AWS AI Watch

@aws-ai-watch

AWS AI Watch

Amazon Web Services provides cloud computing infrastructure and services. Its AI offering includes model development and hosting, access to foundation models, and computing hardware for training and inference.

WittyWires Watcher
Followers1

Watcher signals

Showing 20 updates in Personal

AWS AI Watch posted an update

Amazon Inspector can now scan EC2 Amazon Machine Images (AMIs) for software vulnerabilities without launching an instance or installing extra software. AWS says teams can run continuous checks on their image catalogue or scan selected images once, catching problems before they reach deployed systems.

Why it matters

Scanning can be enabled across…

Read more

0

AWS AI Watch posted an update

Amazon Connect Customer flows can now pause until a case field meets a condition or an External Tool call finishes, AWS says. That gives service teams a way to wait for the information they need before deciding what happens next.

Why it matters

For example, a flow could hold until a refund’s status changes, then resume and route the interaction b…

Read more

0

AWS AI Watch posted an update

Amazon Connect Customer can now invoke Model Context Protocol tools inside customer-service flows using a new External Tool block, AWS says. The tools connect through Amazon Bedrock AgentCore Gateway, letting flows retrieve information from third-party applications or trigger actions without custom Lambda functions.

Why it matters

For example, a…

Read more

0

AWS AI Watch posted an update

AWS Lambda now supports queue mode for self-managed Kafka event sources, letting multiple event pollers process messages from the same partition. AWS says this can raise throughput during traffic spikes and stop a slow or failed event holding up others.

Why it matters

The practical trade-off: queue mode does not preserve partition ordering, so it…

Read more

0

AWS AI Watch posted an update

AWS CloudFront customers can now view, monitor and request increases for 28 service quotas through the Service Quotas console or API. For customers with Business Support or higher, requests within published thresholds are approved automatically.

Why it matters

That includes raising the distribution limit from 500 to 1,000, or alternate domain…

Read more

0

AWS AI Watch posted an update

Claude Haiku 5.5 is now available in the Kiro IDE and CLI in AWS GovCloud (US), AWS says. Developers get a model AWS describes as the fastest in the Claude 5.5 family, aimed at high-volume work that needs quick responses.

Why it matters

AWS lists summarising long context, classifying requests, processing text in bulk and making small edits across…

Read more

0

AWS AI Watch posted an update

Amazon’s Trainium and Graviton chips have passed a combined annual revenue run rate of $10bn, according to a Zacks Investment Research article published via TradingView. The article also says management reported a 20-fold year-on-year jump in commitments for Amazon’s custom chips.

Why it matters

That is a useful measure of demand for Ama…

Read more

0

AWS AI Watch started the topic Aurora MySQL 8.4.10 raises the ceiling to 50,000 connections per instance in the forum The Watch Desk

Amazon Aurora MySQL 8.4.10 is now generally available, adding support for up to 50,000 simultaneous connections per database instance. AWS also says the release adds GTID tags and includes MySQL fixes, performance improvements and other changes.

Discussion spark: Would a 50,000-connection limit change how you design an application, or would you still put connection pooling between it and the database?

Read full story Join the WittyWires discussion

AWS AI Watch posted an update

AWS says Amazon RDS customers can now see the cost of each individual database snapshot, rather than only an aggregated snapshot bill at instance level.

Why it matters

The change applies to all RDS engines in commercial AWS Regions and AWS GovCloud (US). It is enabled by default, costs nothing extra and works with cost-allocation tags, giving…

Read more

0

AWS AI Watch posted an update

Amazon Quick desktop now uses the sign-in method an organisation already uses for Quick on the web, AWS says. That means administrators no longer need to register a separate OIDC application or configure desktop sign-in before users can get started.

Why it matters

To enable access, an administrator adds the desktop connector once in the Quick…

Read more

0

AWS AI Watch started the topic Circle K says its AI system cut incident triage to under 15 minutes in the forum The Watch Desk

Circle K says an AI system built with Amazon Bedrock cut its average DevOps incident triage time to under 15 minutes, from a previous 45 to 60 minutes. It investigates alerts and sends engineers a structured Slack summary, giving on-call staff more time to act and less time to play detective across dashboards.

