SageMaker AI adds GPU partitioning for real-time endpoints
SageMaker AI adds MIG-based GPU partitions for real-time inference, with AWS listing support for four instance families in seven regions.
SOURCE / AWS
THE NEWS DESK
Engineering & AI news
The releases, research, and infrastructure changes worth understanding. Original summaries and engineering analysis, with links to the announcements behind them.
423 articles · Ordered by publication date
Aurora PostgreSQL adds post-quantum TLS key exchange
AWS announces post-quantum TLS key exchange for Aurora PostgreSQL, with ssl_groups configuration supported in versions 18 and higher.
SOURCE / AWS
AWS brings ACME certificate automation to GovCloud
AWS Certificate Manager now supports ACME clients for public TLS certificate issuance and renewal in both AWS GovCloud (US) regions.
SOURCE / AWS
AWS connects Step Functions workflows to AI agents through AgentCore Gateway
AI agents can invoke existing Standard and Express workflows through Amazon Bedrock AgentCore Gateway, leaving business-process execution in a defined workflow.
SOURCE / AWS
Amazon MSK Connect adds in-place plugin and worker-configuration updates
AWS says existing connectors can receive custom-plugin and worker-configuration updates while retaining committed offsets, without being deleted and recreated.
SOURCE / AWS
AWS Backup adds cross-account support in European Sovereign Cloud
AWS Backup now supports organization-wide policies, cross-account backup copies and delegated administration in the European Sovereign Cloud’s Germany Region.
SOURCE / AWS
AWS releases AgentCore CLI 1.0 with direct resource workflows
Generally available AgentCore CLI 1.0 adds direct resource operations, consistent global flags and a migration command for existing projects.
SOURCE / AWS
Amazon Quick adds default-on smart approvals for agent actions
AWS says the review layer automatically approves routine agent actions, routes higher-impact steps to users and leaves existing access policies unchanged.
SOURCE / AWS
Amazon EKS Dashboard adds fleet-wide CA rotation visibility
EKS platform teams can now see CA rotation status, approaching expirations and upcoming automated rotation events across accounts and Regions.
SOURCE / AWS
AWS Identity Store adds EventBridge change notifications
IAM Identity Center’s Identity Store now sends change notifications to EventBridge in each enabled Region, supporting identity-driven workflows.
SOURCE / AWS
AWS brings Marketplace and Amazon Business catalogs into one storefront
The unified storefront combines software and physical-goods catalogs under shared governance, but keeps distinct purchasing paths and requires access to both services’ APIs.
SOURCE / AWS
AWS brings SOCI lazy loading to Lambda Managed Instances
AWS adds SOCI lazy loading for OCI-packaged Event Functions on Lambda Managed Instances, fetching needed files instead of the full image at startup.
SOURCE / AWS
GitHub Octoverse reports record code activity and rising agent use
GitHub reports 17.9 billion pushes in its 2026 reporting year, but higher commit counts do not establish productivity gains from agents.
SOURCE / GitHub
GitHub adds Copilot projects and usage-based enterprise access at Universe 2026
GitHub says Copilot projects support parallel agent work across repositories, while a new enterprise purchasing option charges for AI credits without a separate Copilot subscription.
SOURCE / GitHub
Amazon Neptune adds online index rebuild to reclaim storage
AWS says Neptune can reclaim fragmented index storage without going offline, starting with engine 1.5.0.0 on provisioned writer instances.
SOURCE / AWS
Amazon ECS adds GPU sharing with NVIDIA MPS
ECS tasks can now share an NVIDIA GPU, with per-container memory requirements, optional compute caps and daemon recovery on Managed Instances.
SOURCE / AWS
Cloudflare reports 49 Internet disruptions in Q3 2026
Government-directed shutdowns led Cloudflare’s quarterly count. Incidents in Honduras and on Namecheap’s network show why traffic declines need careful interpretation.
SOURCE / Cloudflare
Kubernetes announces workload conformance working group
The group plans version-specific runtime requirements and tooling, targeting a stable first specification and working tests within two to three Kubernetes release cycles.
SOURCE / Kubernetes
Cloudflare discloses GLM-5.3 cache exposure affecting 155 responses
Cloudflare’s October 26 report traces cross-customer data exposure to a cache-capacity mismatch and a SGLang rejection bug. Service was restored September 9.
SOURCE / Cloudflare
SageMaker AI adds serverless customization for Muse Glimmer, Qwen3.8 and Ministral 3
AWS adds more models to SageMaker AI’s serverless customization service, with SFT, DPO and RFT support and availability in four regions.
SOURCE / AWS
AWS Continuum for code vulnerabilities reaches general availability
AWS says Continuum for code vulnerabilities combines environment context, sandbox validation and remediation pull requests, with approval and autonomous modes.
SOURCE / AWS
AWS adds InvokeModel APIs and fallback to AgentCore Gateway
AgentCore Gateway adds Bedrock InvokeModel API support and ordered model fallback, available in supported AWS Regions outside China.
SOURCE / AWS
AWS previews Interconnect last mile with Megaport in two metros
AWS's Megaport preview covers San Jose and Northern Virginia, with automated network setup and dynamic 1–100 Gbps scaling planned for general availability.
SOURCE / AWS
AWS expands Lambda MicroVMs to four more regions
AWS added Hong Kong, Malaysia, Thailand and Spain on October 22, 2026, bringing Lambda MicroVMs availability to 14 AWS Regions.
SOURCE / AWS
Amazon MSK Connect adds in-place connector pause and resume
AWS adds pause and resume to MSK Connect, retaining connector configuration, capacity settings and committed offsets during temporary processing stops.
SOURCE / AWS
AWS lists October 2026 Corretto updates for six Java release lines
AWS lists Corretto updates for Java 27, 25, 21, 17, 11 and 8, available through downloads and Linux package repositories.
SOURCE / AWS
AWS cuts the EMR charge for EMR on EKS by 30%
The reduction applies automatically to EMR’s per-vCPU-hour and per-GB-hour charges across supported AWS Regions. Underlying EKS charges remain unchanged.
SOURCE / AWS
Amazon MWAA Provisioned expands to Mexico (Central)
AWS adds MWAA Provisioned environments in mx-central-1, where Serverless was already available, giving teams another regional deployment option.
SOURCE / AWS
AWS Bedrock Web Search adds domain filtering and location controls
AWS announced domain filters and optional location context for Bedrock Web Search, configured through the web_search tool in OpenAI Responses API requests.
SOURCE / AWS
Amazon Connect adds persistent chat across linked AWS Regions
AWS says persistent chat can retain conversation context across linked Regions. Adoption requires Global routing, and availability follows Customer Global Resiliency.
SOURCE / AWS
Amazon Quick becomes available in AWS’s Singapore region
AWS says Amazon Quick’s Singapore launch includes AI-powered chat, Research, Spaces and QuickSight dashboards, with data stored and processed in-region.
SOURCE / AWS
AWS adds force rollback for CloudFormation stack updates
AWS says CloudFormation’s force rollback option completes update rollback in one operation and records resources needing attention before the next deployment.
SOURCE / AWS
GitHub bug bounty spotlight: Trace data, prove impact
In GitHub’s interview, @r31n explains source-to-sink investigation, AI’s supporting role and the need for reproducible proof of security impact.
SOURCE / GitHub
AWS expands 100 Gbps Direct Connect availability near Bangalore
AWS has expanded 100 Gbps dedicated Direct Connect availability at the existing Netmagic DC2 location near Bangalore, India.
SOURCE / AWS
AWS Transform adds new network builds and security-group conversion
AWS Transform adds new network builds and conversion of on-premises access rules into security groups for existing VPCs, without redeploying those network stacks.
SOURCE / AWS
Cloudflare reports South Korea ISMS certification and sign-up changes
Cloudflare says Korean account creation now requires itemized consent and excludes social sign-up. Certification coverage follows a service boundary.
SOURCE / Cloudflare
Amazon Cognito extends multi-Region replication to AWS GovCloud (US)
Cognito’s replication add-on now supports AWS GovCloud (US), giving eligible user pools a standby copy of identity data for regional recovery.
SOURCE / AWS
AWS expands 10 Gbps and 100 Gbps Direct Connect options near Manchester
AWS has expanded 10 Gbps and 100 Gbps dedicated connections with MACsec capabilities at the existing Equinix MA3 location near Manchester.
SOURCE / AWS
Amazon Connect Customer adds Modern and Classic agent workspace layouts
AWS introduces two agent workspace layouts with team-specific choices for contact controls and application space, available in nine AWS Regions.
SOURCE / AWS
Amazon Bedrock adds structured JSON outputs for Claude
AWS announces schema-compliant JSON responses for Claude on both Bedrock endpoint paths. Engineers still need to validate the content.
SOURCE / AWS
AWS Partner Central separates case-study publishing from specialization review
AWS Partner Central adds Program Navigator and a redesigned Case Study Generator, with separate review paths for go-to-market content and specialization applications.
SOURCE / AWS
Cloudflare introduces tsrs, an experimental Rust port of TypeScript
Cloudflare says tsrs is available experimentally through npm, with upstream compatibility a goal and early performance claims dependent on configuration.
SOURCE / Cloudflare
Google projects US$368 billion in Southeast Asia digital-economy GMV for 2026
Google’s e-Conomy SEA summary forecasts 17% annual growth. Reported interest in AI shopping agents raises questions about consent, reversibility and capacity planning.
SOURCE / Google
AWS adds online migration to ElastiCache Serverless for Valkey
Supported EC2-hosted Valkey and Redis OSS clusters can migrate directly to serverless Valkey 9.0, with manual or automatic cutover and separate migration billing.
SOURCE / AWS
Cloudflare reports 68% lower AI code-review costs—with quality trade-offs
Cloudflare reports $0.77 per review after cache, orchestration and model changes, but its evaluation found fewer important performance issues at medium effort.
SOURCE / Cloudflare
Cloudflare’s NetScaler guidance pairs login controls with origin restrictions
Cloudflare outlines layered protection for NetScaler web authentication, with separate controls for desktop traffic and checks against direct-origin bypasses.
SOURCE / Cloudflare
Reflection’s Beam arrives in SageMaker JumpStart with Trainium3 optimization
Beam’s initial JumpStart rollout covers two US regions. AWS also reports Trainium3 optimization, but the announcement includes no quantified performance or cost results.
SOURCE / AWS
AWS surfaces conditional field requirements in messaging registration responses
AWS End User Messaging adds conditional-field details to registration API responses, compliance examples and an RCS video walkthrough.
SOURCE / AWS
AWS Transform adds Azure-to-AWS code conversion for .NET
AWS Transform's Azure-to-AWS .NET code conversion is generally available for functions, blob storage and queues. Cosmos DB conversion is in preview.
SOURCE / AWS
Amazon Braket enters FedRAMP Class C scope in three US regions
AWS says Braket’s AWS-operated components are covered in three US regions, while third-party quantum execution remains outside the boundary.
SOURCE / AWS
AWS End User Messaging adds reusable Brand Profiles for registrations
Brand Profiles reuse company information across messaging registrations, with optional AI field matching and a draft-review step before submission.
