In 2023, with the global popularity of ChatGPT, generative AI has also attracted widespread attention from people from all walks of life. All of a sudden, from international technology giants to domestic artificial intelligence companies, almost all the technology companies we are familiar with have entered the generative AI market one after another.
As the pioneer and leader of global cloud computing technology, Amazon Cloud Technology is no exception. So what does Amazon Cloud Technology think of the rocket-like generative AI? What preparations have you made for this? What advice do you have for businesses and individuals?
Technological innovations on generative AI
"We are at an exciting technological inflection point. I believe that generative AI will bring many innovations, reshape a large number of customer experiences and applications, and bring new opportunities." Amazon Cloud Technology Greater China Product Manager Chen Xiaojian, the manager, told Ququ Technology, “Generative AI is actually a type of artificial intelligence. At three levels of artificial intelligence, machine learning, and stack, Amazon Cloud Technology has a very large product portfolio and has been using artificial intelligence. and machine learning technology to empower its own business and customers from all walks of life."
Chen Xiaojian pointed out that regarding generative AI, Amazon Cloud Technology currently has four technological innovations:
1. Amazon Cloud Technology has launched a new fully managed service, Amazon Bedrock, which can be used to build and expand generative AI applications, allowing customers to access AI start-ups such as AI21Labs, Anthropic and Stability AI through APIs. Pre-trained basic models , also provides exclusive access to Amazon Titan FMs, the base model series developed by Amazon Cloud Technology. It is also the easiest way for customers to build and extend generative AI applications using the base model.
2. Generative AI has extremely high requirements for reasoning workloads. For this reason, Amazon Cloud Technology announced that the Amazon EC2 Inf2 instance based on the Amazon Inferentia2 chip is fully available to create the most cost-effective generative AI cloud infrastructure. Running generative AI workloads at scale provides the best performance, most energy efficiency, and lowest cost.
3. Amazon Cloud Technology announced the official availability of a new Trn1n instance based on the Amazon Trainium chip, which doubles the network bandwidth and provides higher performance for training generative AI models.
4. Amazon CodeWhisperer, Amazon Cloud Technology's AI-based programming assistant, is officially available. This gives individual developers free access to Amazon CodeWhisperer without any usage restrictions.
Benefits of Amazon Bedrock
Seeing this, I believe that many readers and friends have already developed an interest in Amazon Bedrock. Chen Xiaojian said that Amazon Bedrock will bring customers five benefits:
1. Fully managed experience: Bedrock provides a serverless hosting experience through the API, and customers do not need to worry about infrastructure details such as instance types, network topology, and endpoints.
2. Easy Customization: Customers can easily customize Bedrock's base model for their specific use case by providing a small number (as low as 20) of labeled examples in Amazon S3.
3. Constructed with privacy in mind: the customer’s proprietary data used for model fine-tuning will not be used to train the basic model, ensuring that the customer’s work results will not benefit other customers (including competitors).
4. Amazon Titan base model: Bedrock has exclusive access to the most advanced base model developed by Amazon Cloud Technology.
5. Customer choice: Customers can choose between a number of different base models from Amazon Cloud Technology and third-party partners, all of which can be further customized to support their specific use cases.
On the Value Proposition of Generative AI
Chen Xiaojian revealed that regarding generative AI, Amazon Cloud Technology has given five value propositions:
1. Amazon Cloud Technology is committed to promoting the inclusiveness of generative AI technology: by launching four new innovations in the field of machine learning to support Amazon Cloud Technology's generative AI applications, customers can choose more flexibly the way to build generative AI, Companies of any size can benefit from it.
2. The easiest way for Amazon cloud technology to build and expand enterprise-level generative AI applications: Launch Amazon Bedrock preview, which provides easy-to-use API access to pre-trained basic models (such as FM from Anthropic, AI21 and Stability AI), And the Amazon Titan large language base model.
3. Amazon cloud technology provides the most cost-effective generative AI cloud infrastructure: launched Amazon EC2 Inf2 instances equipped with AWS Inferentia2 chips, and announced the general availability of Amazon EC2 Trn1n instances equipped with AWS Trainium. These instances provide the highest performance and lowest cost for running compute-intensive generative AI workloads at AWS scale.
