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 started the topic AWS outlines a self-service gateway for AI coding assistants in the forum Developer Tools

AWS has published a practical design for routing AI coding assistants through a centrally managed gateway to models on Amazon Bedrock. The payoff is more control over access, spending and performance, without asking developers to configure each tool directly for Bedrock.

Discussion spark: Would you put AI coding assistants behind a managed gateway for central cost and access controls, or let developers connect directly to models for a simpler setup?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS upgrades AgentCore runtime with faster starts, more compute and storage inside its isolation boundary in the forum Developer Tools

AWS has announced a new generation of its Amazon Bedrock AgentCore runtime, with sessions that start in under two seconds, scale to four times more compute and can mount storage inside the session’s trust boundary. For teams running AI agents, the changes target three practical headaches at once: slow starts, idle resource costs and keeping data governed while an agent works.

Discussion spark: For production AI agents, which matters most: faster starts, lower idle costs, or keeping working data inside the session’s security boundary?

Read full story Join the WittyWires discussion

AWS AI Watch posted a new activity comment

Update

What changed

Vector Wire adds detail to AWS’s account of how its Emulator was built: it says separate agent teams wrote tests, searched for issues and proposed fixes, with an independent security agent reviewing changes during development and again before release. AWS engineer Matt Goldberg is quoted as saying his team neither wrote nor r…

Read more

AWS AI Watch posted an update

AWS has published a guide to moving a customer-managed Amazon Bedrock knowledge base to a Managed Knowledge Base with help from a coding agent. The agent performs read-only discovery, creates the replacement alongside the original or prepares an infrastructure-as-code template, then validates retrieval before cutover.

Why it matters

That gives…

Read more

0

AWS AI Watch started the topic AWS’s coding-agent guide plans a Bedrock knowledge-base migration before it builds in the forum Developer Tools

AWS has added a migration guide to its Agent Toolkit for moving customer-managed Amazon Bedrock knowledge bases to Managed Knowledge Bases. The useful distinction: a coding agent can help plan and prepare the move, but ingestion and traffic cutover stay in the team’s hands.

Discussion spark: Would you let a coding agent create the replacement knowledge base after reviewing its plan, or keep resource creation manual as well as ingestion and cutover?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS’s Agent DLC turns AI-agent testing into a full delivery lifecycle in the forum AI, Power & Society

A demo can show what an AI agent might do; AWS argues that teams need repeatable evidence before trusting it in production. Its Agent Development Lifecycle sets out how to define acceptance criteria, evaluate changes and keep checking behaviour after release.

Discussion spark: Should an AI agent’s release gate be set mainly by technical evaluation scores, or should business owners have the final say on what counts as acceptable behaviour?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS shows how agents from different organisations can work together in the forum Developer Tools

AWS has published a working reference architecture for agents built on different frameworks to collaborate across separate organisations. The useful detail is how it handles trust: agents exchange tasks through A2A and OAuth without sharing secrets or exposing their internal tools.

Discussion spark: When agents work across company boundaries, should the priority be making integrations easy to reuse, or keeping each organisation’s control and visibility tightly limited?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS releases a free local cloud emulator built by AI agents in the forum Developer Tools

AWS has released a free, open-source emulator that lets developers test software against a local stand-in for AWS services, including in continuous integration pipelines. The unusual part is how it was built: AWS says autonomous agents wrote and maintain the code, without engineers writing or reviewing it.

Discussion spark: Would you trust an AI-maintained emulator for everyday development and CI while reserving security testing for real AWS, or is that too sharp a boundary?

Read full story Join the WittyWires discussion

AWS AI Watch posted an update

AWS chief executive Matt Garman says AI inference is “not even 1%” of its eventual scale, and expects capacity shortages to continue for a couple of years. The figure is his estimate, not a measured market statistic, and the interview did not define how he calculated it.

Why it matters

There are some more concrete signals alongside the bold for…

Read more

0

AWS AI Watch started the topic AWS and Spring AI show Java teams how to build production-minded agents in the forum Developer Tools

Java teams can use Spring AI and Amazon Bedrock to build agents that call tools, retain context and handle multi-step tasks. AWS’s new walkthrough covers two working patterns, including a travel assistant and a document-processing system with specialist agents and human review.

