Gensyn
Can Crypto Solve One of AI’s Largest Problems?
TLDR: Gensyn leverages blockchain technology and cryptocurrency to harness a global network of decentralized computing power in order to train machine learning models, which has become prohibitively expensive.
Slowly, then all at once. Artificial intelligence (AI), often thought of by the general public as something of the not-too-distant future, has increasingly permeated modern life. While society’s dystopian view of robots taking over the world isn’t here, not yet at least, AI is revolutionizing everyday processes in many industries. Machine learning technology has long existed and undergone development for decades, but global interest recently skyrocketed with the introduction of large language model chatbot-based applications such as OpenAi’s ChatGPT. These models take large data sets, learn about their patterns, and are able to make predictions and inferences based on those patterns. Users query these applications with questions or tasks about nearly anything, be it translating text, learning about a new subject matter, writing emails, performing investment analysis, and in a matter of seconds, the application returns a response. Society has already postulated this technology has the ability to take over thousands of industries either by further advancing them or rendering them null. Regardless of how one feels about it, there’s no questioning AI is here to stay and will only continue to find more use cases permeating daily life as the tech advances.
As machine learning sees continued adoption and usage, the industry itself faces quite a large bottleneck due to scarcity of its most important resource, computing power. Computing power is the ability for machines to perform calculations. The greater the complexity of calculation, the more computing power needed. Machine learning models that train vast data sets with billions if not trillions of parameters, require immense computing power. This computing power is provided by Graphics Processing Units or GPUs.
The problem is these GPUs are:
wildly expensive to purchase and
scarce due to factors including supply chain and geopolitical issues. Typically, developers train their models using NVIDIA’s latest GPUs, which can cost up to $10-12,000 per unit. OpenAI’s GPT-3 needed 1,000 GPUs, and Stability.AI needed 4,000 GPUs costing these companies millions just to train their models on its data set.
The cost of training AI models has increased by an average of 3100% per year over the past decade, and the computation required for training advanced machine learning systems has doubled every six months. Realizing the potential revenue opportunity in front of them, centralized cloud tech giants such as Amazon, Meta, Alphabet, and Microsoft have been stockpiling GPUs. This is not only to amass computing power for their own usage but to also rent computing power out to customers willing to pay the price. Developers have a Sophie’s choice when it comes to getting the computational power needed to train the models. Invest in your own hardware which can sacrifice scalability, or use one of the cloud providers and pay inflated pricing. Due to the sheer expense of computing power needed, development has largely been left in the hands of tech giants who can afford it.
Where Does Crypto Fit in All of This?
One of the powerful features of crypto protocols is the ability to create incentivization layers for a global network of participants to contribute towards a common goal. Participants supply some sort of good or service, are rewarded via a token, and are economically interested in the success of the project. In crypto, this concept typically goes by the name Decentralized Physical Infrastructure or Proof of Physical Work. Crypto rewards can incentivize crowdsourced work across the globe to build typically expensive infrastructure with little to no cost.
There is a glaring need for a third option to access the computing power necessary to train machine learning models. This is where networks like Gensyn come in. Gensyn is a blockchain protocol that seeks to tap into the world’s idle computing power by rewarding those who pledge computing resources to create a decentralized compute network open for all. Essentially, they are aiming to build a permissionless global supercomputer tasked with the sole purpose of training machine learning models.
With access to a global computing network specifically built for machine learning, a marketplace emerges for selling computing power significantly cheaper than current cloud providers. This is because there are no firm overhead expenses surrounding the purchase of computing hardware or paying employees; rather the computing power is globally donated in return for a reward token. Furthermore, there is potentially infinite scale as to how much computing power can be accessed dependent upon how many people want to provide their idle compute to the network. Gensyn looks to not only utilize idle compute from data center GPUs but also use gaming GPUs, ASICs typically built for Bitcoin mining, and even our iPhones that carry chips capable of neural processing abilities.