4. Amazon Cloud Technology provides transformative generative AI programming language tools to significantly improve productivity: Amazon CodeWhisperer, a programming language generation tool, is now open to all individual developers for free, supporting more than 10 programming languages. CodeWhisperer dramatically increases developer productivity, with 57% faster task completion and 27% greater task success during the preview period.
5. Amazon cloud technology continues to innovate, making next-generation technologies such as generative AI more accessible, helping customers restructure products and business models: Amazon cloud technology has been continuously innovating in the field of generative AI, providing infrastructure for customers from all walks of life And AI/ML services, support customers to embrace the new era of AI, provide extensively verified and easy-to-deploy advanced AI algorithm models, and provide rich and cost-effective cloud resources to help customers grasp new opportunities in the AI era.
A New Wave of Machine Learning Applications
Since generative AI is also a type of artificial intelligence, of course, generative AI is the same as all artificial intelligence technologies, and its capabilities must be provided by machine learning models. These models are pre-trained large models based on a large amount of data, often referred to as foundation models (Foundation Models).
Chen Xiaojian pointed out that since the birth of Amazon, artificial intelligence has been the focus of its attention. Over the past 20-plus years, Amazon has invested heavily in the development and deployment of artificial intelligence and machine learning for customer service and internal operations. So far, Amazon has helped more than 100,000 customers of different sizes and industries to use industry-leading capabilities for machine learning and artificial intelligence innovation.
With the development of generative AI today, it is foreseeable that generative AI and machine learning will reshape applications and customer experience in the future. Amazon cloud technology will make basic models and generative AI benefit a wide range of customers. Make generative AI simpler, more practical, and more cost-effective for customers, so that customers of different sizes and developers of different levels can benefit from it.
With the help of Amazon cloud technology, customers can use the purpose-built machine learning infrastructure to build their own basic models, use pre-trained basic models to build applications, and use the built-in generative AI services to build generative AI in your own way.
"The new wave of machine learning applications has just begun, and the future is full of infinite possibilities. Amazon Cloud Technology also hopes to create more value for customers in this regard." Chen Xiaojian said.
How Businesses and Individuals Can Respond to Generative AI
The outbreak of ChatGPT in 2023 has triggered a carnival of generative AI all over the world. On the one hand, technology giants have entered the competition; on the other hand, many people are also worried that they will lose their jobs. What views and suggestions does Amazon Cloud Technology have on this?
"Amazon Cloud Technology believes that the application of generative AI will not only rely on one large generative model, but will have multiple large models for different scenarios to support applications from all walks of life. The wave of generative AI may just begin .Including Gartner has a similar judgment. Many technologies in this industry are still in a very early stage. Therefore, when choosing a suitable scenario and technology, users need to first consider clearly what problems they need to solve most, and then based on this Make a route and plan that is most suitable for your own business direction. Because only the user can clearly know the pain points of the current business and the problems that need to be solved urgently, and by formulating the route and plan that is most suitable for the enterprise, you can choose what kind of What kind of tools, what kind of cloud platform, what kind of big model can help you achieve your goals.”
Chen Xiaojian pointed out that in essence, artificial intelligence is mainly composed of three most important factors: computing power, algorithms, and data. Among them, the computing power and algorithm are universal, while the data is related to the industry. The richness and accuracy of data will directly affect the final model effect. Therefore, users need to use their own data to feed the large model to serve their own specific scenarios, and they must also be able to help the large model to retrain, so as to find the most suitable solution for their business scenarios.
When talking about whether generative AI will bring unemployment, Chen Xiaojian said that generative AI will inevitably eliminate many repetitive, inefficient, and low-value jobs in the future, but at the same time, it will also release unprecedented productivity. In this way, human beings can do many things that could not be done before, and bring many innovations that were unimaginable before, thus creating more job opportunities. Therefore, generative AI is a very valuable thing for human society. It will bring more value and opportunities to human beings, and make many things that were impossible before possible.
"All the releases of Amazon Cloud Technology this time are aimed at helping users lower the underlying computing power and technical barriers as much as possible. For example, Amazon Bedrock allows customers to customize and train their own models more easily and conveniently." University of Amazon Cloud Technology Wang Xiaoye, director of the data technology expert team in China, said, "For users who want to make their own generative AI models, considering that computing power, algorithms and data are the three most important cores, they should look for the most valuable cloud Service providers. But it should be noted that performance is not the only indicator at this time, cost and cost performance are also very important, that is, it should be realized in the most cost-effective way as possible.”
