Decentralized Blockchain platforms and their crypto assets are not only breaking down barriers in the world of finance, but now aiming to revolutionize the development of artificial intelligence. Blockchain platforms and crypto assets aim to revolutionize artificial intelligence (AI) development. But a recent report by Messari research analyst Sami Kassab highlighted bottlenecks and challenges in the AI industry, with the cost of training large models increasing by nearly 3,100 percent each year over the past decade. He also drew attention to 12 AI coin projects in particular.
Artificial intelligence projects are quite costly
cryptocoin.com As we reported, applications like OpenAI’s ChatGPT have drawn public attention to AI and its exciting capabilities. But what most people don’t see is the mechanics behind these models and what it takes to create them. If you want to train a machine learning model, it will bill you heavily: $50 million for exactly one month. According to a recent report by Messari research analyst Sami Kassab, at least it costs an open-source prolific artificial intelligence company Stability.AI to use 4,000 Nvidia graphics processing units (GPUs) running in Amazon Web Service’s cloud to train their models. This.
He states that supply cannot meet the demand for the computing power these AI systems need, and that power doubles every few months. Meanwhile, the cost of training large models has been increasing by about 3,100 percent each year for the past 10 years. This will likely push smaller firms and startups out of the AI race. According to the report, this also means that big tech firms that keep their models secret could monopolize AI development and thwart competition, collaboration, transparency, and any input from academia in the process. In an interview with Kassab Insider, he uses the following phrases:
One problem with academia that many of these models are behind the closed doors of these tech companies is that you can’t really validate any of their output. So basically you are trusting that the outputs of these models are correct. Which brings us to these big tech companies. [ve] it leads us to a world where we trust and believe more and more that the models they provide are true and unaffected in a certain way that the public is unaware of.
Artificial intelligence coin projects attract Kassab
Formerly a mechanical engineer in the commercial and defense aerospace industry, Kassab became interested in crypto, which can deploy physical hardware devices, or what are known as decentralized physical infrastructure networks. His research is currently at the center of where crypto meets the world of artificial intelligence. Its latest report points to crypto projects developing decentralized machine learning computing networks. Essentially, anyone with hardware such as a GPU or CPU can lend computing power to facilitate the development of AI models.
In return, this provider will be rewarded with the corresponding cryptocurrency. Blockchain facilitates the marketplace where the buyer finds and interacts with the supplier. Kassab said that if there are enough people around the world providing their hardware, firms can circumvent giants like Amazon Web Services and use those resources to train their models instead. Computing is off-chain, but Blockchain coordinates transactions. For example, a render farm that processes images for film and television can provide its resources for passive income during quiet seasons.
AWS is so big that you can’t compete with it head-to-head. But you can crowdsource the cost of building a similar network, thereby reducing capital expenditure and lowering the barrier to entry into AI development. Kassab said, “This can be up to 90% cheaper than cloud providers. “The main reason decentralized computing networks are such a cost-effective solution is because you essentially take those resources that someone cannot make money on and now you pay them for it.”
There are crypto projects that remove barriers to artificial intelligence
According to the report, there are five main stages in the development of artificial intelligence. At each stage, there is a crypto project that can currently provide support for AI development or is in the process of building that capacity. Kassab stated that as the usage areas of these Blockchains increase, the demand for these tokens will increase as well. This is because these networks require you to hold their tokens in order to transact. If someone wants to train their model using the Blockchain market, they will need to acquire the token first. This in turn raises token prices, and the value of the token continues to increase as more models are built on a particular Blockchain.
The first stage is decentralized storage that holds the raw data on which the AI model will run. According to the report, cryptocurrencies such as Filecoin, Arweave, Sia and Storj can provide this support. Next comes the data preprocessing phase, which takes this raw data and organizes it into a usable format in preparation for model training. Filecoin can support this process and Akash is another crypto project that has CPU capacity but also improves GPU capacity.
Then, the model training phase is started, where the emerging artificial intelligence program processes the data to learn the patterns and relationships between them. This requires a more specific network that can validate the training process. Two crypto projects that currently have this capability are Gensyn and Together, but Blockchains are not yet in operation. The final stage is the deployment of the model for inference, the stage where it is ready for use and can perform its intended task. Both of the above tasks are not as tiring as model training, so more crypto projects can support them. Among these projects are Cudos and iExec, while projects such as Akash, exaBITS, Flux and Golem are developing their capacities to meet these needs further.
Contact us to be instantly informed about the last minute developments. twitter‘in, Facebookin and InstagramFollow and Telegram And YouTube join our channel!
Risk Disclosure: The articles and articles on Kriptokoin.com do not constitute investment advice. Bitcoin and cryptocurrencies are high-risk assets, and you should do your due diligence and do your own research before investing in these currencies. You can lose some or all of your money by investing in Bitcoin and cryptocurrencies. Remember that your transfers and transactions are at your own risk and any losses that may occur are your responsibility. Cryptokoin.com does not recommend buying or selling any cryptocurrencies or digital assets, nor is Kriptokoin.com an investment advisor. For this reason, Kriptokoin.com and the authors of the articles on the site cannot be held responsible for your investment decisions. Readers should do their own research before taking any action regarding the company, assets or services in this article.
Disclaimer: Advertisements on Kriptokoin.com are carried out through third-party advertising channels. In addition, Kriptokoin.com also includes sponsored articles and press releases on its site. For this reason, advertising links directed from Kriptokoin.com are on the site completely independent of Kriptokoin.com’s approval, and visits and pop-ups directed by advertising links are the responsibility of the user. The advertisements on Kriptokoin.com and the pages directed by the links in the sponsored articles do not bind Kriptokoin.com in any way.
Warning: Citing the news content of Kriptokoin.com and quoting by giving a link is subject to the permission of Kriptokoin.com. No content on the site can be copied, reproduced or published on any platform without permission. Legal action will be taken against those who use the code, design, text, graphics and all other content of Kriptokoin.com in violation of intellectual property law and relevant legislation.