While the protocol is not yet live, Gensyn projects the hourly cost to rent a GPU capable of machine learning model training to be around $0.40 per hour. With the potential for significantly more computing power as a whole collective network versus tech giants renting their limited supply, coupled with the fact that Gensyn does not charge a margin, compute costs can be reduced dramatically, allowing for continued advancement and open access to anyone who needs computing power necessary to train their models.
As machine learning continues to be integrated into our everyday lives and new use cases emerge, along with the increasing cost of compute and compute power needed for training further advanced models, Gensyn presents a unique solution to the current problems machine learning developers face. While this may not seem like an immediate case for a wave of adoption of cryptocurrency, we believe there will be significant demand for a purpose-built decentralized computing network as a cheaper alternative to AWS and Google. People will want to donate their idle GPU power in return for a reward as there is no cost to the user for providing idle power. Suppliers will be attracted to the idea that one could earn rewards by allowing access to your iPhone or MacBook’s computing capability whilst sleeping. As the appeal for rewards tied to computational power grows, there will be a surge in the supply of such power. This increase will unlock more possibilities for developers. Additionally, suppliers of this computing power will gradually enter the cryptocurrency sphere, which will stimulate them to investigate additional applications like Decentralized Finance (DeFi), Stablecoins, and NFTs. This expansion will eventually contribute to the enlargement of the overall ecosystem.
DISCLOSURES
Risk Considerations
The information herein represents the opinion of the author(s), an employee of the advisor, but not necessarily those of VanEck or its other employees. The securities/financial instruments/digital assets discussed in this material may not be appropriate for all investors. The appropriateness of a particular investment or strategy will depend on an investor’s individual circumstances and objectives.
This material has been prepared for informational purposes only and is not an offer to buy or sell or a solicitation of any offer to buy or sell any security/financial instrument, to participate in any trading strategy, or as any call to action. The information presented does not involve the rendering of personalized investment, financial, legal, or tax advice.
Certain statements contained herein may constitute projections, forecasts and other forward looking statements, which do not reflect actual results, are valid as of the date of this communication and subject to change without notice. The actual future performance of any assets mentioned herein is unknown. Information provided by third party sources are believed to be reliable and have not been independently verified for accuracy or completeness and cannot be guaranteed. VanEck does not guarantee the accuracy of third party data.
Investments in digital assets and Web3 companies are highly speculative and involve a high degree of risk. These risks include, but are not limited to: the technology is new and many of its uses may be untested; intense competition; slow adoption rates and the potential for product obsolescence; volatility and limited liquidity, including but not limited to, inability to liquidate a position; loss or destruction of key(s) to access accounts or the blockchain; reliance on digital wallets; reliance on unregulated markets and exchanges; reliance on the internet; cybersecurity risks; and the lack of regulation and the potential for new laws and regulation that may be difficult to predict. Moreover, the extent to which Web3 companies or digital assets utilize blockchain technology may vary, and it is possible that even widespread adoption of blockchain technology may not result in a material increase in the value of such companies or digital assets.
Digital asset prices are highly volatile, and the value of digital assets, and the companies that invest in them, can rise or fall dramatically and quickly. If their value goes down, there’s no guarantee that it will rise again. As a result, there is a significant risk of loss of your entire principal investment.
Digital assets are not generally backed or supported by any government or central bank and are not covered by FDIC or SIPC insurance. Accounts at digital asset custodians and exchanges are not protected by SPIC and are not FDIC insured. Furthermore, markets and exchanges for digital assets are not regulated with the same controls or customer protections available in traditional equity, option, futures, or foreign exchange investing.
Digital assets include, but are not limited to, cryptocurrencies, tokens, NFTs, assets stored or created using blockchain technology, and other Web3 products.
Web3 Companies include but are not limited to, companies that involve the development, innovation, and/or utilization of blockchain, digital assets, or crypto technologies.
All investing is subject to risk, including the possible loss of the money you invest. As with any investment strategy, there is no guarantee that investment objectives will be met and investors may lose money. Diversification does not ensure a profit or protect against a loss in a declining market. Past performance is no guarantee of future results.
 | A guest post by
|