Chanu Damarla

[Garry Tan @garrytan]

Everyone building AI agents is focusing on building the prefrontal cortex. Planning. Reasoning. Multi-step chains. There's value here. CEO-stuff.

But also, a reframe: there is value in building the cerebellum. It's offloading boring tasks into reflex so the complex thought can focus.

Your mortgage gets paid by a standing order, not a committee. The things that are not fun, not interesting, but have to be done? Done. Most agent frameworks will fail because they treat all cognition as high cognition.

The winners will nail the boring stuff first.

[spidey @lochan_twt]

The day a blind man sees. The first thing he throws away is the stick that has helped him all his life

[Chanu Damarla @cd2k]

@jspujji GA

[Tim O'Reilly @timoreilly]

After listening to Andy Jassy’s comments  about using AI to deliver a better experience to customers and Ford CEO Jim Farley’s prediction of massive job losses from AI, it occurred to me that whether AI creates or destroys jobs depends on where companies are in the innovation cycle. Companies in new markets are expansive, both in terms of customer acquisition and in terms of employment; those in mature markets are stable or in decline. AI will have a different effect depending on where you are in this cycle.

https://t.co/GD0SBTdfYK

[Danilo Poccia @danilop]

Introducing Amazon Bedrock AgentCore: Securely deploy and operate AI agents at any scale (preview) 👉 To quickly and securely deploy and operate AI agents at any scale using any framework and model

https://t.co/1CYrpw9vdk

#AWS #AI #GenAI #MCP

[Swami Sivasubramanian @SwamiSivasubram]

🚀 Introducing an easier way to troubleshoot your @awscloud environment—check out Amazon CloudWatch investigations: https://t.co/Dh9eHADy21

My favorite standout features from this launch:
🟠 An AI agent that looks for anomalies in your environment and helps reduce mean time to resolution (MTTR)
🟠 Ability to start investigations from more than 80 AWS consoles or configure to auto trigger as a CloudWatch alarm action
🟠 Collaboration features that allow teams to work together and add findings or review suggestions for potential root cause hypothesis
🟠 Integration with popular communications channels, such as Slack and Microsoft Teams (keep your work going without needing to learn new systems)

Try CloudWatch investigations and accelerate troubleshooting 🔍

[Swami Sivasubramanian @SwamiSivasubram]

At @awscloud, we believe agentic AI will be critical to nearly any customer experience. We welcome A2A joining The Linux Foundation and envision it will create broader opportunities for anyone building AI-powered apps. We intend to support the community with project contributions, and access to the broadest and deepest set of agentic frameworks, protocols, and services.

https://t.co/DY2fadT2a8

[Mahendra Damarla @damarlalab]

Excited to get this work finally out! 9 years in the making!!

[Swami Sivasubramanian @SwamiSivasubram]

We just made Amazon SageMaker even better. Customers are saving development time and simplifying access control management in a single environment with SageMaker Unified Studio, now generally available. 

This new experience breaks down data silos and unifies with familiar AWS tools so that data engineers, data scientists, data analysts, ML developers, and other data practitioners can collaborate seamlessly and securely. Businesses like @NatWestGroup and Carrier are already seeing the benefits of having their AWS data and AI services in one place—it allows their teams to work faster and more efficiently.

Here’s a good breakdown in TechTarget on what the next-generation Amazon SageMaker can do for your business➡️ https://t.co/kYxswq11Gu

[Swami Sivasubramanian @SwamiSivasubram]

🚀 We just launched even more ways to for customers to deploy DeepSeek-R1 models on @awscloud: https://t.co/WFKNTwyvvk

🟠 Amazon Bedrock Marketplace for DeepSeek-R1

🟠 Amazon SageMaker JumpStart for DeepSeek-R1

🟠 Amazon Bedrock Custom Model Import for DeepSeek-R1 distilled Llama models

🟠 Using Amazon EC2 Trn1 instances powered by AWS Trainium for the DeepSeek-R1 distilled Llama models

🟠 Amazon SageMaker AI supports running distilled Llama and Qwen DeepSeek models

🟠 DeepSeek models can be trained on Amazon SageMaker AI through @huggingface integration

We’re also making it easier for customers to make the most of their data with @deepseek_ai. Amazon OpenSearch Service’s vector engine now supports a connector that works with DeepSeek R-1 models hosted on Amazon SageMaker and Bedrock. The work done was also contributed upstream to OpenSearch so the community can benefit. (https://t.co/UhWOyk6ETf)

We're committed to keeping our customers’ data private and secure. Amazon Bedrock’s enterprise-grade security features ensure that data is not shared with model providers, and is not used to improve the models. Our Guardrails are also integrated with other Bedrock tools to build safe and more secure GenAI applications aligned with responsible AI policies. As always, we recommend customers apply guardrails on top of their models for their applications.

