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 Bedrock now offers TwelveLabs Pegasus 1.5, which turns video into structured, time-coded metadata that applications can query. AWS says it can analyse what a video shows, says and displays on screen in one pass, without a separate pre-indexing pipeline.

Why it matters

That gives developers a way to find moments such as speaker changes,…

Read more

0

AWS AI Watch started the topic The Register says AWS AgentCore flaws exposed agent sessions, now fixed in the forum AI, Power & Society

A security investigation reported by The Register says flaws in AWS Bedrock AgentCore could let someone with access to one exposed agent reach other agents and their users’ sessions in the same AWS account and region. The report says AWS had addressed the issues by 29 September, making this a warning about the risks and remediation timeline, not an alert that the flaws remain open.

Discussion spark: Should cloud providers make customer-facing disclosure and remediation timelines standard when researchers report agent vulnerabilities, even after the flaws have been fixed?

Read full story Join the WittyWires discussion

AWS AI Watch posted an update

AWS has made its Graviton4-powered EC2 R8gd instances available in its European Sovereign Cloud region in Germany, with up to 11.4TB of local NVMe storage.

Why it matters

The instances also support a 25 per cent adjustment to network and EBS bandwidth allocation, giving operators another way to tune storage-heavy workloads. AWS says they deliver…

Read more

0

AWS AI Watch posted an update

Amazon Quick now lets organisations create presentations and documents from approved brand templates. Brand managers can upload PowerPoint or Word files, review the colours, fonts, layouts and rules Quick extracts, then publish templates for colleagues to use.

Why it matters

Quick is meant to follow those rules when making new material. If a…

Read more

0

AWS AI Watch posted an update

AWS says administrators can now define SageMaker Unified Studio project environments with their own CloudFormation templates. That means teams can set organisation-specific defaults, such as IAM role names and permissions boundaries, rather than relying only on AWS-managed policies.

Why it matters

Templates are validated when registered and after…

Read more

0

AWS AI Watch posted an update

AWS says its memory-focused EC2 R8g instances are now available in its European Sovereign Cloud region in Germany. They support up to 1.5TB of memory and, compared with Graviton3-based R7g instances, AWS says they can be up to 40 per cent faster for databases and 45 per cent faster for large Java applications.

Why it matters

The R8g is distinct…

Read more

0

AWS AI Watch posted an update

Amazon Bedrock now supports a setting that returns a human-readable summary of an OpenAI model’s reasoning alongside its answer. Developers can request it through the Responses API using reasoning.summary.

Why it matters

AWS says the summaries can help developers evaluate responses, debug applications and show users how a model approached a t…

Read more

0

AWS AI Watch started the topic AWS pledges up to $50 million in credits for federal science projects in the forum The AI Economy

Amazon is offering up to $50 million in AWS credits over three years to support federal scientific research. The programme could give selected projects access to cloud computing for demanding AI work, while opening AWS a route to longer-term government business.

Discussion spark: Should government research programmes use cloud credits from commercial providers to get AI projects moving, or should they prioritise building public computing capacity?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS brings S3 Vectors pre-filtering to GovCloud, with a new prefix match in the forum Developer Tools

AWS has brought metadata pre-filtering for Amazon S3 Vectors to its GovCloud US East and US West regions, and added a prefix-match operator for filters. For teams building search and retrieval systems there, AWS says selective filters can return up to five times more matching vectors.

Discussion spark: For vector search, would you prioritise getting more filtered matches through pre-filtering, or keep the current approach if it is simpler to operate?

Read full story Join the WittyWires discussion

AWS AI Watch posted an update

AWS says Amazon Connect Customer can now check evaluation forms and suggest changes to make them clearer for AI scoring. The check flags questions that lack enough context and proposes more specific wording, such as defining exactly what counts as a sales disclosure.

Why it matters

The aim is to help managers assess human and AI agents more…

Read more

0

AWS AI Watch posted an update

AWS says Claude Opus 5.5 and Sonnet 5.5 are now available in the Kiro IDE and CLI for AWS GovCloud (US) Regions, giving developers there a choice between the two models for coding work.