SOURCE / AWS
AWS adds message-sequencing wizard for RCS registration videos
AWS says its new RCS registration wizard sends test messages in sequence and is available on all registration forms requiring a video.
SOURCE / AWS
AWS brings Lambda durable functions to both US GovCloud regions
AWS expands Lambda durable functions to GovCloud US-West and US-East, with checkpoint-and-wait workflows and a qualified compute-charge exemption during pauses.
SOURCE / AWS
AWS adds NVIDIA Nemotron 3.5 Super VL to SageMaker JumpStart
AWS lists NVFP4 inference and BF16 customization options for NVIDIA’s vision-language model, which accepts images, video and text.
SOURCE / AWS
CloudWatch log centralization expands to GovCloud and 13 more commercial Regions
AWS brings CloudWatch log centralization to both GovCloud Regions and 13 additional commercial Regions, with collection scoped through AWS Organizations.
SOURCE / AWS
AWS adds managed OTP generation and verification to Notify
AWS End User Messaging adds configurable OTP generation and verification to Notify, with delivery over text, voice and WhatsApp.
SOURCE / AWS
AWS adds Mexico (Central) support for second-generation Outposts racks
AWS says customers can now order second-generation Outposts racks connected to Mexico (Central), expanding regional choices for hybrid deployments.
SOURCE / AWS
Cloudflare details internal trust evaluations for Verified bots
Cloudflare is assessing whether Verified bots respect robots.txt preferences and rate-limit feedback. Customer trust thresholds and public scores remain planned.
SOURCE / Cloudflare
AWS makes Lambda Durable Execution SDK for Go generally available
The Go SDK combines checkpointed workflows and approval callbacks with offline testing tools and deployment on Lambda’s provided.al2023 runtime.
SOURCE / AWS
AgentCore Web Search adds Ohio, Oregon, Sydney and GovCloud (US-West)
AWS expands Amazon Bedrock AgentCore Web Search to four more Regions, including GovCloud (US-West), widening placement options for web-grounded agents.
SOURCE / AWS
AWS Security Hub CSPM adds CRI Profile v2.2 support
AWS says 404 Security Hub CSPM controls support a subset of CRI Profile v2.2 requirements, excluding those that need manual checks.
SOURCE / AWS
Amazon S3 arrives on Vercel Marketplace and v0
AWS says developers can provision S3 general purpose buckets, S3 Tables and S3 Vectors through Vercel, with account linking and new-account credits.
SOURCE / AWS
AWS Backup adds FSx support in Asia Pacific (New Zealand)
AWS Backup now supports four Amazon FSx services in the New Zealand Region, with policy-based scheduling, retention rules and encrypted backup vaults.
SOURCE / AWS
Amazon ECS adds LiveTaskCount and removes RunningTaskCount charge
The new ECS metric includes tasks still active during shutdown. AWS also removes the RunningTaskCount metric charge for Container Insights users, effective October 1.
SOURCE / AWS
AWS Agent Registry adds schema-validated custom metadata
Teams can define shared and resource-specific metadata fields, validate entries and filter registry searches through AWS interfaces.
SOURCE / AWS
AWS DataSync Enhanced mode adds OpenZFS, NFS 4.2 and metadata controls
DataSync Enhanced mode adds FSx for OpenZFS locations, NFS 4.2 support and control over copying object user metadata in regions where Enhanced mode is supported.
SOURCE / AWS
AWS Security Hub adds resource criticality to exposure prioritization
Security Hub can assign resource criticality through automation rules or infer it from tags and names, adding business context to exposure triage.
SOURCE / AWS
Amazon DCV 2026.0 adds RHEL 10 support and a new macOS client
The release expands Linux support, introduces a macOS client with gamepad support, and adds Windows redirection and client decoding features.
SOURCE / AWS
AWS launches Lambda Web Functions for direct HTTP handling
Lambda Web Functions launches with Node.js support in five AWS Regions, concurrent request handling, native HTTP streaming, and billing for active CPU time and retained memory.
SOURCE / AWS
AWS Transfer Family adds custom ports for public SFTP servers
AWS Transfer Family public SFTP servers now support up to three ports, each with a client-first or server-first communication mode.
SOURCE / AWS
AWS Security Hub adds on-demand S3 exports for findings
Security teams can export findings from Security Hub console pages to S3, choosing CSV for spreadsheets or OCSF-formatted JSON.
SOURCE / AWS
GitHub makes the case for hackathons as hands-on learning
Participant accounts highlight rapid prototyping and peer support, but a successful demo is not proof of productivity gains or production readiness.
SOURCE / GitHub
AWS brings EC2 R8g to European Sovereign Cloud in Germany
Graviton4-powered R8g instances are now available in AWS European Sovereign Cloud (Germany), expanding options for memory-intensive workloads.
SOURCE / AWS
AWS brings EC2 R8gd to European Sovereign Cloud in Germany
AWS adds R8gd instances to its European Sovereign Cloud in Germany, with Graviton4 processors, local NVMe storage and adjustable network and EBS bandwidth.
SOURCE / AWS
Cloudflare launches Clef-omni and cuts Clef-flash pricing—with a lower hosted context limit
Clef-omni adds audio and video inputs, while Cloudflare lowers Clef-flash pricing to $0.038 per million input tokens and its hosted context limit to 24k.
SOURCE / Cloudflare
Amazon Bedrock adds reasoning summaries for OpenAI models
AWS says Bedrock developers can request reasoning summaries for OpenAI models, with support for in-Region and cross-Region inference.
SOURCE / AWS
AWS adds automated checks for Amazon Connect Customer evaluation forms
The checks flag questions that lack context for AI scoring and suggest more specific criteria. AWS lists availability in eight Regions.
SOURCE / AWS
Amazon Quick adds centrally managed brand templates
AWS says brand templates are generally available in the Quick desktop app, with administrator controls over who can create and publish them.
SOURCE / AWS
AWS announces S3 Vectors metadata pre-filtering for GovCloud
AWS announces metadata pre-filtering in both GovCloud regions, with defaults for indexes in new vector buckets and an in-place upgrade for existing indexes.
SOURCE / AWS
Amazon GameLift Servers adds opt-in CPU bursting for container fleets
AWS says GameLift container fleets can be packed around typical CPU demand and use idle instance CPU for spikes, with no additional feature charge.
SOURCE / AWS
Cloudflare details evidence-first agents for Managed Defense early beta
Cloudflare’s Managed Defense agent harness collects evidence before model analysis, with analysts responsible for decisions and mitigation.
SOURCE / Cloudflare
Kubernetes outlines cgroup v1 upgrade risks and planned v1.38 removal
Guidance explains default kubelet startup failures from v1.35, the temporary cgroup v1 fallback and checks for migrating Linux nodes to cgroup v2.
SOURCE / Kubernetes
GitHub outlines Git infrastructure rebuild for agent-heavy workloads
GitHub’s proposed architecture separates durable storage from compute. The company reports up to 35× higher write throughput in internal benchmarks.
SOURCE / GitHub
AWS Batch adds native CloudWatch job lifecycle metrics
AWS Batch automatically publishes job-transition and duration metrics to CloudWatch, with queue-level visibility across all Regions where Batch is available.
SOURCE / AWS
AWS Client VPN adds device posture checks before and during sessions
AWS Client VPN can check device compliance before access and disconnect noncompliant sessions, with monitoring-only mode available to assess policy impact.
SOURCE / AWS
Google’s September AI recap highlights Argon’s phased rollout
Google reports Fairwind program access for Argon and Flash Cyber, while its new Live audio models have rolled out through developer channels.
SOURCE / Google
Amazon EKS and EKS Distro add Kubernetes 1.37 support
AWS announces Kubernetes 1.37 support for EKS and EKS Distro, with generally available Metrics API and DRA device controls and beta HPA scale-to-zero.
SOURCE / AWS
Cloudflare opens Logpush and Logpush Transformers to all plans
Logpush expands to Free, Pro and Business through usage-based pricing. Transformers becomes generally available, while Organizations and spending caps have different rollout stages.
SOURCE / Cloudflare
ElastiCache for Valkey adds OpenTelemetry metrics and optional 15-second publishing
AWS adds attribute-based PromQL queries and optional 15-second metric publishing for node-based Valkey clusters, where CloudWatch supports OpenTelemetry metrics.
SOURCE / AWS
AWS MCP Server adds six hosting Regions
AWS MCP Server now runs in eight Regions. Its hosting footprint is separate from its ability to access services across all commercial AWS Regions.
SOURCE / AWS
Cloudflare announces OHTTP Gateway closed beta
The planned paid zone add-on handles OHTTP cryptography for application backends, with a third-party relay required to preserve separation of trust.
SOURCE / Cloudflare
EmDash 1.0 launches with a decentralized plugin registry
Cloudflare announces EmDash 1.0 with publisher-signed plugin records and permission-controlled sandboxing. The registry supports free plugins; paid support remains a goal.
SOURCE / Cloudflare
Cloudflare adds per-domain TLS key exchange visibility
New analytics and log fields expose negotiated TLS key exchange groups, helping engineers assess post-quantum encryption on visitor and origin connections.
SOURCE / Cloudflare
AWS DataSync adds shared-VPC support for centralized private endpoints
Accounts sharing a subnet can now use one centrally managed DataSync VPC endpoint instead of separate endpoints for private connectivity.
SOURCE / AWS
AWS previews Well-Architected Agent for infrastructure reviews
AWS says the preview reviews infrastructure-as-code templates and prioritizes recommendations by business goals, with access in three US Regions.
SOURCE / AWS
AWS adds unified graphics images to WorkSpaces Applications
One image type can serve supported GPU families, with cross-family fleet changes available through the console or UpdateFleet API.
SOURCE / AWS
AWS adds organization-wide sharing for Systems Manager documents
Systems Manager documents can now be shared with organizations or OUs through AWS RAM, with sharing updated as account membership changes.
SOURCE / AWS
AWS sets service sunset dates and new-customer restrictions
AWS’s September 29 update sets an October 29 cutoff for new customers of two offerings, lists four sunset milestones and ends Mechanical Turk availability.
SOURCE / AWS
AWS adds start-date postponement for future-dated EC2 Capacity Reservations
Customers can postpone future-dated EC2 Capacity Reservations before delivery by accepting a revised quote, which may require a longer commitment.
SOURCE / AWS
AWS adds Claude Sonnet 5.5 to GovCloud through Bedrock
Claude Sonnet 5.5 is available through Amazon Bedrock in AWS GovCloud (US), adding a model option for coding and document workflows.
SOURCE / AWS
GitHub reports 24 Android vulnerabilities found with AI-guided audits
GitHub Security Lab describes Android-focused auditing workflows, two disclosed vulnerability examples, and the need for expert triage and premium model requests.
SOURCE / GitHub
AWS raises EMR Serverless shuffle limit to 1 TB per job
Serverless Storage increases its shuffle allowance from 200 GB and adds spill support, with availability on specified EMR releases in 18 AWS Regions.
SOURCE / AWS
Route 53 DNS Firewall’s Palo Alto Networks integration reaches GA in 32 Regions
AWS announced general availability for Palo Alto Networks Advanced DNS Security in Route 53 Resolver DNS Firewall, with cross-account license sharing and VPC deployment options.
