A well-known mathematician from Greece, Archimedes, is recognized for his contributions to various scientific concepts, including pi, spirals, exponents for large numbers, hydrostatics, levers, and gravity. He is famously associated with the discovery of the volume of an irregular shape due to water displacement in his bathtub, leading him to exclaim “Eureka!” as he ran through the streets naked.
In the years following Archimedes’ exclamation, numerous unexpected inventions, revelations, and innovations may have warranted a similar cry of “Eureka!” Today, the advancements in AI and its impact on visual connectivity and the internet raise the question of whether they too deserve such recognition. Additionally, a novel technology platform utilizing electro-optical polymers is speculated to facilitate the rapid dissemination of AI online.
Assessing the Impact of AI
In the past, dial-up computers took an average of 10 to 20 minutes to load a single image, a process that continues today. Just a decade ago, waiting 10 to 20 hours to access a short video or TV show was common. However, with the reduction in latency times for images and videos, online traffic has surged. The recent surge in AI and raw computing power is poised to further escalate internet usage.
For instance, while it took nearly 3.5 decades for Netflix to reach one million subscribers after launching its streaming service, Threads by Meta reportedly attracted close to 100 million users in just over a week, showcasing the rapid adoption of AI-driven platforms. This exponential growth in AI is anticipated to have profound societal implications.
Over the past 60 years, the computing power of high-performance systems has traditionally doubled every three to five years. However, since approximately 2020, this progress has accelerated significantly, with computing power now doubling every three to four months (measured in petaflops). This surge is primarily driven by neural networks within AI utilizing GPUs and microprocessing units, placing strain on the underlying computational infrastructure.
The increased demand for visual data transmission capacity necessitates enhanced capabilities within data centers, where visual routing and switching occur. Consequently, operators of visual networks and the internet may face escalated financial burdens.
The Intersection of the Internet and AI
AI integration into everyday applications aims to enhance productivity and potentially intelligence levels. As these AI-powered devices become ubiquitous, it is imperative to consider the implications of widespread machine intelligence adoption. The visual system-based internet infrastructure, responsible for routing and switching data, is undergoing significant transformations to accommodate the growing demands of AI-generated data.
Present-day data facilities are undergoing unprecedented upgrades to meet the evolving requirements of increased traffic and data generated by AI. Three key areas of focus driving innovation in the online landscape are density, speed, and energy efficiency. Designers are leveraging these markets to develop advanced visual and optoelectronic components for fiber optic cables connecting data centers.
To address the need for efficient, low-power consumption, faster data transmission, and higher switch density, next-generation visual components like polymer modulators are being developed. These components, situated in front of lasers that transmit light through fiber-optic cables, represent a promising avenue for enhancing visual infrastructure in response to the demands of AI-driven technologies.