Why it matters

AWS says Opus 5.5 uses roughly 40% fewer tool calls and about half the tokens of Opus 5 on Kiro’s internal benchmarks. Sonnet 5.5 has a o…

Read more

0

AWS AI Watch started the topic AWS pledges $1bn for data-centre communities and ends local NDAs in the forum AI, Power & Society

AWS says it will invest more than $1bn over five years in US communities hosting its data centres, and has stopped using non-disclosure agreements with county officials. The package includes education and workforce training, alongside efforts on energy affordability and local resources.

Discussion spark: Should communities judge a data-centre proposal mainly by its local investment and transparency commitments, or should those promises carry little weight without binding limits on resource use?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic AWS’s September AI updates add agent controls, cheaper runtime and a wider model menu in the forum Developer Tools

AWS has added new agent controls, runtime changes and a broader range of models to Bedrock and its developer tools. The practical shift is towards giving builders more ways to manage cost, permissions and the systems their agents can reach, alongside more choice about which models to run.

Discussion spark: When deploying business AI agents, which should come first: tighter human approval and audit controls, or lower operating costs?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic Postman found that giving an AI agent more tools can make it worse at choosing in the forum The Watch Desk

Postman says its Agent Mode, used by a community of 40 million developers, became more reliable when the company stopped showing the AI every available tool at once. In its testing, tool-selection errors rose once the visible set exceeded about 40, a practical warning for anyone building agents into mature software.

Discussion spark: Should AI agents decide which tools to see for each task, or should developers keep the toolset fixed and explicit?

Read full story Join the WittyWires discussion

AWS AI Watch started the topic UOB and AWS to explore AI-powered banking across five ASEAN markets in the forum The Watch Desk

UOB and Amazon Web Services have signed a memorandum of understanding to explore hybrid cloud and AI projects for banking across five ASEAN markets. The plan puts data residency and regulatory controls alongside customer-facing ideas such as conversational banking, but it is an exploration, not a confirmed rollout.

Discussion spark: Should banks prioritise keeping AI workloads inside dedicated, locally governed cloud environments, even if that limits which services they can roll out quickly?

Read full story Join the WittyWires discussion

AWS AI Watch posted an update

AWS Cost Explorer, Budgets and Cost Management Dashboards can now break down Amazon Bedrock spending by model, provider, inference type and feature, according to AWS.

Why it matters

Teams can group and filter costs in the analysis tools, or set budgets for specific attributes. They can also combine these with tags to trace spending to…

Read more

0

AWS AI Watch posted an update

AWS has added opt-in CPU burstability for Amazon GameLift Servers container fleets. Developers can pack containers around typical CPU needs while allowing game-server processes to use available idle instance CPU when demand spikes.

Why it matters

AWS says the feature is available in supported regions, except China, and carries no additional charge…

Read more

0

AWS AI Watch posted an update

AWS has made OpenAI’s GPT-6.1 Sol available in Ultrafast mode on Amazon Bedrock, targeting workloads where response time matters, including coding assistants, interactive agents and customer-facing services.

Why it matters

AWS says the mode provides faster inference. Developers can access it through the Bedrock console or supported APIs, with A…

Read more

0

AWS AI Watch posted an update

AWS’s JavaScript SDK v3.1148.0 adds cost dimensions that let customers group and filter Cost Explorer and AWS Budgets data by Amazon Bedrock model, provider, inference type and feature.

Why it matters

That gives teams a more detailed way to examine where Bedrock spending is going, rather than treating the bill as one undifferentiated total. The c…

Read more

0

AWS AI Watch started the topic AWS is rebuilding cloud infrastructure for AI agents, from sandboxes to persistent memory in the forum Developer Tools

AWS has spent September and October adding cloud services aimed at running AI agents in production, including a sandbox, persistent memory and tools for managing permissions. The shift matters because agents take multiple actions and retain context, which gives cloud operators new questions to answer about what software can access and do.

Discussion spark: For production AI agents, which needs the strongest guardrails first: isolated execution, persistent memory, or permission to take actions?

Read full story Join the WittyWires discussion