Shortly after the OpenAI mystery project "Q*", Amazon Web Services launched Q, an enterprise-level generative AI assistant with a similar name.
In a recent keynote at the Re:Invent conference in Las Vegas, Amazon Web Services CEO Adam Selipsky officially announced the launch of Amazon Q, "You can use Q to easily have conversations, generate content, and take action. qFully understand your systems, data repositories, and operational needs. Selipsky said.
Amazon Q is trained by 17 years of knowledge and experience from Amazon Web Services to make recommendations on cloud infrastructure that meets business needs, output blog posts, help applications**, and search and analyze enterprise data.
How to help developers
Accelerate business development
Amazon Q is an expert in building, deploying, and operating applications and workloads on Amazon Web Services. According to the official introduction, users can connect Q to applications and software specified by the organization (such as Salesforce, Jira, Zendesk, Gmail, and Amazon S3 storage instances, etc.), and customize configurations accordingly.
Q is able to index against all relevant data and content, and "Xi" all aspects of the current business, including organizational structure, core concepts, and product names.
For example, a company can use a web application to ask Q to analyze which features users are experiencing problems with and how they can improve themIt is also possible to upload files directly as you would with ChatGPT (Word documents, PDFs, electronic **, etc.) and ask content-related questions. q provides responses and references through connections, integrations, and data, including business-specific data. Based on these questions, Amazon Q gives a clear answer and lists the citations.
Users can ask as many rounds of questions as they want to get more detailed answers, find the best option for their workload, and get step-by-step guidance on basic operations.
Q not only answers questions, but also acts as an assistant to generate or summarize blog posts, newsletters, and emails. It also provides a set of configurable plugins for general actions at work, including automatic ticket creation, through specific teams in Slack, and updating dashboards in ServiceNow, among other things.
To prevent errors, Q asks users to check their recommendations before taking action and to present the results for verification.
qIt can be accessed through the Amazon Web Services management console, various web applications, and chat applications such as Slack, and has a thorough understanding of the products and services of the Amazon Web Services family. According to Amazon Web Services, Q can understand the nuances of various application workloads on Amazon Web Services, and even applications that only need to run for a few seconds or have little access to stored content can accept Q's guidance and operations.
Q can also solve common problems such as network connectivity, analyze network configurations to provide repair recommendations. "If there is an error in the console, you can press the Amazon Q button to troubleshoot. Q will look into the bug and suggest how to fix it. Amazon Q also understands networking and "can help resolve connectivity issues quickly," Selipsky said.
"I'm a firm believer that this will be a productivity change, and I hope that people from different industries and in different roles will benefit from Amazon Q." ”
Generate, Interpret**
Q is combined with the Amazon CodeWhisperer service to build and interpret applications**. In supported IDEs (e.g., Amazon CodeCatalyst), Q can generate tests for users to gauge their level of quality.
In addition, Q can create new software features, perform transformations, and update drafts and documentation for packages, repositories, and frameworks, using natural language to refine and execute plans.
Selipsky said that a small team within Amazon Web Services managed to upgrade thousands of applications from J**A 8 to J**A 17 using Q in just two days, and even completed the corresponding tests.
Q's conversion function only supports upgrades from J**A 8 and J**A 11 to J**A 17 (to follow.) .NET Framework to cross-platform. .NET conversion), and all related features, including conversions, are required with the CodeWhisperer Professional subscription service.
Selipsky added that Amazon Q will be able to transfer applications from Windows.NET Framework migration to cross-platform on Linux. .net, which is a great idea, but often challenges in practice due to its dependency on Windows alone.
Amazon Web Services says it is also using Q to enhance more first-party products, such as Supply Chain and QuickSight, a business analytics service.
Q can provide visualization options for business reports in QuickSight, automatically adjust formatting, or answer user questions based on citations and data in reports. In the Supply Chain, Q is able to respond with the latest analysis results such as "Why is my delivery order delayed?"." and so on.
Q is also gradually moving into the contact center software Amazon Connect. With Q's support, agents can now quickly get answers to user questions, corresponding steps, and links to background information, so that they can do away with tedious and inefficient manual searches. Q can also generate call summaries to help supervisors track service progress in the future.
Privacy and security
Throughout the presentation, Selipsky repeatedly emphasized that Q's answers and suggestions were completely controllable and supportive of screening. Q will only return information that the user has permission to view, and admins can restrict sensitive topics and require Q to filter out inappropriate questions and answers if necessary.
To alleviate the hallucination problem (i.e., the fabrication of facts that is common in generative AI systems), administrators can require Q to extract knowledge only from internal company documents and not use knowledge from the underlying model. Selipsky said that the underlying model that drives Q is a combination of models provided by Amazon's AI development platform Bedrock, including Amazon's original Titan series, and will never use user data for model training.
Currently, more than a dozen companies have explicitly banned or restricted the use of ChatGPT, reflecting concerns about the risk of leaks that could result from entering data into chatbots. Selipsky emphasizes, "If your users don't have access to something in the first place, they still won't have access to it after using Q." q Understand and respect the user's current identity, role, and permission ......We will also never use business content to train the underlying model. ”
In addition to the high emphasis on privacy, Q seems to be a strong response from Amazon Web Services to Microsoft's Azure Copilot, which in turn is Microsoft's response to Google's Duet AI. Both Azure Copilot and Duet AI are chat assistants that provide cloud users with recommendations on application and environment configurations, as well as assist with troubleshooting by identifying potential issues and providing answers.
Off topic
Ray Wang, founder and principal analyst at Constellation Research, said in an interview that he thinks Q is the most important launch at Re:Invent. "It's arming developers with AI to help them succeed. ”
It's clear that Amazon Web Services sees the key conclusion of a recent survey that most vendors experimenting with generative AI don't know how to incorporate new technologies into business use cases and turn them into productivity.
Christoph Albrecht, Data Engineering and Analytics Consultant at BMW Group, said: "The BMW team needed to quickly extract and interpret new data to deliver the precise experience that customers expect. The new Amazon Q feature in Amazon QuickSight helps our analysts build dashboards in hours instead of days. ”
Notably, Anthropic CEO and co-founder Dario Amodei also took the stage alongside Selipsky at Re:Invent. In September, Amazon Web Services invested up to $4 billion in the AI startup to provide Anthropic with more cloud infrastructure and chips to train and run its models. Anthropic's Claude LLM will power a range of AWS products, including AppFabric.
The seven founders of Anthropic are all from OpenAI and have been deeply involved in many studies such as OpenAI's GPT-3 and the introduction of strong chemical Xi with human preferences. The reason for leaving OpenAI is said to be because it "had a different vision in terms of model security from the beginning." "Anthropic's appearance at Re:Invent shortly after the end of OpenAI's high-level battle shows that Amazon Web Services is committed to generative AI.