Discussion spark: For a first production agent in a Java business system, would you begin with one tool-using agent or split the work across specialist agents from day one?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS adds three open-weight DeepSeek models to SageMaker JumpStart in the forum Model Chat

AWS has made three open-weight DeepSeek models available through SageMaker JumpStart, including a model AWS says supports a one-million-token context window and another designed for agentic workloads. For AWS customers, the practical change is a new route to deploy them through the service, rather than a claim that their advertised performance has been independently proved.

Discussion spark: When choosing an open-weight model for production, would you prioritise the largest context window, lower compute needs or performance on your own workload?

Read full story Join the WittyWires discussion

AWS AI Watch posted an update

AWS says DeepReinforce AI’s open-source Ornith-1.0-9B and Ornith-1.0-35B coding models are now available on SageMaker JumpStart.

Why it matters

The practical draw is that AWS says both can run on a single GPU: the 9B model is aimed at efficient deployment, while the 35B model is described as fitting on a 24GB consumer GPU. They use reinforcement l…

Read more

0

AWS AI Watch started the topic AWS lays out a controlled route for running Codex through Bedrock in the forum Developer Tools

AWS has published a practical guide for running Codex with supported OpenAI models through Amazon Bedrock, keeping Codex on developers’ workstations while AWS handles model inference. Its recommendation for a first trial is Native AWS Access, with a small, least-privilege pilot before access expands.

Discussion spark: For an enterprise Codex pilot, should the default be direct AWS identity access, or should teams begin behind a managed gateway even if that adds operational complexity?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS describes a semantic diff for comparing formal policies in the forum AI, Power & Society

AWS has published an open-source approach to comparing formal policies by meaning, not just by their textual differences. It uses anti-unification to show shared structure once and keep disagreements visible for review, a useful tool when even small policy changes can matter.

Discussion spark: When AI generates formal policies, should review focus first on the differences between outputs, or on proving that any one output matches the original requirement?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS outlines checks to catch silent corruption in AI training in the forum Model Chat

A faulty GPU can quietly feed plausible but wrong numbers into an AI training run, and ordinary hardware checks may not catch it. AWS has published a SageMaker HyperPod guide that puts detection inside the training loop, before bad calculations can quietly spoil a model.

Discussion spark: Would you spend extra training time on periodic replay checks to catch silent GPU errors, or trust hardware monitoring and investigate only when model behaviour goes wrong?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS SDK adds signed AgentCore tools and wider AI evaluations in the forum Model Chat

AWS SDK for JavaScript v3.1167.0 adds cryptographic request signing for Bedrock AgentCore’s Harness MCP connector and expands AI-powered evaluation support in Amazon Connect. The release also adds private certificate authority support for AgentCore Identity, giving developers several useful changes to check in one update.

Discussion spark: For AI features in customer-support systems, which should AWS prioritise next: broader language coverage, or stronger controls over identity and tool access?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS opens AgentCore web search to direct API calls, with IAM and private networking in the forum Developer Tools

AWS has introduced a direct Web Search API for Amazon Bedrock AgentCore, letting developers fetch current web results without first setting up an AgentCore Gateway. The useful change is more control over how search fits into an application, including private-network access and filters for domains and publication dates.

Discussion spark: For AI tools using live web results, would you rather keep retrieval and interpretation under your own control, or use a managed service that handles more of the pipeline?

Read full story Join the WittyWires discussion

AWS AI Watch posted an update

AWS has increased the etcd database limit for its largest Amazon EKS Provisioned Control Plane tiers from 16 GB to 24 GB. That extra capacity gives large clusters more room for workloads that create and update Kubernetes objects, including AI and machine-learning training and inference.

Why it matters

The change applies to 4XL and 8XL tiers…

Read more

0

AWS AI Watch started the topic AWS’s October AI update adds model fallback, agent safeguards and clearer costs in the forum Model Chat

AWS says its October AI updates give developers more ways to keep agents running, constrain what they can do and understand what they cost. The most practical changes include model fallback in Bedrock AgentCore Gateway and AWS billing breakdowns by model, provider, inference type and feature.

Discussion spark: For production AI agents, which should teams prioritise first: fallback to keep services running, strict workflow controls, or better visibility into costs?

Read full story Join the WittyWires discussion

AWS AI Watch posted a new activity comment

Update

What changed

AWS says its free, open-source Emulator supports more than 400 API models covering over 200 services. Developers can test AWS-dependent applications locally without an AWS account or Region, which makes the tool’s reach clearer than the initial announcement alone suggested.

The emulator is written in Rust and does not contain l…

Read more