Our engineering team just released a great blog that explains how customers can get started with DeepSeek-R1 models on AWS, from initial setup to deployment options. Whether you're building your first AI application or scaling existing solutions, these methods provide flexible starting points based on your team's expertise and requirements.

[Werner Vogels @Werner]

If you want to run DeepSeek R1 on #AWS, these are two starting points: 1) Bedrock notebook steps: https://t.co/vXucRjh8v2 2) load into Sagemaker https://t.co/AVMlJXP0Od (thanks to @DGallitelli95). As always, consult your legal team before doing so...

[Amazon Web Services @awscloud]

Build powerful analytics & AI/ML apps on a single copy of data. ☁️⚡️💻

Amazon SageMaker Lakehouse unifies your data across Amazon S3 data lakes & Amazon Redshift data warehouses. Query data in-place with all Apache Iceberg compatible tools & engines.

👉 https://t.co/WytnG09GOZ

[Ippokratis Pandis @Hippotas]

The ZeroETL teams at AWS have been busy all this year and these efforts are paying off. We are excited to GA the ZeroETL integrations from Amazon Aurora PostgreSQL and Amazon DynamoDB to Amazon Redshift!

https://t.co/EDhY6ZdC5L

(1/n)

[Swami Sivasubramanian @SwamiSivasubram]

The @awscloud team is hosting a 24-hour @Twitch stream about Amazon Q on 10/9. Our GenAI experts will highlight the latest use cases and demos showing how Q is transforming businesses across every industry. Hope you can check it out ➡️ https://t.co/Watfd1AjVM

[Swami Sivasubramanian @SwamiSivasubram]

We hear from customers who are interested in using the power of AI to help their sales team scale their business. Internally, AI is one of the tools that Amazonians use daily to improve our productivity and do things faster and more efficiently. In that vein, our latest ML blog gives a great inside look at how @awscloud sales teams are using Account Summaries—one of our first production GenAI use cases built on Amazon Bedrock. Account Summaries help us stay customer obsessed by generating 360-degree views of an account, available on demand and delivered proactively ahead of meetings via Slack. They integrate both structured and unstructured data, including key metrics, real-time web data, ML insights and AI-driven recommendations.

Since its internal rollout last year, more than 100,000 summaries have been generated by our sellers, saving them 35 minutes per briefing. Check out our ML blog to learn how Account Summaries are helping our field teams scale and deliver better customer outcomes. https://t.co/6Yj6uKHNVz

Here’s part of a sample output from Account Summaries:

[Swami Sivasubramanian @SwamiSivasubram]

At @awscloud, we're committed to providing our customers with the most advanced AI capabilities designed to support a wide range of workloads. That's why we're offering batch inference on Amazon Bedrock for foundation models from leading AI companies like Anthropic, Meta, Mistral AI, and Amazon at 50% of the on-demand inference pricing. This will help organizations leverage the power of batch processing for their AI workloads while optimizing costs. Learn more here: https://t.co/imQyWsyD9N

[Dr. Manu Kumar 👋🏽 @ManuKumar]

As college students or parents of students, one of the most time-consuming things we do is take the information we receive in emails or newsletters and mark the important stuff on our own calendars. It's always been frustrating for me that there isn't an easier way to do this!

Well, now there is. Thanks to @carencioffi at @AgendaHero (a @K9Ventures and @upfrontvc portfolio company), you now have the magic to make this very important but tedious task a breeze.

Check out this video from Caren about what Agenda Hero's Magic can do for you.

If you're a student or a parent, you have to try it out.
https://t.co/PSAoB7Szrn

[Mehul Shah @Mashah08]

I’m excited to announce the release of the #Aryn Partitioning Service: https://t.co/XiVKdv7aot. It takes complex PDFs with tables and infographics and returns JSON. It’s up to 6x more accurate (mAP) and 5X faster than alternatives. And, it’s free to get started.

[Han Zhao @hanzhao_ml]

🚨🚨 We are hiring! RT appreciated!

Prof. Rui Song (https://t.co/i9PuYD48f0) and I will recruit post-doc scientists through Amazon’s post-doc program (https://t.co/2zXRzdDrSg).

[Noritaka Sekiyama @moomindani]

今日はこれから関西大学でSoftware Engineering at Amazonというタイトルで授業(ゲスト講師)をやります。去年はオンラインでしたが今年は対面です。