI initiatives. If your company is looking to deploy an LLM model, you would likely seek assistance from one of these marketplaces. Golem, Akash, IONet, and Aethir represent examples of such entities. They have dedicated teams looking to connect with projects requiring GPU resources.

Our goal is to forge connections with these marketplaces while they concentrate on uncovering demand. We provide the supply, creating a cycle that links demand with supply and the necessary protocols. Other participants may exist in niche spaces, but they are not central to this discussion. While some smaller projects are innovating within AI, delving into every single one would be too intricate for this context.

Could you clarify how gamers contribute their GPUs on your platform and the operational process?

Our guiding principle from the very beginning has been to simplify everything for our users. The goal is always to ensure a plug-and-play experience. Our platform functions seamlessly; gamers activate it in the background without any disruption to their gaming experience or computer functionality. We emphasize ease and accessibility in all aspects.

GamerHash's Evolution: Harnessing AI to Shape the Future of the Web3 Ecosystem - Metaverse Post

Once established as a cryptocurrency mining platform utilizing GPUs, GamerHash has now embraced AI to redefine its mission and assert itself as a pivotal player within the Web3 landscape.

GamerHash's Evolution: Harnessing AI to Shape the Future of the Web3 Ecosystem