Brian Jeffrey
Boosting Business Revenue with AI
A recent report by Accenture reveals that businesses leveraging advanced AI technologies, like large language models (LLMs) and generative AI, can see a revenue increase of 10 percent. This figure is 2.6 times greater compared to companies that do not utilize these technologies.
The Integration Challenge
As AI and LLMs become more prevalent, integrating data science into daily operations poses challenges for many businesses. Implementing AI models effectively in various business functions is often difficult.
Supply and Demand in AI Equipment
According to Nguyen Van Tuan, CEO of Hyratek, there is a significant gap between the demand and supply of equipment for training AI systems. As a result, companies must place orders with providers six months ahead of time.
High Costs of AI Systems
Most AI systems are trained centrally, incurring substantial costs that can impede their adoption in business operations.
Cloud Services and Their Limitations
In Vietnam, some companies are utilizing cloud services for deploying AI models, yet the expenses associated with large-scale operations and the lack of adaptability in work processes pose significant challenges.
Shift Toward Integrated Workstations
Nguyen Van Giap, CEO of Lenovo Vietnam, highlighted that businesses are increasingly leaning towards AI-integrated workstations to facilitate more frequent use of AI technologies.
Private LLMs and SMLs for Security
Many organizations are opting to store and develop their LLMs and small language models (SMLs) as a strategy to enhance data security and reduce training expenses.
Optimizing Processes and Innovation
This approach not only streamlines workflows but also enables decision-makers to act quickly and fosters innovation across various sectors.
Workstations Designed for AI Training
High-performance CPUs and GPUs in workstations are tailored to support the development of AI models, allowing for fine-tuning and training on a smaller scale at a lower cost than cloud options.
Enhanced Data Security and Efficiency
Working with data on-site provides heightened security and allows data scientists to train AI models more swiftly and efficiently, thus shortening the time to achieve optimal results.
Localizing Language Models
The variety of LLMs is expanding globally. Robert Hallock from Intel suggested that to enhance digital transformation, nations could develop their own LLMs, such as a Vietnamese version, incorporating local linguistic elements.
Positive Attitudes Towards Generative AI
A survey by Finastra indicates that Vietnam is at the forefront of interest in generative AI, with approximately 91 percent of respondents believing that it can yield positive outcomes.
Development of Custom LLMs
Numerous institutions are creating their own LLMs and SMLs using AI-integrated workstations to safeguard data security and minimize training costs.