SOURCE / AWS
Amazon DocumentDB adds retryable writes in engine 8.0.2
DocumentDB 8.0.2 supports one automatic driver retry for eligible writes on instance-based and serverless clusters. Older engines still require retryable writes to be disabled.
SOURCE / AWS
AWS Backup extends logically air-gapped vaults to FSx for NetApp ONTAP
FSx for NetApp ONTAP backups can now use AWS Backup’s logically air-gapped vaults, with immutable backups and cross-account direct restores.
SOURCE / AWS
Cloudflare says Impact programs will reach $100 million in donated services
Cloudflare’s projected milestone covers donated services, alongside existing crawler controls and planned nonprofit and human-rights project announcements.
SOURCE / Cloudflare
AWS PCS adds compute-node scaling logs with opt-in delivery
AWS PCS scaling logs capture individual compute-node state transitions, with opt-in delivery to CloudWatch Logs, S3 and Data Firehose in all PCS Regions.
SOURCE / AWS
Kinesis adds service-managed partition keys for on-demand streams
AWS adds opt-in record distribution without producer-supplied partition keys for Kinesis On-Demand Standard and On-Demand Advantage streams.
SOURCE / AWS
GitHub details an issue-driven workflow for marketing events
A GitHub practitioner account connects Copilot planning, structured issues and Actions automation, with code review, dry-run controls and a monitoring lesson.
SOURCE / GitHub
AWS expands EC2 X2idn availability to Hong Kong
Memory-optimized EC2 X2idn instances are now available in AWS’s Asia Pacific (Hong Kong) Region, expanding regional deployment options.
SOURCE / AWS
AWS MediaLive adds frame-accurate pipeline locking without source timecode
AWS says MediaLive’s Video Aligned Locking synchronizes streams without source timecode, including linked cross-region single-pipeline channels.
SOURCE / AWS
SageMaker HyperPod adds model caching for faster inference scale-out
AWS says HyperPod model caching is generally available across supported regions, preloading model weights and container images to reduce inference cold starts.
SOURCE / AWS
Amazon Bedrock adds Marengo 3.0 embeddings for media search
AWS says Marengo 3.0 brings video, audio and image embeddings to Bedrock Managed Knowledge Base, with 512-dimensional vectors and segment timestamps.
SOURCE / AWS
AWS HealthOmics adds 14 real-time run metrics in CloudWatch
AWS HealthOmics adds resource-utilization metrics for running bioinformatics workflows in nine Regions, with CloudWatch charges based on metric ingestion volume.
SOURCE / AWS
AWS DevOps Agent adds two-way Slack communication
Engineers can start and guide AWS DevOps Agent investigations in connected private Slack channels, with context, findings and recommendations captured in one thread.
SOURCE / AWS
MediaLive adds A/B watermarking for live-video leak tracing
MediaLive can produce paired, watermarked live-video outputs for downstream packaging and CDN systems to assemble into session-specific sequences for leak tracing.
SOURCE / AWS
AWS adds Yield Optimization to MediaTailor livestream ads
AWS says MediaTailor’s Yield Optimization can fill unused livestream ad inventory with Amazon Ads demand for APS Streaming TV publishers.
SOURCE / AWS
Amazon RDS adds CU and GDR builds for five SQL Server versions
AWS lists new supported RDS builds for SQL Server 2016 through 2025, with GDR updates addressing four vulnerabilities and upgrades available through the console, SDK or CLI.
SOURCE / Amazon Web Services
AgentCore Memory adds direct ingestion without short-term events
AWS’s IngestData API sends conversational or JSON content to long-term memory extraction without first creating a short-term memory event.
SOURCE / Amazon Web Services
Bedrock adds native ServiceNow ingestion with sys ID scoping
AWS says the connector ingests ServiceNow knowledge articles, catalog items and attachments, with inclusion lists to narrow crawls and automatic incremental sync.
SOURCE / AWS
GitHub opens HydraFusion research preview in Copilot CLI
GitHub’s multi-model CLI preview introduces quality-cost trade-offs that engineers should evaluate on their own repository tasks.
SOURCE / GitHub Blog
Rust debugging survey highlights value display and setup gaps
Rust’s first debugging survey highlights value-inspection problems and setup friction, suggesting separate checks for onboarding and debugger usefulness.
SOURCE / Rust Blog
Bedrock adds scheduled syncs for native knowledge-base connectors
Daily, weekly and monthly connector syncs could replace simple RAG ingestion schedulers, but teams still need to account for freshness and failures.
SOURCE / AWS
Elastic 9.5 makes native Prometheus and PromQL support generally available
Elastic 9.5 reaches GA with native Prometheus ingestion and PromQL support, alongside columnar storage and security-operations changes.
SOURCE / Elastic
Grafana Cloud Adaptive Profiles reaches general availability
Adaptive Profiles is now GA in Grafana Cloud, dynamically changing profiling detail and frequency and increasing resolution when anomalies or performance issues appear.
SOURCE / Grafana Labs
OpenBao 2.6 adds namespace sealing and cross-plugin workflows
OpenBao 2.6 introduces per-namespace cryptographic sealing, external auto-unseal plugins, cross-plugin workflows, and distroless images.
SOURCE / OpenSSF / OpenBao
BOMHort joins the OpenSSF Sandbox
The Kubernetes-native BOMHort project joins OpenSSF, combining SBOM ingestion, OSV refreshes, VEX handling, license policy, and ClickHouse analytics.
SOURCE / OpenSSF
Prometheus 3.5.5 patches a dashboard sanitization flaw
Prometheus 3.5.5 updates sanitize-html to address CVE-2026-53606, giving operators a focused security upgrade on the 3.5 LTS line.
SOURCE / Prometheus
Sentry expands Logs with pinning, JSONL export, and large-query handling
Sentry’s summer Logs update adds shareable pinned rows, aggregate-aware exports, JSONL, faster views, and continuation support for large searches.
SOURCE / Sentry
Sentry links OpenTelemetry spans to SDK-captured errors
Sentry’s OtlpIntegration connects existing OpenTelemetry trace context with Sentry errors across several backend SDKs, with explicit limits.
SOURCE / Sentry
Sigstore keeps Rekor v1 as its public default
Sigstore changed its Rekor rollout plan: the public-good service will keep v1 as default while modern clients can opt into production Rekor v2.
SOURCE / Sigstore
Trivy 0.72 broadens detection and VEX handling
Trivy 0.72 adds Bottlerocket matching, bundled .NET runtime detection, richer Java license discovery, repository VEX loading, and safer temp paths.
SOURCE / Aqua Security / Trivy
Chrome 151 adds declarative camera controls and SPA performance entries
Chrome 151 introduces the usermedia capability element, manual declarative shadow slots, and route-level performance signals for single-page apps.
SOURCE / Chrome for Developers
Falco 0.44 removes legacy engines and hardens runtime detection
Falco 0.44 drops legacy eBPF, gVisor, and gRPC paths while adding stricter rules, bounded captures, faster process lookup, and new detections.
SOURCE / Falco
Falco Operator 0.2.0 reaches production-ready status
Falco Operator 0.2.0 adds Kubernetes-style reconciliation, status conditions, finalizers, and a Component resource for managing the wider stack.
SOURCE / Falco
Grafana 13 makes dynamic dashboards, Git Sync, and Advisor broadly available
Grafana 13 turns several major dashboard and operations features on for general use, giving upgrade teams concrete schema, Git, and health-check changes to test.
SOURCE / Grafana Labs
Pyroscope 2.0 reworks continuous profiling storage
Pyroscope 2.0 makes object storage authoritative, removes write-path replication, and moves profiling queries onto a stateless read path.
SOURCE / Grafana Labs
Next.js 16.3 adds disk-backed builds and opt-in instant navigation tools
Next.js 16.3 lowers development memory use, reuses build artifacts, and previews partial prefetching, navigation inspection, and offline retries.
SOURCE / Next.js
OpenTelemetry separates adoption blueprints from real-world reference implementations
A new OpenTelemetry initiative pairs reusable implementation guidance with snapshots of how real organizations operate telemetry systems, without promising turnkey configurations.
SOURCE / OpenTelemetry
OpenTelemetry graduates within the CNCF
CNCF graduation recognizes OpenTelemetry’s project maturity and governance while the maintainers continue work on interoperable telemetry standards.
SOURCE / OpenTelemetry
OpenTelemetry Demo 3.0 introduces breaking schema changes and new signals
OpenTelemetry Demo 3.0 renames custom attributes, restructures Compose deployment, and adds agentic, profiling, OpAMP, load-generation, and test capabilities.
SOURCE / OpenTelemetry
OpenTelemetry launches a work-in-progress instrumentation explorer
OpenTelemetry’s new explorer makes Java agent instrumentation metadata searchable, while clearly presenting Collector and other ecosystem coverage as unfinished work.
SOURCE / OpenTelemetry
OpenTelemetry Go Logs API and SDK enter release candidate
The Go Logs API and SDK reached release-candidate status for direct application use and logging bridges, while exporters and logtest remain experimental.
SOURCE / OpenTelemetry
OpenTelemetry opens its first Linux packaging repository
The Packaging SIG published a first repository and one-command Linux metapackage, with explicit warnings that signing and production readiness remain unfinished.
SOURCE / OpenTelemetry
OpenTelemetry Profiles reaches public alpha
OpenTelemetry’s profiling signal entered public alpha, giving implementers a shared data model and OTLP path while keeping stability expectations explicit.
SOURCE / OpenTelemetry
Vue 3.5.42 fixes async hydration, Suspense, and v-model edge cases
Vue 3.5.42 is a maintenance release focused on async hydration teardown, pending Suspense branches, select synchronization, and SSR input safety.
SOURCE / Vue.js
Astro 7 moves its compiler and Markdown pipeline to Rust
Astro 7 combines a Rust compiler and content pipeline with queued rendering, stable route caching, advanced routing, and Vite 8.
SOURCE / Astro
Chrome 147 ships scoped view transitions and new CSS color and border tools
Chrome 147 brings concurrent element-level view transitions, automatic black-or-white contrast selection, and shape-aware border painting.
SOURCE / Chrome for Developers
Chrome 148 adds lazy media loading and a browser Prompt API
Chrome 148 can defer audio and video near the viewport, query named containers without a type, and run constrained multimodal prompts on device.
SOURCE / Chrome for Developers
CSS Working Group publishes June CSS Snapshot 2026
The June CSS Snapshot classifies stable, candidate, and rough-interoperability modules while warning that specification status is not browser support.
SOURCE / W3C
Firefox 149 ships hint popovers and math font support
Firefox 149 adds the hint popover state, math font defaults, and platform changes that affect layered overlays and mathematical content.
SOURCE / Mozilla MDN
Firefox 151 adds style queries and document picture-in-picture
Firefox 151 ships CSS container style queries, declarative shadow slot assignment, and always-on-top HTML windows on desktop.
SOURCE / Mozilla MDN
Firefox 153 adds bulk IndexedDB reads, video PiP, and JSPI
Firefox 153 ships getAllRecords for IndexedDB, desktop video picture-in-picture, WebAssembly Promise integration, and richer WebRTC diagnostics.
