This begins with the communication with customers and suppliers and continues to taking in and processing orders, manufacturing, despatch and on-time delivery. This development can greatly benefit the facade and glass industry, by enhancing efficiency and safety through advanced AI technologies.
However, using AI often raises concerns about the security of confidential or proprietary information because it can be used for AI model training and may be exposed in some way.
See also: Interview with Val Voznyi: Trends of the key trade fairs GlassBuild and glasstec
How is this possible? Nvidia has introduced new AI hardware by the name of Project DIGITS, marking a significant step towards accessible and affordable AI capabilities for small and medium-sized companies. In addition, this solution will keep your data safe.
Generated using DALL-E
Background: How does AI work for you?
There are primarily two ways to utilize AI, technically referred to as large language models (LLMs), which are the core of AI chatbots such as ChatGPT:
Cloud-based AI models: These are cost-effective or even free, but they involve sharing data with the cloud-based AI model. While many companies already use different cloud-based services, concerns arise with AI models since they may use your data for training, potentially leading to data leaks.
This often raises concerns about the security of confidential or proprietary information because it can be used for AI model training and may be exposed in some way.
In-house AI models: With that in mind, another option is running LLMs on local hardware, which keeps all data within the company, preventing data from leaving the premises. However, the hardware required to run powerful LLMs can be quite expensive.
At the recent presentation during glasstec 2024, Thomas Häuser of A+W Software highlighted the high costs of hardware for AI models, estimated at around 50,000 dollars or more. This expense makes local AI models unaffordable for many small and medium-sized enterprises (SMEs), which constitute most glass-processing companies.
One cost-effective solution mentioned by Häuser was the OpenAI API Enterprise, a cloud-based service that ensures data is excluded from AI model training or retention by default, offering enhanced data security for companies. This is quite an important and interesting part, but I see you would like to omit it.
glazingAI.com
A safe and affordable solution
Now Nvidia offers another option that is both powerful and affordable: the new AI computer Project DIGITS. This device can run AI models with 200 billion parameters, and two of them linked together can handle up to 405 billion parameters, making it suitable for most well-known AI models, including the Meta's largest Llama model.
Also interesting: AGP takes advantage of the Glaston Autopilot on its tempering lines
In addition, this system can serve more than adequately to run various AI models if the latest news from DeepSeek is accurate.
In-house AI for just 3,000 dollars
The AI computer is compact, about the size of two smartphones, and can operate on a regular electrical network without any special equipment. This solution allows for the local use of AI models without concerns about data security. The best part? It's priced at 3,000 dollars, making it an affordable option even for smaller companies.
Over the coming weeks and months, we will be further expanding on the topic of AI, sharing our expertise on how this technology can enhance the efficiency and safety in many sectors of the glass and facade industries.
For more enquiries on how AI is transforming the glazing industry, feel free to contact Val Voznyi directly: val@glazingAI.com
Generated using DALL-E
