• Cloudbites
  • Posts
  • AI Updates you Need to Know from AWS re:Invent

AI Updates you Need to Know from AWS re:Invent

PLUS: Missed re:Invent? Catch-up on the AWS Events Youtube Channel

In Today’s Cloudbites:

💡 AI Updates you Need to Know from AWS re:Invent

📚 This Week’s Top AI Resources

☁️ PLUS: Missed re:Invent? Catch-up on the AWS Events Youtube Channel

Read time: 4 minutes

Hi friends, welcome back to Cloudbites. Today, I’ll be sharing the key AI updates you need to know from AWS re:Invent. I’ll also be recommending some of my favorite talks from the week which you can watch on-demand.

💡 AI Updates you Need to Know from AWS re:Invent

  1. Amazon Q: A New Type of Generative AI Assistant.

    Amazon Q is a new AI assistant designed to simplify cloud development for businesses.

    It's been trained on 17 years worth of AWS knowledge and data, and goes beyond just answering simple questions.

    This new service can help with network troubleshooting, conversational Q&A, and selecting optimized EC2 Instance types (Find out more here).

  2. New Features Announced for Amazon Sagemaker.

    SageMaker Canvas is a new visual point and click interface, which allows business analysts to create a machine learning model which generates predictions without coding.

    To address the growth of Large Language Models (LLMs), Amazon has also launched a new version of Large Model Inference Deep Learning Containers.

    Its new inference capabilities mean you can now deploy multiple foundation models from the same endpoint. (Find out more here).

  3. Powerful New Capabilities for Amazon Bedrock.

    Launched earlier this year, Amazon Bedrock is an AWS service that offers foundation models (FMs) through a single API. In re:Invent, there were 3 major additions to Bedrock:

    • Guardrails for Amazon Bedrock - Helps users implement customized safeguards and responsible AI policies.

    • Agents for Bedrock  - Allows generative AI to complete business processes using existing company systems e.g. taking sales orders or checking product availability.

    • Fine-tuning for Bedrock - Customers can now customize FMs using their own data to build applications specific to their use-case.

  4. You can now use Plain Language with Amazon Lex.

    You can now build a conversational bot using the new descriptive bot building feature in Amazon Lex. Instead of writing code, you can describe in plain English what you want the bot to accomplish.

    Amazon Lex will then automatically generate intents, training utterances, prompts and a conversational flow to bring the bot to life.

    This will save many hours of manual work for developers. (Find out more here).

  5. The Next Generation of AWS-Designed Chips.

    During re:Invent, Amazon announced two new AWS-designed chips - AWS Graviton4 and AWS Trainium2.

    These chips will help with training AI models faster, with lower energy usage and costs. (Find out more here).

📚 This Week’s Top AI Resources

  1. Recall: Summarize online content and generate a repetition-based study-plan with the help of an AI tutor.

  2. ExplainThis: Access explanations of key cloud concepts without leaving your webpage.

  3. Watto AI: Integrate your data from multiple platforms to automatically generate summarized, high-quality articles.

  4. aiPDF: Summarize, get key insights from, and chat with any file. 

  5. Simplescraper: Extract data from any source and easily compare data between different sources.

☁️ [PLUS] Missed re:Invent? Catch-up on the AWS Events Youtube Channel

Last week, AWS re:Invent saw over 50,000 attendees, with over 2,900 technical sessions. If you didn’t get a chance to join in-person, many of the talks are now available on the ‘AWS Events’ YouTube Channel.

Here are my top picks from the week:

  1. Keynote - Swami Sivasubramanian (VP of Data & AI at AWS).

  2. Keynote - Adam Selipsky (CEO at AWS).

  3. Innovation Talk - Innovate faster with Gen AI.

  4. Innovation Talk - From hype to impact: Building a Gen AI architecture.

  5. AI Talk - Use Gen AI to name your pet after your favorite song.

That's all for now. Thanks for reading! 😊

P.S. How was today's email? Reply directly with your feedback, or DM me on Twitter @techwithlucy