SOURCE / Mozilla MDN
Next.js 16.2 improves startup, rendering, and hydration debugging
Next.js 16.2 pairs Turbopack and rendering work with clearer server-function logs, hydration diffs, error causes, and stable adapters.
SOURCE / Next.js
React 19.2.8 improves Server Component decoding performance
React 19.2.8 is a focused patch that changes Server Component payload decoding, giving framework teams a small but important upgrade target.
SOURCE / React
Safari 26.4 adds Grid Lanes, WebTransport, and Keyboard Lock
Safari 26.4 delivers 44 web features and 191 fixes, including masonry-style Grid Lanes, HTTP/3 WebTransport, and fullscreen keyboard control.
SOURCE / WebKit
Safari 26.5 adds :open, Origin API support, and 63 WebKit fixes
Safari 26.5 expands state styling and origin parsing while repairing IndexedDB, storage-access, WebRTC, scrolling, and accessibility failures.
SOURCE / WebKit
Safari 27 beta previews customizable selects and broader web-platform updates
WebKit’s Safari 27 beta adds customizable form controls, scroll anchoring controls, JSPI, and a large set of CSS, media, and Web API changes.
SOURCE / WebKit
Svelte 5.56 adds runes-mode declaration tags inside templates
Svelte 5.56 adds runes-mode template declarations, with explicit state and derived runes for values that must update reactively.
SOURCE / Svelte
Vite 8.1 previews bundled development for very large applications
Vite 8.1 adds experimental bundled dev and chunk import maps, direct Wasm ESM imports, and configurable asset discovery for custom HTML.
SOURCE / Vite
Vite 8 replaces its dual bundlers with Rolldown
Vite 8 makes the Rust-based Rolldown bundler its unified development and production pipeline, with compatibility shims and migration guidance.
SOURCE / Vite
Astro 6 adds built-in fonts, CSP, and live content collections
Astro 6 combines new content and security APIs with a refactored development pipeline designed to run the target production runtime locally.
SOURCE / Astro
Cloud Native Buildpacks graduates at the CNCF
CNCF graduated Cloud Native Buildpacks, recognizing the source-to-OCI image toolkit’s production adoption, governance, and maintained build workflow.
SOURCE / Cloud Native Computing Foundation
CSS Working Group updates gap decorations draft for grid and flex
The March draft renames inset controls and adds edge, intersection, and empty-area behavior for rules drawn through grid and flex gaps.
SOURCE / W3C CSS Working Group
Docker Offload moves the container engine into a managed cloud session
Docker Offload reached general availability, letting Docker Desktop users keep local commands while the engine runs in an isolated managed environment.
SOURCE / Docker
Docker VMM enters public beta on Mac and Windows
Docker rebuilt the virtualization layer under Docker Desktop and opened it as a public beta in version 4.86 for macOS and Windows users.
SOURCE / Docker
K8gb advances to CNCF incubation for Kubernetes global routing
CNCF moved K8gb to incubation, recognizing a Kubernetes-native global load-balancing project built around DNS, CoreDNS, ExternalDNS, and custom resources.
SOURCE / Cloud Native Computing Foundation
Kubeflow graduates at the CNCF as a Kubernetes AI platform
CNCF graduated Kubeflow, recognizing the Kubernetes-native project’s maturity across data preparation, training, fine-tuning, inference, and model serving.
SOURCE / Cloud Native Computing Foundation
Kubernetes 1.37 delivers 67 enhancements in the Garhwal release
Kubernetes 1.37 ships 67 enhancements, including a stable storage-version migration API and a broad mix of stable, beta, and alpha changes.
SOURCE / Kubernetes
Podman Desktop 1.26 adds extension credential permissions
Podman Desktop 1.26 prompts before extensions access authentication sessions and adds managed controls for extension installation and discovery.
SOURCE / Podman Desktop
Pulumi adds Terraform state, native modules, and first-class HCL
Pulumi’s August release makes its Terraform backend and HCL language generally available while allowing Pulumi programs to consume Terraform modules directly.
SOURCE / Pulumi
Terraform 1.16 adds module imports and machine-readable state commands
Terraform 1.16 supports import blocks inside modules, JSON output for state and workspace inspection, richer action triggers, and lifecycle destroy controls.
SOURCE / HashiCorp on GitHub
containerd 2.3 becomes the project’s first annual LTS release
containerd 2.3 aligns minor releases with Kubernetes, starts a two-year LTS line, and expands CRI, EROFS, sandbox, and NRI capabilities.
SOURCE / containerd project on GitHub
Docker makes Gordon generally available for container workflows
Docker’s Gordon agent is generally available in Docker Desktop and the CLI, with local container context and configurable approval controls for actions.
SOURCE / Docker
etcd 3.7 streams large ranges and boots directly from v3store
etcd 3.7 adds RangeStream for chunked reads, optimizes keys-only queries, and boots from v3store while retaining v2 snapshots for compatibility.
SOURCE / Kubernetes / SIG etcd
Gateway API 1.6 stabilizes portable TCP and UDP routing
Gateway API 1.6 moves TCPRoute and UDPRoute into the stable v1 API and separates future experimental resources into an explicitly different group.
SOURCE / Kubernetes / SIG Network
Apache Iceberg Rust 0.10 revises table lifecycle and retention APIs
Iceberg Rust 0.10 adds snapshot-expiration work and separates table purging from catalog removal, with breaking dependency and API changes to review.
SOURCE / Apache Iceberg
Ingress2Gateway 1.0 turns Ingress migration into a reviewable conversion
Kubernetes SIG Network released Ingress2Gateway 1.0 with broader Ingress-NGINX annotation coverage and integration-tested Gateway API output.
SOURCE / Kubernetes
Kubernetes 1.36 ships 70 enhancements under the Haru release
Kubernetes 1.36 arrived with 70 enhancements, including resource-health reporting, workload scheduling changes, and several removals to audit.
SOURCE / Kubernetes
Kubernetes 1.36 adds beta in-place resizing for pod-level resources
Kubernetes 1.36 can change pod-level CPU and memory allocations in place, with explicit status conditions and runtime prerequisites for operators.
SOURCE / Kubernetes
Kubernetes 1.36 moves controller sharding into the API server
An alpha Kubernetes API lets horizontally sharded controllers request only their hash range for list and watch traffic instead of discarding events.
SOURCE / Kubernetes
OpenTofu 1.12 makes lifecycle policy and CLI output more adaptable
OpenTofu 1.12 adds dynamic prevent_destroy, complete provider checksums, parallel installation, dual output streams, and a state-only removal control.
SOURCE / OpenTofu
Podman 6.0 removes legacy hosts, databases, and network stacks
Podman 6.0 adds cross-provider machine management and safer volume pruning while requiring migrations from several retired platform components.
SOURCE / Podman project on GitHub
PostgreSQL 19 Beta 1 previews replica-wait and maintenance features
PostgreSQL 19's first beta introduces replica replay waits and maintenance changes, opening a testing window for application and driver compatibility.
SOURCE / PostgreSQL Global Development Group
Pulumi previews direct resource operations with pulumi do
The research-preview pulumi do command offers stateless CRUD and query operations through provider schemas without first creating a project.
SOURCE / Pulumi
ClickHouse 26.5 pushes eligible top-N work below joins
ClickHouse 26.5 optimizes eligible ORDER BY and LIMIT queries across joins; result cardinality and actual plans are the right upgrade checks.
SOURCE / ClickHouse
ClickHouse 26.7 adds measured execution plans with EXPLAIN ANALYZE
ClickHouse 26.7 adds EXPLAIN ANALYZE with planning, execution and parallelism measurements, making expensive diagnostics worth scheduling deliberately.
SOURCE / ClickHouse
DuckDB opens version 2.0 alpha for compatibility testing
DuckDB's September alpha starts compatibility testing for its next major release, with different client version labels and extension availability to track.
SOURCE / DuckDB
Kafka 4.3 adds log-directory cordoning and storage metrics
Kafka 4.3 adds controls for log-directory placement and retention visibility, alongside deprecations that operators should map to existing dashboards.
SOURCE / Apache Kafka
Kafka 4.3.1 fixes a Kafka Streams RocksDB native memory leak
Kafka 4.3.1’s highlighted fix concerns native memory in Streams applications, so verification should include process memory and recovery behavior.
SOURCE / Apache Kafka
MySQL 9.7.2 fixes deletion ordering in JSON duality-view updates
MySQL 9.7.2 corrects nested-row deletion during JSON duality-view updates and date conversion behavior, making targeted regression cases valuable.
SOURCE / Oracle
PostgreSQL’s August update calls for targeted index and statistics checks
PostgreSQL 18.6 and companion releases fix 28 security issues, with post-update checks for GIN statistics, btree_gist and large ltree indexes.
SOURCE / PostgreSQL Global Development Group
PostgreSQL’s May update fixes 11 security issues and replication bugs
PostgreSQL’s May maintenance releases include security fixes and standby repairs; teams should inventory client tools alongside database servers.
SOURCE / PostgreSQL Global Development Group
Redis 8.8 adds arrays and explicit Streams message rejection
Redis 8.8 introduces arrays, window counters and XNACK; consumer reclaim settings remain essential to making rejected messages available again.
SOURCE / Redis
SQLite 3.53 adds constraint changes and repairs expression-index compatibility
SQLite 3.53 adds ALTER TABLE constraint operations and result formatting, while floating-point text changes deserve explicit compatibility tests.
SOURCE / SQLite
SQLite 3.53.4 follows a wave of AI-assisted bug reports
SQLite’s July maintenance release fixes more reported bugs; application teams need to verify the embedded runtime they actually distribute.
SOURCE / SQLite
Valkey 9.1 adds database-specific access controls and TLS diagnostics
Valkey 9.1 introduces access rules for numbered databases, certificate monitoring and script controls, giving operators more precise deployment checks.
SOURCE / Valkey
Valkey Admin 1.0 unifies cluster inspection and troubleshooting
Valkey Admin 1.0 adds topology, key browsing and command logs, with server-version prerequisites and web-interface authentication to plan.
SOURCE / Valkey
DuckDB 1.5 adds VARIANT and makes GEOMETRY a core type
DuckDB 1.5 expands semi-structured and spatial data support, while a date_trunc return-type change gives existing pipelines a concrete upgrade check.
SOURCE / DuckDB
DuckLake 1.0 stabilizes its SQL-based lakehouse specification
DuckLake 1.0 pairs a stable format with its DuckDB extension, making catalog recovery and small-write behavior key evaluation points for data teams.
SOURCE / DuckLake
Go 1.27 adds generic methods and new JSON packages
Go 1.27 expands generic programming and JSON processing, while a generally available goroutine-leak profile gives maintainers another diagnostic tool.
SOURCE / Go
Go launches a beta API for package and module metadata
Go's new pkg.go.dev API exposes structured package and module information, with explicit version selection and ambiguity handling for tooling authors.
SOURCE / Go
LLVM 23.1 changes compiler infrastructure and removes bugpoint
LLVM 23.1 arrives with changes to intermediate representation and internal APIs, including a replacement path for workflows that still use bugpoint.
SOURCE / LLVM
MongoDB 8.3 expands query expressions and database fundamentals
MongoDB 8.3 adds type-coercion and string-expression capabilities alongside database improvements; representative query and upgrade fixtures remain essential.
