The huge data centers of the ugly duck of the world of technology infrastructure, are profitable but boring, although they are present today as a ball’s replacement – serve as some of the largest profit drivers for Google, Microsoft and Amazon. Flexera to survey 759 cloud decision makers and found that 30 % of companies spend up to 12 million dollars annually on cloud services, and 11 % of 60 million dollars.
The demand for large computing power witnesses tremendous growth after the emergence of effective artificial intelligence – but with the increasing demand for this around the world, the central infrastructure developed by large technology companies over the years, which is located in rural and low -density areas, means that both retail and companies ultimately will face more expensive and slow performance.
To help understand the size of the next challenge, I am sitting with Botyan Lee, the highest technology specialist in the world of computing power. She worked as WABI CEO and was an investor in Lightspeed and NGC Ventures. Lee is strongly involved in the democratic character of the technology economy, in an attempt to overcome large weapons at one time, and to convert devices in our homes into small data centers.
Let’s dive.
ADE: What is the new mania with data centers? Is it new? Is it required?
Boutian: Big Tech converts billions to computer infrastructure to ensure distribution control and access to the spine of the Internet. Its edge is central – it gives them absolute control over efficiency, which is the general story, but the prices and data ownership are the private that you do not hear. The challenge lies in this data centers site – where more and more people use more and more computing for increasingly complex tasks, as databases hundreds of miles are no longer viable. It is like a taxi service that works in a different city. Instead of building huge central data centers, we need to improve a counterpart connection to an counterpart, reduce cumin, and ensure that the availability of the account is smooth and automatic such as cloud services. This is how we deal with most digital problems today – the borders are the final limits.
ADE: The appearance of artificial intelligence has increased sharply from computing/data costs for startups. How can one keep high -speed access to millions of users without the need to convert data to large technology?
Boutian: The key is decentralized for both computing and data. Edge computing means treating these artificial intelligence tasks near the end user, reducing cumin costs and lowering costs. These accounts are scheduled to be published on the user -owned devices instead of corporate data centers. Instead of running everything according to AWS, accounts can instead occur on a global network of pre -existing infrastructure, such as consumer devices, bypassing the search form for the lease of the cloud. The developers do not need to pay exorbitant fees to run the work burdens; Instead, they can take advantage of this network of “unintended account”, these users pay fees for the work burden of their devices, and provide most of their accounting budget to a variety of other startups that you need desperate.
ADE: If the inactive laptops and personal computing devices can help create the next generation of decentralized data saving services, is Bitcoin mining again? Will the big players kidnap this initiative as well?
Boutian: Today’s decentralized computing is very different from bitcoin mining. Mining was a race to dominate retail, which led to industrial farms. On the other hand, he blesses the increase in computing distributed from daily consumer devices on the edge. It comes to enabling participation on a large scale so that Internet operations operate faster and cheaper.
ADE: The use of indiscriminate strangers to process decentralized data is a great concept, but it represents more specific challenges for the final user. How do you have a privacy and security of lead -resistant, so no person’s data is not linked, especially for sensitive things such as artificial intelligence tasks?
Boutian: We conducted the tests using standard devices of five million swamps that have opened access to us, and a association, which are approved among the best 3 giant computers in the world. Other super computers on that list can only be reached by a few scientists and researchers who are spent by the government. Keep in mind that we have operated a dynamic consensus mechanism, which means that customers can choose a dynamic if they want to process data using zero knowledge guides, data collection, or any other method, and not every user device has any idea of the processor already due to safe operating time.
What we found through these tests is that you can get performance, you can have more control over your privacy, and you can do these things in a cheaper and faster way and distribute them completely.
ADE: Decentralized networks can lead to large bottlenecks that are difficult to overcome. We have heard about this in the context of projects such as Ethereum, Solana and many other Blockchains that have faced the main suffocation points over the years. Have you faced similar bottlenecks, and how did you solve them?
Boutian: Our biggest challenge in Barts was to improve the loading budget and direct analogous. This means, what devices do the work burden? We have built a mechanism similar to the dating application in the Black Mirror “Hang the DJ” – it simulates in a second millimeter the performance and computing mission with consumer devices in competing for this task, taking into account geography, internet speed, processing power, and other devices that currently work on them. Fortunately, we are not Blockchain – we do not deal with blocks! We are dealing with computing and things that occupy the Internet.
ADE: Five years outside, if users have the backbone of the Internet account, what is the largest industry that you bet on? Artificial intelligence, metaverse, web hosting?
Boutian: Treating artificial intelligence. Computing is located locally, which means faster speeds, cheaper access, and most importantly, it can be open source, maintain privacy, and available to all. The biggest challenge here is not actually a technical challenge-there are already excellent open source models there that can be connected to a local device (if you do not mind waiting for the length of the Netflix episode to inquire) but it is a financial device. With companies like Microsoft that support Chatgpt with huge supplies of computing, users will continue to work as free passengers, extract value from them until the tap dries and prices go to sustainable levels.
Q7: With Park, you are working on the first decentralized data resource in the world. It has 4.3 million contracts now. How can you turn on such a large distributed entity and lay the foundation to reach artificial intelligence in the future?
Boutian: By automating everything. Bisor works on an independent network, where the contract connects and customizes the resources without manual intervention. We are not a market through which the machines you can require. Everything is automatically determined by what we call “The Orchestrator” – a mechanism that decomposes the devices of each device, the processing capacity, geographical positioning and simulation of how to perform it – all of this in millimeters again. The result is a truly decentralized and self -developmental mathematical network.
