In June 2023, McKinsey, a world-renowned consulting firm, released the report "The Economic Potential of Generative AI: The Next Wave of Productivity", which pointed out that if the 63 types of generative AI analyzed are applied to various industries, it will bring $2.6 trillion to $4.4 trillion of annual growth to the global economy. This prediction does not take into account all generative AI applications, and the economic impact of generative AI could double if applications that have not yet been studied.
In fact, since the global explosion of generative AI, it has quickly penetrated into all walks of life, bringing considerable efficiency and effectiveness improvements. In the field of architectural design, for example, AIRI lab. Leverage generative AI technology to power OBO Design, a global design services firm. This helped more than 400 designers reduce their drawing time from three weeks to just one day, which is expected to increase the team's workforce efficiency by 25%. How did they get to this remarkable achievement?
A pair of eyes that are good at spotting opportunities
As the use cases of artificial intelligence (AI) technology in different industries continue to grow, AIRI lab. The team also sees new opportunities to apply AI to the architectural sector, with plans to launch generative AI-based services to support designers in their daily work.
AIRI lab. Co-founder Fung Dayuan said, "AIRI lab. The team's goal is to be the first to apply AI technology in the field of architecture, so that generative AI can become an assistant to designers, changing the lengthy and complex workflow of traditional concept drawings and renderings, reducing project time and simplifying work complexity. ”
For AIRI lab. For the team, it would be challenging to deploy a model that generates images from text on its own using cloud service offerings from cloud service providers. AIRI lab. At the beginning of the team, the team was a team of only 4 people, and with limited human resources, it not only needed to consider the scheduling management and operation and maintenance of basic resources, but also needed to fine-tune the model in service design to meet the needs of the business. AIRI lab. It can perfectly meet the needs of design companies in the architectural field, and can effectively improve efficiency while controlling costs by using cloud hosting, elastic resource scaling, and large model inference capabilities.
A spirit that is not afraid of challenges
After comprehensive consideration, AIRI lab. AMAZON WEB SERVICES AND VERYCLOUD PROFESSIONAL CLOUD SERVICES TECHNICAL TEAMS WERE SELECTED AS PARTNERS. In April 2023, AIRI lab. The team worked with the Amazon Web Services team to begin the migration of the application.
First, the switch from the original on-premise cloud server environment to the use of Amazon SageMaker cloud-hosted machine learning Xi service was completed. AIRI lab. The team deployed large models based on Amazon SageMaker, took full advantage of Amazon SageMaker's fully managed service, selected the appropriate GPU instance and volume for the model in a simple instance setup wizard, and relied on Amazon Auto Scaling's elastic resource scaling and combined with the asynchronous inference capabilities provided by Amazon SageMaker. The team was able to implement incoming requests for text-generated images, queue them, and output them asynchronously, striking a balance between cost and efficiency.
Secondly, on the basis of the managed machine Xi, AIRI lab. The team used Amazon Elastic Container Registry (Amazon ECR), a fully managed container registry, to provide container image management and deployment of application images and artifacts for generative AI applications, making it easier to publish containerized applications. Amazon Relational Database Service (Amazon RDS) is used to provide a managed database service for generative AI applications to meet the needs of user data management for generative AI applications, among other things. With the ease of integration between Amazon Web Services services, AIRI lab. The team outputs and stores large model results running on Amazon SageMaker's managed machine Xi service directly into the Amazon Simple Storage Service (Amazon S3) service, providing designers and customers with persistent renderings storage and fast access experience through unified storage and global access.
AIRI lab. Collaborate with Amazon Web Services and VERYCLOUD teams to achieve better architecture planning by designing business requirements, AIRI lab. The team used Amazon Web Services' global multi-region GPU instances, Amazon Auto Scaling, to develop a scale plan that aligned with business needs, enabling generative AI to provide compute power to meet business needs during business hours and automatically shut down at other times to reduce costs.
A fruitful collaboration
AIRI lab. After migrating and completing a proof-of-concept built on Amazon Web Services managed services, the team completed a generative AI renderings application that was originally planned to take two to three months to implement in just one month, and the service has provided more than 400 designers with a convenient application of Wensheng diagrams, and more than 50 designers are using generative AI every week Not only that, the system has also changed the traditional drawing method and communication process, shortening the rendering time from three weeks to one day, and is expected to improve the team's human efficiency by 25%.
AIRI lab. THE TEAM'S COLLABORATION WITH AMAZON WEB SERVICES AND VERYCLOUD HAS ENABLED ARCHITECTURAL DESIGN FIRMS TO ACHIEVE EFFICIENCY GAINS AND TIME AND COST SAVINGS. The Future, AIRI lab. The team will continue to cooperate with Amazon Web Services to jointly promote the innovation of artificial intelligence in the design industry, and plans to use Amazon Web Services' managed services to complete the basic model in the field of generative pre-trained transformer model (GPT), and realize fine-tuning and customization according to the needs of the architectural design industry, so that artificial intelligence can better serve the business and bring more creative architectural design works to the world.
About AIRI lab. & Ruihong shares VERYCLOUD
AIRI lab. is an online AI assistant developed by architects, for architects, designed to provide efficient, accurate, and professional-grade design support. AIRI lab. After more than 5,500 hours of training, using more than 3 million curated data, create 1 unique architecture model. The model generates 1K resolution from base and 4K resolution for the final output, offering unlimited design possibilities and professional-grade image quality.
AIRI lab. Not limited to specific image input, adaptable to different design stages. Whether it's a manuscript or a detailed model, the final result can be quickly brought to life, beyond the imagination of the eyes. Whether it's day or night, spring, summer, autumn and winter, there are all kinds of renderings. We are committed to providing the best quality services for architects, helping designers better complete their design work and improve work efficiency. Whether it's architectural style or sustainable design, whether it's landscape or urban planning issues, AIRI lab. can provide valuable insights and expertise and quickly visualize design ideas. AIRI lab. Become a "better" design assistant and integrate perfectly into your current architectural design workflow to support all aspects of your design work.
As a senior consulting partner of Amazon Web Services, Jiangsu Ruihong Network Technology Co., Ltd. has a professional technical team that is deeply involved in the cloud product market, and can provide enterprises with intelligent, safe and efficient cloud services. The current solutions cover games, cross-border e-commerce, pan-entertainment, metaverse, online education, financial technology, new energy, new retail, intelligent manufacturing and other fields, and are committed to leading the future of enterprises through digital thinking and practice.
Let's witness one small step for Amazon together
A big step forward for cloud computing