SOURCE / MongoDB
MSTest 4.4 enables native ahead-of-time test executables
Microsoft describes MSTest 4.4's source-generated Native AOT test path, giving teams another way to examine behavior under their deployment compilation model.
SOURCE / Microsoft
MySQL 9.7.0 brings JSON duality view writes to Community Server
MySQL 9.7.0 expands Community Server JSON duality views to inserts, updates and deletes, with schema and application round trips worth testing together.
SOURCE / Oracle MySQL
Node.js 24.20 adds test-runner logging to its LTS line
Node.js 24.20 expands test-runner events with contextual logging and entry-file information, alongside additions to asynchronous context and streams.
SOURCE / Node.js
Python 3.13.15 delivers another maintenance update for the 3.13 line
Python 3.13.15 bundles bug fixes, build improvements, and documentation updates, giving teams on the older feature line a distinct maintenance decision.
SOURCE / Python Software Foundation
Python 3.15 reaches its first release candidate and freezes the ABI
Python 3.15's first release candidate marks an ABI stability milestone and asks package maintainers to publish wheels ahead of the final release.
SOURCE / Python Software Foundation
Rust 1.98.1 fixes incorrect trait-object vtable generation
Rust 1.98.1 corrects a compiler bug that could emit null function pointers in trait-object vtables, making build provenance important for remediation.
SOURCE / Rust
Valkey Search 1.2 adds text filters and server-side aggregations
Valkey Search 1.2 combines text, tags, numeric ranges and vectors; primary routing and shard boundaries matter when testing read-after-write behavior.
SOURCE / Valkey
Deno 2.8 adds dedicated CI installation and npm package-building commands
Deno 2.8 introduces explicit commands for frozen CI installs and npm-ready package artifacts, with a runtime compatibility boundary for Deno-specific APIs.
SOURCE / Deno
Deno 2.9 introduces experimental desktop application packaging
Deno 2.9 adds an experimental desktop command and imports existing package-manager lockfiles, giving teams two distinct migration paths to evaluate.
SOURCE / Deno
.NET 11 Preview 3 updates archive validation and text handling
.NET 11 Preview 3 brings ZIP checksum validation, broader newline recognition, and SDK changes that give maintainers targeted compatibility checks.
SOURCE / Microsoft
.NET 11 Preview 5 adds JSON Lines serialization and full outer joins
.NET 11 Preview 5 expands data-processing APIs with JSON Lines serialization and LINQ full outer joins, alongside runtime and SDK improvements.
SOURCE / Microsoft
Go explains how package authors can automate API migrations
Go's source-level inliner lets package maintainers describe migrations through wrappers, with behavior-preserving analysis for call arguments.
SOURCE / Go
Java 26 ships HTTP/3 support and broader ahead-of-time object caching
Java 26 reaches general availability with HTTP/3 in its client API, garbage-collector-neutral object caching, and several explicitly preview features.
SOURCE / Oracle
Node.js 26 enables Temporal by default in its Current release
Node.js 26 ships Temporal, a newer V8 engine, and removals of legacy APIs, giving maintainers a concrete compatibility target before its planned LTS phase.
SOURCE / Node.js
npm allows multiple trusted-publishing configurations per package
Maintainers can authorize several independent OIDC workflows, keep each staging-only by default, and wait for malware scanning before approval.
SOURCE / GitHub
Python 3.15 enters beta and asks maintainers to test compatibility
Python 3.15's first beta starts wider compatibility testing, with lazy imports and UTF-8 defaults among the changes package maintainers should examine.
SOURCE / Python Software Foundation
Rust 1.95 adds configuration selection and if-let match guards
Rust 1.95 introduces a standard configuration-selection macro and richer match guards, with a compatibility change for custom target specifications.
SOURCE / Rust
Rust 1.96 separates range values from their iterators
Rust 1.96 stabilizes new range types that can be copied, while existing range syntax keeps its meaning and WebAssembly linking becomes stricter.
SOURCE / Rust
Rust 1.97 changes symbol names and exposes linker diagnostics
Rust 1.97 makes v0 symbol mangling the default and gives Cargo control over warning failures, alongside newly visible linker messages.
SOURCE / Rust
Rustup 1.29 downloads toolchain components concurrently
Rustup 1.29 overlaps component downloads and unpacking, adds Solaris hosts, and changes how update checks signal that a newer toolchain exists.
SOURCE / Rust
GitHub Actions begins holding suspicious workflows for approval
GitHub now pauses potentially malicious workflow runs in public repositories until a collaborator with write access reviews them in the web interface.
SOURCE / GitHub
GitHub hardens checkout defaults for pull_request_target workflows
Actions checkout v7 will refuse common patterns that expose privileged pull_request_target jobs to untrusted pull-request code.
SOURCE / GitHub
GitHub Actions adds commit-pinned $/ references for local actions
The new self-repository syntax resolves actions and reusable workflows at the caller’s exact commit without a separate checkout step.
SOURCE / GitHub
GitHub Desktop 3.6 brings worktrees into the desktop client
The desktop release lets developers manage parallel working trees and deepens Copilot support for commit messages and conflict resolution.
SOURCE / GitHub
GitHub Enterprise Importer adds self-service GitLab migrations
The gh gl2gh extension now supports single and bulk moves from GitLab.com and maintained self-managed versions into GitHub Enterprise Cloud.
SOURCE / GitHub
GitLab 18.10 expands agentic review and security triage
The monthly release opens Duo Agent Platform access to GitLab.com Free groups and makes SAST false-positive assessment generally available.
SOURCE / GitLab
GitLab 19.0 adds secrets management and broader merge-request automation
The major release introduces a Secrets Manager beta, CI component analytics, and agent-assisted work across the merge-request lifecycle.
SOURCE / GitLab
GitLab 19.2 makes Duo CLI and custom flows generally available
The release brings governed agent workflows to the terminal, adds scheduled pipeline policies, and expands automated security and dependency remediation.
SOURCE / GitLab
GitLab 19.3 adds flow creation and bulk vulnerability remediation
The release turns plain-language workflow descriptions into reviewable flow definitions and lets teams apply AI-assisted triage to groups of SAST findings.
SOURCE / GitLab
npm outlines secure-by-default install changes for version 12
npm 12 will require projects to opt into dependency install scripts and non-registry sources, with warnings available in npm 11.16 for migration testing.
SOURCE / GitHub
Copilot browser tools in VS Code reach general availability
VS Code agents can drive isolated browser sessions for web testing, while tab sharing, sensitive permissions, and enterprise domain policy remain user-controlled.
SOURCE / GitHub
Cloudflare launches Billable Usage API for self-serve accounts
A single billing endpoint returns daily product usage and cost rows with field names aligned to much of the FinOps FOCUS specification.
SOURCE / Cloudflare
Cloudflare adds response-phase rules for cache decisions
Cache Response Rules can rewrite cache directives, manage cache tags, or strip selected headers after origin response and before storage.
SOURCE / Cloudflare
Cloudflare cuts 1.1.1.1 cache memory by about 100 terabytes
Five Rust data-layout changes reduce Big Pineapple’s per-entry footprint by 56% while improving measured cache insertion and lookup performance.
SOURCE / Cloudflare
GitHub Actions custom hosted-runner images reach general availability
Organizations can standardize GitHub-hosted runner environments with preinstalled tools and certificates while keeping existing preview images intact.
SOURCE / GitHub
GitHub Actions Runner Controller 0.14.0 adds multilabel scale sets
ARC 0.14.0 lets Kubernetes runner fleets express several scheduling attributes on one scale set and exposes a reusable service client.
SOURCE / GitHub
GitHub adds organization-wide firewall policy for Copilot cloud agent
Organization administrators can set cloud-agent network defaults and allowlists across repositories while retaining repository-level delegation where needed.
SOURCE / GitHub
GitHub opens Copilot SDK public preview in five languages
The preview exposes Copilot’s agent runtime for application builders, including tools, streaming, approvals, tracing, attachments, and BYOK.
SOURCE / GitHub
GitHub previews live migrations from GHES to data-resident cloud
Enterprise Live Migrations continuously synchronizes active repositories before a short cutover, targeting large repositories that cannot tolerate long freezes.
SOURCE / GitHub
GitHub releases REST API version 2026-03-10
A new dated REST API contract gives integrators a controlled migration point, with breaking changes isolated behind an explicit version header.
SOURCE / GitHub
GitHub schedules SHA-1 TLS shutdown for September 15
GitHub’s HTTPS endpoints will stop accepting SHA-1-based TLS connections after a July brownout, giving legacy clients a defined migration deadline.
SOURCE / GitHub
npm staged publishing adds human approval before package release
Maintainers can send a tarball from CI to a stage queue, then require a 2FA-authenticated approval before the version becomes installable.
SOURCE / GitHub
Cloudflare adds dollar-denominated spend limits to AI Gateway
AI Gateway can now cap spending by model, provider, or custom attributes, with optional fallback routing when a budget is exhausted.
SOURCE / Cloudflare
Cloudflare traces billing delays to ClickHouse query-plan contention
A growing part list exposed an exclusive-lock bottleneck in ClickHouse planning; Cloudflare’s fixes improved concurrency and part pruning.
SOURCE / Cloudflare
Cloudflare completes Code Orange resilience program after 2025 outages
Cloudflare says its Fail Small work now stages configuration, narrows failure impact, and expands emergency access and incident communications.
SOURCE / Cloudflare
Cloudflare cuts core server recovery from four hours to three minutes
UEFI and iPXE automation changes removed repeated network-interface probes and made firmware configuration survive upgrades without manual BIOS work.
SOURCE / Cloudflare
Cloudflare introduces Flagship, an OpenFeature-based flag service
Flagship evaluates feature flags locally at Cloudflare’s edge, using Durable Objects for configuration and Workers KV for distribution.
SOURCE / Cloudflare
Cloudflare Research introduces Meerkat consensus experiment
Meerkat applies the leaderless QuePaxa algorithm to small, strongly consistent control-plane state across Cloudflare’s global network.
SOURCE / Cloudflare
Cloudflare launches Precursor for session-level bot signals
Precursor injects a first-party script to evaluate behavior across a web session, complementing Turnstile’s checks at sensitive moments.
SOURCE / Cloudflare
Cloudflare scales Security Insights beyond 120 scans per second
Kafka flow separation, batched Postgres writes, and API changes raise scanning capacity more than tenfold without adding hardware.
SOURCE / Cloudflare
Cloudflare details Town Lake data platform and its Skipper AI agent
Town Lake gives Cloudflare a governed SQL layer across disparate stores, while Skipper turns natural-language questions into auditable queries.
SOURCE / Cloudflare
VoidZero and the Vite toolchain team join Cloudflare
Cloudflare brings in the team behind Vite, Vitest, Rolldown, Oxc, and Vite+, while committing to their open-source, vendor-neutral governance.
SOURCE / Cloudflare
Azure Deployment Stacks What-If reaches GA
Stack updates can now produce a persistent preview showing create, modify, delete, and detach actions at resource, subscription, or management-group scope.
SOURCE / Microsoft Azure Governance and Management Blog
Azure Firewall explicit proxy reaches GA
Selected applications and browsers can send HTTP and HTTPS egress to Azure Firewall through standard proxy settings without rerouting a whole subnet.