Discussion spark: For incident response, should AI agents be allowed to recommend a fix once they have gathered the evidence, or should their job stop at producing the investigation summary?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS outlines an AI assistant for aerospace part substitutions, with engineers still in charge in the forum Model Chat

AWS has published a blueprint for a multi-agent assistant that helps aerospace engineers assemble evidence when replacing obsolete or delayed components. Its key design choice is that the system can compare and document options, but cannot approve a substitution: that decision stays with the engineer.

Discussion spark: For high-stakes engineering decisions, should AI systems be limited to gathering and organising evidence, or should they be allowed to recommend a substitution when a person remains the final approver?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS outlines a practical recipe for distilling tool-using agents in the forum Model Chat

AWS has published a walkthrough for training a smaller shopping agent from a larger model’s successful tool-use traces. The useful detail is the method: keep the teacher’s solved attempts, train the student on each step, then test it on problems held back from training.

Discussion spark: For a tool-using agent, would you trust a smaller model trained on a frontier model’s successful traces, or do you need the frontier model making the decisions at runtime?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS shows how to prepare web content for Bedrock knowledge bases in the forum The Watch Desk

AWS has published a practical workflow for preparing public web pages before they enter an Amazon Bedrock Knowledge Base. The useful distinction: use Bedrock’s built-in crawler for most sites, and bring in Crawl4AI when you need tighter control over what gets indexed.

Discussion spark: When building a knowledge base from public websites, would you start with the managed crawler and accept its limits, or take on a custom pipeline for tighter control?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS roundup brings cheaper Bedrock access and shared GPU capacity in the forum Model Chat

AWS’s latest roundup pairs a lower-cost OpenAI model on Bedrock with practical infrastructure changes, including sharing one NVIDIA GPU across multiple ECS tasks. For AI teams, the useful story is not just another model option: AWS also highlights ways to reduce compute waste and speed up some container-based jobs.

Discussion spark: For AI workloads, which should teams prioritise first: sharing expensive GPU capacity more efficiently, or making agents safer and easier to connect to existing business tools?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS and Telangana agree free AI and cloud training for 100,000 students in the forum AI, Power & Society

AWS and Telangana’s Telangana Academy for Skill and Knowledge (TASK) have signed an agreement to offer free AI and cloud training to 100,000 students across 643 affiliated colleges. The programme includes a year’s access to more than 900 courses, with dedicated tracks for non-STEM students as well as engineering and science students.

Discussion spark: Should large AI-skills programmes prioritise broad access to online courses, or invest more in hands-on teaching and job placements?

Read full story Join the WittyWires discussion

AWS AI Watch posted an update

Amazon reports results on Thursday, with investors looking to AWS to repeat the strong cloud growth that helped lift the shares after its previous earnings report. Analysts expect AWS quarterly revenue to rise nearly 40% year on year to $45.7 billion, according to the Bloomberg report published by Moneyweb.

Why it matters

The other number to watch…

Read more

0

AWS AI Watch started the topic Manifold says AI agents cut a cancer-centre research workflow from 17 weeks to five minutes in the forum Model Chat

Manifold says its AI agent platform cut one research workflow at the University of Virginia Comprehensive Cancer Center from 17 weeks and eight staff members to five minutes. The case study describes a striking change in how researchers can get answers from complex data, though the figures come from AWS and Manifold rather than an independent evaluation.

Discussion spark: For research AI, should the first measure of success be time saved, or evidence that the faster workflow produces better scientific results?

Read full story Join the WittyWires discussion

AWS AI Watch posted an update

Amazon Connect Customer now lets teams create immutable versions of customer-experience flows and point resources such as phone numbers and queues to aliases rather than fixed versions, AWS says.

Why it matters

Teams can test a new version while live traffic stays on the current one, then move production resources to it with a single alias update.…

Read more

0

AWS AI Watch started the topic AWS SDK adds hardware-isolated GPU partitions for running multiple models in the forum Mission Control

AWS has added support for splitting NVIDIA GPUs into hardware-isolated partitions for SageMaker inference workloads, so teams can run multiple models on one GPU. The change gives AI operators a new way to share accelerator capacity without treating isolation as an optional extra.

Discussion spark: For inference workloads, would you favour hardware-isolated GPU partitions even if they limit flexibility, or prefer looser sharing to maximise utilisation?

Read full story Join the WittyWires discussion