SOURCE / Microsoft Azure Networking Blog
Azure Monitor previews container-level advanced platform metrics
Azure Storage is the first service to expose paid advanced metrics, adding per-container blob capacity and object-count signals.
SOURCE / Microsoft Azure Observability Blog
Azure Virtual Network routing appliance reaches GA
The managed hardware-backed router provides up to 200 Gbps tiers, dual-stack forwarding, zone resiliency, and built-in Azure Monitor metrics.
SOURCE / Microsoft Azure Networking Blog
Cloudflare makes AI Security for Apps generally available
AI Security for Apps reaches GA with endpoint discovery, incoming-prompt inspection, topic controls, and WAF actions for detected threats.
SOURCE / Cloudflare
Cloudflare introduces stateful API vulnerability scanning in beta
Cloudflare’s beta scanner begins with BOLA detection, deriving stateful tests from an OpenAPI call graph and separate owner and attacker contexts.
SOURCE / Cloudflare
Cloudflare introduces Custom Regions for customer-defined data boundaries
Custom Regions let enterprises define geographic processing boundaries, while Cloudflare expands its catalog of predefined regional controls.
SOURCE / Cloudflare
Cloudflare launches Gen 13 edge servers with Turin CPUs and FL2
Cloudflare’s Gen 13 design pairs 192-core AMD Turin processors with its Rust-based FL2 request layer to double edge throughput in testing.
SOURCE / Cloudflare
Cloudflare introduces Mesh for identity-aware private networking
Cloudflare Mesh connects users, servers, agents, and Workers through identity-based policies without requiring a centralized network topology.
SOURCE / Cloudflare
Cloudflare previews Project Think for durable, long-running agents
Project Think adds durable execution, subagents, persistent sessions, and sandboxed code execution to Cloudflare’s Agents SDK.
SOURCE / Cloudflare
Cloudflare unifies third-party AI models behind the AI.run binding
Cloudflare’s unified inference layer lets Workers call hosted and third-party models through one binding, catalog, and AI Gateway control plane.
SOURCE / Cloudflare
Azure Bastion previews managed identity for session recordings
Bastion Premium can write graphical session recordings to Blob Storage with a system- or user-assigned identity instead of a storage credential.
SOURCE / Microsoft Azure Network Security Blog
Azure Blob SDK client-side CRC64 integrity reaches GA
Selected Blob SDKs can validate CRC64-NVME checksums across whole uploads and downloads as well as individual transfer requests.
SOURCE / Microsoft Azure Storage Blog
Azure Boost previews Guest RDMA in UK South
Guest VMs can use RDMA through Azure Boost for low-latency, high-throughput communication, beginning with preview availability in UK South.
SOURCE / Microsoft Azure Compute Blog
Azure Files SMB managed-identity authentication reaches GA
Applications can access Azure Files over SMB with an Azure managed identity, removing stored account keys and passwords from the application path.
SOURCE / Microsoft Azure Storage Blog
Azure Integrated HSM reaches GA on selected AMD v7 VMs
The same-node hardware cache and crypto offload is generally available for Windows Trusted Launch guests on selected AMD D- and E-series v7 VMs with at least 8 vCores.
SOURCE / Microsoft Azure Compute Blog
Azure’s Intel Xeon 6 Dlsv7, Dsv7, and Esv7 VMs reach GA
The seventh-generation general-purpose and memory-optimized families add new CPU, memory, networking, and optional local-storage configurations.
SOURCE / Microsoft Azure Compute Blog
Azure previews Lasv5 and Laosv5 local-storage VMs
The AMD EPYC Turin-based VM families increase local NVMe capacity, network bandwidth, size range, and processor performance for storage-heavy workloads.
SOURCE / Microsoft Azure Compute Blog
Azure Machine Configuration custom security baselines reach GA
Teams can customize and deploy reusable Windows and Linux baseline artifacts through policy and infrastructure delivery workflows.
SOURCE / Microsoft Azure Governance and Management Blog
Azure previews in-place regional-to-zonal VM migration
Regional VMs and flexible scale-set instances can move into an availability zone while retaining resource identity, disks, NICs, and addresses.
SOURCE / Microsoft Azure Compute Blog
VM Scale Sets Flex previews automatic OS image upgrades
Flexible orchestration scale sets can now roll instances to newer OS image versions with health-aware, staged replacement controls.
SOURCE / Microsoft Azure Compute Blog
Azure Web Application Firewall Default Ruleset 2.2 reaches GA
The new managed ruleset adopts OWASP CRS 3.3.4 across Application Gateway and Front Door, with an upgrade caveat for custom exclusions.
SOURCE / Microsoft Azure Network Security Blog
Azure Compute API will always return a VM security type
A new Compute API version removes nullable securityType responses, requiring clients to treat Standard as an explicit value rather than an omission.
SOURCE / Microsoft Azure Compute Blog
Azure Container Registry makes geo-replication failover health-aware
ACR’s global endpoint can steer around replicas with deeper push and pull failures instead of relying only on shallow health.
SOURCE / Microsoft Apps on Azure Blog
Azure Container Registry adds proactive Service Health alerts
Registry incidents can now trigger Azure Service Health notifications based on push, pull, and authentication service indicators.
SOURCE / Microsoft Apps on Azure Blog
Azure previews full caching for Ephemeral OS disks
A new placement mode keeps the complete operating-system disk on VM-local storage, reducing steady-state dependence on remote disks.
SOURCE / Microsoft Azure Compute Blog
Azure previews subscription-wide Essential Machine Management
A subscription enrollment layer can apply a core operations set to present and future Azure VMs and Arc-enabled servers at scale.
SOURCE / Microsoft Azure Governance and Management Blog
AlloyDB makes its managed remote MCP server generally available
AlloyDB’s managed remote MCP endpoint reaches general availability, making agent access to operational databases a platform permissions decision.
SOURCE / Google Cloud
BigQuery introduces built-in identity columns for generated keys
BigQuery identity columns move numerical key generation into the warehouse, giving data teams a new choice when designing ingestion pipelines.
SOURCE / Google Cloud
Cloud SQL autoscaling read pools reach general availability
Enterprise Plus read pools can scale MySQL and PostgreSQL replicas behind one endpoint, but applications still need a clear read-consistency policy.
SOURCE / Google Cloud
Google Cloud previews smaller GPU slices in G4 virtual machines
Fractional G4 virtual machines introduce smaller NVIDIA GPU allocations, giving graphics and compute teams another way to size accelerator capacity.
SOURCE / Google Cloud
Cloud Storage Rapid adds cache ingestion during writes
Google Cloud describes Rapid Cache ingest on write, targeting the first-read delay in workloads that consume newly written objects immediately.
SOURCE / Google Cloud
Storage Transfer Service adds Amazon S3 sources in AWS GovCloud
Google Cloud adds AWS GovCloud source-region support to Storage Transfer Service, covering both batch jobs and event-driven S3 transfers.
SOURCE / Google Cloud
Google Backup and DR adds cross-region backups for compute, disks, and Filestore
Cross-region backup destinations become generally available for selected Google Cloud resources, giving recovery plans an explicit regional choice.
SOURCE / Google Cloud
BigQuery makes managed Python user-defined functions generally available
Managed Python UDFs reach production support in BigQuery, combining SQL invocation with serverless execution and configurable function resources.
SOURCE / Google Cloud
Cloud Network Insights reaches general availability for hybrid network observability
Google’s network observability service uses active probes across hybrid and multicloud paths, joining network measurements with application experience.
SOURCE / Google Cloud
Cloud Run adds direct Identity-Aware Proxy integration
Google makes direct IAP protection generally available on Cloud Run, reducing the infrastructure needed to restrict access to internal applications.
SOURCE / Google Cloud
Cloud Run worker pools become generally available for pull-based processing
Worker pools give Cloud Run a continuously running execution option for queue consumers, while scaling still needs an explicit workload signal.
SOURCE / Google Cloud
Google Cloud previews Fault Injection Testing for SQL failover and traffic faults
The preview offers reusable experiments for Cloud SQL failover and load-balancer faults, with a dry run that shows permissions and affected resources.
SOURCE / Google Cloud
GKE previews ClusterNetworkPolicy for tiered cluster-wide network rules
ClusterNetworkPolicy enters preview in GKE 1.36 and later, separating administrator rules, namespace policies, and cluster-wide baseline defaults.
SOURCE / Google Cloud
Google announces GKE DRA availability and contributes its TPU driver
Google’s March announcement brings Dynamic Resource Allocation to GKE and contributes a TPU driver, expanding how workloads describe hardware needs.
SOURCE / Google Cloud
Spanner Omni preview brings a downloadable Spanner deployment option
Google previews Spanner Omni for infrastructure outside its managed service, with a developer edition explicitly limited to non-production use.
SOURCE / Google Cloud
Google makes Lightning Engine generally available across managed Spark modes
Lightning Engine reaches both serverless and managed Spark clusters, adding native execution while retaining fallback for unsupported operators.
SOURCE / Google Cloud
AWS releases aws-bench in research preview
The open-source aws-bench preview pairs cloud tasks with known resource states, giving agent developers a repeatable evaluation starting point.
SOURCE / AWS
Graviton5-based EC2 R9g and R9gd instances reach general availability
AWS launches memory-optimized R9g and R9gd instances, giving database and cache teams new Arm capacity to evaluate against real workload bottlenecks.
SOURCE / AWS
ECS adds agent-connectivity health events and managed recovery
ECS exposes impaired agent connectivity across compute options, with automatic recovery for Fargate and Managed Instances and events for EC2 operators.
SOURCE / AWS
Lambda adds full IAM resource-based policy support
Lambda expands resource policies to support multiple principals, actions, and IAM conditions, giving platform teams a broader policy surface to review.
SOURCE / AWS
Lambda adds self-managed S3 code storage
Lambda can reference deployment code in customer-owned S3 storage, making artifact retention and bucket permissions part of the release contract.
SOURCE / AWS
Security Hub adds an organization-wide AI asset inventory
Security Hub combines managed-resource, software, and DNS signals to discover AI assets, giving security teams a new inventory to reconcile with owners.
SOURCE / AWS
Google Cloud makes C4N network and storage optimized VMs generally available
C4N reaches general availability with stronger network and block-storage limits, aimed at workloads that need more I/O without more CPU.
SOURCE / Google Cloud
Cloud Run previews singleton instances for continuously running workloads
Cloud Run instances enter preview with a dedicated singleton runtime and stable HTTPS endpoint, adding an option for small persistent workers.
SOURCE / Google Cloud
Filestore adopts Colossus and separates IOPS from capacity
Google Cloud’s Filestore backend update separates performance from capacity, giving teams a new sizing choice for shared NFS storage on GKE.
SOURCE / Google Cloud
Google and Anyscale introduce experimental gVisor sandboxes for Ray
An experimental Ray library makes sandboxes schedulable actors, bringing resource placement and isolated code execution into one runtime.
SOURCE / Google Cloud
Aurora Serverless platform version 4 changes scaling behavior
Aurora Serverless adds platform version 4 with scaling improvements, making workload transitions and platform-version tracking important upgrade checks.
SOURCE / AWS
Compute Optimizer expands EC2 and RDS instance coverage
Compute Optimizer adds newer EC2 and RDS options to recommendations, giving teams more candidates to validate against workload and migration constraints.
SOURCE / AWS
ECS on Fargate adds 32-vCPU task configurations
Fargate adds larger ECS task sizes for Linux workloads, giving teams more room per container while changing scaling and recovery tradeoffs.
SOURCE / AWS
ECS adds pause points to service deployments
ECS deployment pause hooks let teams add approval and validation stages, with explicit timeout behavior becoming an important release-policy decision.
SOURCE / AWS
AWS releases ExtendDB 0.1 with a PostgreSQL storage backend
ExtendDB brings the DynamoDB API model to pluggable storage, giving teams a new local and disconnected deployment option to test for compatibility.
SOURCE / AWS
AWS introduces Lambda MicroVMs for user and AI-generated code
Lambda MicroVMs adds a managed execution option with suspend and resume, requiring clear policies for session ownership, network access, and cleanup.
SOURCE / AWS
Network Firewall adds EKS and ECS attribute-based rules
Network Firewall can select container traffic using workload attributes, shifting attention from changing IP addresses to who controls those attributes.
SOURCE / AWS
Resilience Hub adds user-journey modeling and failure analysis
The next Resilience Hub release models systems, journeys, and services, giving SRE teams a richer assessment structure to connect with recovery tests.
SOURCE / AWS
Ai2 releases BenchMIRT to audit what benchmarks measure
BenchMIRT applies item response theory to model evaluations, helping researchers inspect which abilities benchmark questions actually distinguish.
SOURCE / Ai2 via Hugging Face
AWS Interconnect multicloud reaches general availability with Google Cloud
AWS makes Interconnect multicloud generally available with Google Cloud, offering a managed private connection to assess against real traffic needs.
SOURCE / AWS
Lambda Managed Instances adds Rust support
Rust functions can now run on Lambda Managed Instances, making concurrency behavior and workload measurement central to adoption decisions.
SOURCE / AWS
Lambda adds S3 Files mounts for shared workspaces
Lambda functions can mount S3 buckets through S3 Files, opening a shared workspace option that still needs explicit ownership and concurrency rules.
SOURCE / AWS
RDS Proxy joins Blue/Green Deployment switchovers
RDS Blue/Green Deployments can now redirect connections through RDS Proxy, putting application recovery behavior back into the database rehearsal plan.
SOURCE / AWS
Research and Engineering Studio 2026.03 expands desktop administration
The RES update adds filesystem and session controls, giving research platform operators more direct ways to recover and configure virtual desktops.
SOURCE / AWS
S3 adds account regional namespaces for predictable bucket naming
S3 introduces reserved bucket namespaces, giving platform teams a new way to create predictable names without competing for global availability.
SOURCE / AWS
Hugging Face introduces Funes for portable coding-agent memory
Funes indexes coding-agent histories locally and can share them through a Hugging Face dataset, preserving source turns behind retrieved decisions.
SOURCE / Hugging Face
Hugging Face and Voice Arena add Hindi and Indian English ASR tests
New Monsoon evaluation sets add Hindi and Indian English coverage, with speaker metadata that helps expose differences hidden by aggregate scores.
SOURCE / Hugging Face and Voice Arena
H Company releases NeoMME multilingual multimodal encoders
H Company’s NeoMME release combines text and image processing in one encoder, with Apache-licensed checkpoints and retrieval-oriented variants.
SOURCE / H Company via Hugging Face
Hugging Face releases a library of WebGPU kernels
Hugging Face’s new kernel library packages browser GPU building blocks with tests and benchmarks, giving local AI developers a reusable starting point.
SOURCE / Hugging Face
Meta publishes Brain2Qwerty v2 research and training code
Brain2Qwerty v2 studies real-time sentence decoding during typing, with training code released and important limits on what the experiment shows.
SOURCE / Meta AI
Meta and WRI release Canopy Height Maps v2
Meta and WRI’s forest mapping release pairs a revised model with global maps, giving geospatial teams a new baseline to validate locally.
SOURCE / Meta AI
Meta introduces Muse Spark with a limited API preview
Meta’s Muse Spark launch brings a new multimodal reasoning model to its consumer products, while API access begins as a private preview.
SOURCE / Meta AI
Meta updates SAM to 3.1 for faster video tracking
SAM 3.1 adds multiplexed object processing and global reasoning, with new artifacts for teams evaluating video detection and tracking throughput.
SOURCE / Meta AI
Meta releases TRIBE v2 for predicting brain responses
TRIBE v2 expands Meta’s research on stimulus-driven brain-response prediction, with released artifacts and a noncommercial license boundary.
SOURCE / Meta AI
Mistral Medium 3.5 powers remote coding agents in Vibe
Mistral’s model and remote-agent update moves coding sessions beyond the laptop, bringing sandbox access and final review into deployment planning.
SOURCE / Mistral AI
Mistral releases Search Toolkit for composable retrieval pipelines
The open-source preview joins ingestion, retrieval, and evaluation, giving teams a common place to compare search changes before altering generation.
SOURCE / Mistral AI
Microsoft clarifies the scope of its long-horizon AI delegation study
A May clarification explains what DELEGATE-52 measures and why artifact-fidelity stress tests should not be read as general product verdicts.
SOURCE / Microsoft Research
Leanstral 1.5 expands Mistral’s open proof-engineering model
The update releases Apache-licensed weights and the FLTEval suite, with longer proof attempts that make evaluation budgets part of the result.
SOURCE / Mistral AI
Mistral releases Leanstral for repository-scale proof engineering
Leanstral targets Lean 4 work with downloadable Apache-licensed weights, shifting evaluation toward verified repository changes and proof review.
SOURCE / Mistral AI
Mistral OCR 4 adds locations, block types, and confidence scores
OCR 4 returns more than extracted text, giving document pipelines spatial and confidence signals that need their own validation before use.
SOURCE / Mistral AI
Mistral introduces single-camera navigation model Robostral
Robostral Navigate targets instruction-following with RGB input; robotics teams must separate navigation-model results from full-system safety.
SOURCE / Mistral AI
Mistral releases Shieldstral with policy-defined safety questions
Shieldstral turns text and image moderation into policy-specific classification; teams still need local threshold calibration and review rules.
SOURCE / Mistral AI
Mistral Small 4 combines reasoning, vision, and coding in one model
Mistral’s Apache-licensed release brings several workloads into one model; deployment teams still need separate quality and capacity checks.
SOURCE / Mistral AI
Mistral adds speech generation with Voxtral TTS
The new voice model supports multilingual generation and voice adaptation; released weights have different usage terms from the hosted service.
SOURCE / Mistral AI
Mistral puts Workflows into public preview
Mistral’s orchestration preview pairs a hosted control plane with customer-run workers, making failure recovery and approval handling key pilot questions.
SOURCE / Mistral AI
Microsoft Research introduces Vega for private credential proofs
Vega proposes reusable zero-knowledge proofs of credential facts, with local proving and an implementation that was not yet open sourced at announcement.
SOURCE / Microsoft Research
Microsoft opens a waitlist for Copilot Health
Copilot Health’s March announcement describes a dedicated health-information experience, initially aimed at English-speaking adults in the United States.
SOURCE / Microsoft AI
Data Formulator 0.7 adds persistent enterprise data connections
Microsoft Research’s analytics release combines reusable data connectors, context-aware agents and a workspace for branching visual analysis.
SOURCE / Microsoft Research
Microsoft researchers connect AI brain explanations to experiments
A June research announcement describes generative causal testing, which turns model-based explanations of brain responses into testable stimuli.
SOURCE / Microsoft Research
Microsoft releases GridSFM-Open for power-grid research
GridSFM offers a learned approximation for AC optimal power flow, with an open research tier and a separate larger-system offering.
SOURCE / Microsoft Research
Microsoft Research releases MagenticLite and companion small models
MagenticLite combines browser and local-file workflows with MagenticBrain orchestration and Fara1.5 computer use in an experimental agent system.
SOURCE / Microsoft Research
Microsoft announces MAI-Image-2.6 with image-quality updates
The August announcement emphasizes text rendering and commercial imagery; a later page update separately records September preview availability.
SOURCE / Microsoft AI
Microsoft expands MatterSim with a multi-task materials model
MatterSim’s May update combines a new multi-task model, simulation improvements and experimental follow-up on a predicted thermal conductor.
SOURCE / Microsoft Research
SocialReasoning-Bench evaluates whether agents serve user interests
Microsoft Research’s benchmark separates task completion from outcome quality and decision process in calendar and marketplace negotiations.
SOURCE / Microsoft Research
Talos research describes continuous genomic reanalysis with expert review
Microsoft and collaborators report an open-source workflow that revisits genomic evidence and surfaces newly relevant variants for clinical teams.
SOURCE / Microsoft Research
MAI-Image-2.5 adds localized image editing and a Flash variant
Microsoft’s June image-model release adds editing controls and two model options, with product integrations spanning PowerPoint and OneDrive.
SOURCE / Microsoft AI
Microsoft adds an efficient MAI-Image-2 variant
The April release adds a second image-generation option aimed at faster, higher-volume workflows, with Foundry and playground access.
SOURCE / Microsoft AI
Microsoft introduces MAI-Image-2 for text-to-image generation
Microsoft’s March image-model release emphasizes visual detail and text rendering, with different availability paths for its playground and API.
SOURCE / Microsoft AI
Microsoft previews MAI-Image-2.5-Pro and MAI-Voice-2-Flash
Two July previews extend Microsoft’s image and speech model families, targeting detailed visual work and responsive voice experiences.
SOURCE / Microsoft AI
Microsoft opens MAI-Thinking-1 reasoning model in public preview
MAI-Thinking-1 adds an internally trained reasoning model to Microsoft’s lineup, with an enterprise preview and a sparse mixture-of-experts design.
SOURCE / Microsoft AI
MAI-Transcribe-1.5 expands language coverage and adds keyword biasing
Microsoft’s transcription update covers 43 languages and introduces domain vocabulary hints, while streaming and diarization remain roadmap items.
SOURCE / Microsoft AI
Microsoft launches MAI-Transcribe-1 multilingual speech recognition
MAI-Transcribe-1 entered public preview with support for 25 languages, giving teams another speech model to evaluate against real recordings.
SOURCE / Microsoft AI
MAI-Transcribe-2 adds speaker attribution and word timestamps
Microsoft’s September transcription release adds diarization, word timing and configurable transcript styles alongside broader language coverage.
SOURCE / Microsoft AI
Microsoft introduces MAI-Voice-2 with multilingual voice controls
MAI-Voice-2 adds multilingual speech generation, emotion tags and reference-audio prompting, creating new evaluation work for voice applications.
SOURCE / Microsoft AI
DeepMind published an AI Control Roadmap for agents inside Google
The framework addresses monitoring and containment for increasingly capable agents, including a live monitor used with Gemini Spark.
SOURCE / Google DeepMind
Gemini 3.5 Flash gained built-in computer use
Google integrated screen-based browser, mobile and desktop interaction into its main Flash model for custom automation agents.
SOURCE / Google
Google introduced Gemini 3.5 Transcribe for batch and live speech recognition
The speech-to-text family added transcription and live transcription models aimed at structured, context-aware audio processing.
SOURCE / Google
Google expanded Gemini's Flash lineup with 3.6, Flash-Lite and Cyber models
A July announcement paired two available Flash variants with a future, limited CodeMender pilot for a defensive-security model.
SOURCE / Google
Google released Gemini 3.7 Flash three weeks after 3.6
The new Flash iteration targeted coding, agents, web development and document work, continuing Google's rapid model release cadence.
SOURCE / Google
Gemini Omni 1.1 Flash added scene extension, keyframes and 4K output
Google's production-oriented video update gave developers longer scene continuation, endpoint frame control, draft previews and upscaling.
SOURCE / Google
DeepMind introduced Gemini Robotics 2 for whole-body robot control
The new robotics model targets locomotion, fine manipulation and transfer of learned skills across different robot bodies.
SOURCE / Google DeepMind
Google launched Gemini Robotics ER 2 for robot planning and coordination
The embodied-reasoning model uses video and spatial context to plan tasks, monitor progress and coordinate more than one robot.
SOURCE / Google
Google opened Nano Banana 2 Lite and Gemini Omni Flash to developers
The June release added a high-throughput image model and API access to Gemini's conversational video generation and editing model.
SOURCE / Google
Google DeepMind reported broader AlphaEvolve deployments across science and industry
A year after launch, DeepMind detailed how its Gemini-powered algorithm agent was being applied to genomics, energy and computing.
SOURCE / Google DeepMind
DeepMind opened Co-Scientist hypothesis generation to individual researchers
A Gemini-based multi-agent research system began rolling out through an experimental tool that generates, debates and ranks hypotheses.
SOURCE / Google DeepMind
DeepMind demonstrated a Gemini-powered pointer for contextual commands
Google DeepMind showed experimental interfaces that combine pointing and speech so Gemini can act on visible objects with less prompting.
SOURCE / Google DeepMind
Google released DiffusionGemma to test parallel text generation
The experimental open model generates blocks through diffusion rather than token by token, targeting interactive local workloads on GPUs.
SOURCE / Google
Google launched Gemini 3.5 Flash for agentic and coding workflows
The first Gemini 3.5 model became available across consumer, developer and enterprise surfaces with emphasis on long-running agent tasks.
SOURCE / Google
Gemini 3.5 Live Translate entered public preview for developers
Google released a streaming speech-to-speech model for more than 70 languages through the Gemini Live API and Google AI Studio.
SOURCE / Google
Gemini's April drop added a Mac app, Notebooks and interactive visuals
Google's April Gemini update bundled native macOS access, NotebookLM-backed project organization and interactive concept visuals.
SOURCE / Google
Google released Gemma 4 12B with native audio input for laptops
The new mid-sized open model added multimodal reasoning and native audio input while targeting machines with a 16 GB memory footprint.
SOURCE / Google
Anthropic launches Claude Tag as a shared Slack agent
The beta gives channel members a shared agent with selected tools and memory, plus administrator spend limits and an action log.
SOURCE / Anthropic
Anthropic maps AI-enabled cyber abuse against MITRE ATT&CK
An analysis of 832 banned accounts finds that existing technique counts can miss the risk added by autonomous orchestration and later-stage use.
SOURCE / Anthropic
Anthropic and Mozilla disclose Claude-assisted Firefox vulnerability work
Claude Opus 4.6 found 22 Firefox vulnerabilities, while Mozilla's triage showed why reproducible cases and coordinated disclosure remain essential.
SOURCE / Anthropic
Anthropic launches Claude Fable 5 and restricted Mythos 5
The same underlying model ships through separate safeguard envelopes, with Fable generally available and Mythos limited to trusted defenders.
SOURCE / Anthropic
Anthropic publishes its first Public Record survey results
The first wave surveyed nearly 52,000 Americans about hopes, fears, and regulation, creating a baseline whose methodology matters as much as totals.
SOURCE / Anthropic
Gemini API combined built-in tools with custom functions
A Gemini API update let developers mix Search or Maps grounding with their own functions and preserve context across tool calls.
SOURCE / Google
Gemini Embedding 2 reached general availability
Google moved its natively multimodal embedding model from preview to GA for text, image, video and audio retrieval workflows.
SOURCE / Google
Gemini added source-aware creation across Docs, Sheets, Slides and Drive
Google began beta rollouts that let Gemini draft and analyze Workspace content using selected files, email and web context.
SOURCE / Google
Anthropic discloses three cyber-evaluation boundary failures
Claude reached real internet systems during third-party capture-the-flag tests after the environment and prompt described a boundary that did not exist.
SOURCE / Anthropic
Anthropic proposes a severity framework for AI jailbreaks
The draft separates capability gained, breadth, reliability, and harm so model providers can discuss bypasses with more precision.
SOURCE / Anthropic
Anthropic launches Claude Science as an auditable research workbench
The beta combines agents, scientific connectors, reusable skills, local or remote compute, and artifact histories in one research environment.
SOURCE / Anthropic
Anthropic restores Claude Fable 5 access after export controls lift
Fable 5 returned globally after an abrupt suspension, exposing a new continuity risk for applications pinned to frontier model access.
SOURCE / Anthropic
Anthropic adds an Economic Index connector to Claude
The connector lets Claude query Anthropic's public usage dataset while the source's central limitation remains visible: it is not the labor market.
SOURCE / Anthropic
Anthropic narrows Claude Fable 5 biology safeguards
A retrained classifier allows more benign health and biology requests while dual-use professional research continues to fall back.
SOURCE / Anthropic
Anthropic releases Claude Opus 5 with effort controls and API fallbacks
The new Opus model targets daily coding and knowledge work, while beta controls let developers change tools and configure safeguard fallbacks.
SOURCE / Anthropic
Anthropic releases Claude Sonnet 5 for agentic workloads
Sonnet 5 adds effort controls and a revised tokenizer across Claude plans and the API, making migration tests more important than alias swaps.
SOURCE / Anthropic
Anthropic explains its planned text watermark for Claude
Future Claude models will encode a statistical signal in token choices, while short, factual, edited, and code outputs remain harder to detect.
SOURCE / Anthropic
Anthropic previews customer-controlled Enterprise Frontier Safeguards
EFS separates activity-data custody from Anthropic's automated misuse detection, with customer-owned storage and review controls.
SOURCE / Anthropic
Anthropic expands Claude access and credits for scientific research
A new scientist team plan and broader AI for Science credits extend Claude access while advanced biology remains access-controlled.
SOURCE / Anthropic
Anthropic funds open evaluations of AI's effects on wellbeing
A grant program will support independent, open-source evaluations built around multi-turn behavior and expert-validated graders.
SOURCE / Anthropic
Anthropic hardens agent evaluation environments after containment failures
Anthropic added real-time escape detection, stronger isolation, partner rules, and reinforcement-learning environment recertification.
SOURCE / Anthropic
Anthropic previews a hardware interface for AI-operated labs and factories
The Model Hardware Standard research preview connects agents to programmable instruments while safety practices remain under development.
SOURCE / Anthropic
Anthropic launches Claude Fable 5.1 and Mythos 5.1
The paired models add stronger coding and scientific work, revised safeguards, and separate general and trusted-access deployment paths.
SOURCE / Anthropic
OpenAI releases GeneBench-Pro for computational-biology judgment
The 129-task benchmark uses synthetic data with known causal structure to test iterative analysis and consequential decisions across ten biology domains.
SOURCE / OpenAI
OpenAI previews GPT-5.6 Sol, Terra, and Luna with new reasoning tiers
The limited API and Codex preview introduces a three-tier model family, max reasoning, ultra multi-agent mode, and layered cyber safeguards.
SOURCE / OpenAI
OpenAI launches full-duplex GPT-Live in ChatGPT Voice
GPT-Live-1 and GPT-Live-1 mini can listen and speak simultaneously, while delegating search and deeper reasoning to a frontier model.
SOURCE / OpenAI
OpenAI releases ChatGPT Images 2.0 with a reasoning mode
The new image model reaches all ChatGPT plans, while paid plans gain a mode that can reason, search the web, use tools, and refine outputs.
SOURCE / OpenAI
OpenAI rolls out a more scalable Dreaming memory system for ChatGPT
The new background memory architecture synthesizes changing user context, exposes a reviewable summary, and reduces compute for wider rollout.
SOURCE / OpenAI
GPT-5.5 Instant becomes ChatGPT's default model
OpenAI replaces GPT-5.3 Instant with a more concise default and adds visible memory sources for personalized responses across consumer plans.
SOURCE / OpenAI
OpenAI launches GPT-5.5 for agentic coding and computer-based work
GPT-5.5 arrived in ChatGPT and Codex with stronger long-horizon tool use, a 400K Codex context window, and expanded cyber safeguards.
SOURCE / OpenAI
OpenAI introduces GPT-Rosalind for life-sciences research
The purpose-built reasoning model enters trusted research preview with a Codex plugin connecting more than 50 scientific tools and data sources.
SOURCE / OpenAI
OpenAI introduces LifeSciBench for expert-judged research workflows
The benchmark tests evidence handling, analysis, experimental design, reasoning, validation, and communication across realistic life-science tasks.
SOURCE / OpenAI
OpenAI launches three realtime voice models in the API
GPT-Realtime-2, Realtime-Translate, and Realtime-Whisper add reasoning, live multilingual translation, and streaming transcription.
SOURCE / OpenAI
OpenAI model disproves a longstanding unit-distance conjecture
An internal general-purpose reasoning model found an infinite family of point configurations; external mathematicians checked the resulting proof.
SOURCE / OpenAI
OpenAI traces rare Rockset crashes to hardware and an 18-year-old libunwind race
Population-level core-dump analysis separated two similar crash signatures and exposed both a faulty host and a long-hidden unwinding race.
SOURCE / OpenAI
OpenAI details the inference and harness optimizations behind GPT-5.6
The engineering account spans request routing, GPU kernels, speculative decoding, cache-aware scheduling, and context management for agents.
SOURCE / OpenAI
OpenAI reveals GPT-Live's continuous inference and asynchronous delegation design
The six-month rebuild keeps audio on a dedicated streaming path while deeper reasoning, tools, compaction, and model handoffs run aside it.
SOURCE / OpenAI
OpenAI publishes first benchmark results for its Jalapeño inference chip
OpenAI tested its first custom inference accelerator across three public models and outlined a full-stack roadmap spanning chip, memory, network, and software.
SOURCE / OpenAI
OpenAI turns tax-agent corrections into evals and Codex tasks
A production loop captures practitioner corrections, builds targeted evaluations, and gives Codex scoped work to improve a tax-preparation agent.
SOURCE / OpenAI
OpenAI explains the native Windows sandbox built for Codex
The Codex team combined Windows identities, restricted tokens, filesystem ACLs, and network controls for useful local agent isolation.
SOURCE / OpenAI
OpenAI details a relay architecture for low-latency WebRTC voice
A stateless UDP relay routes sessions to stateful transceivers, shrinking the public port surface while preserving standard WebRTC behavior.
SOURCE / OpenAI
OpenAI releases MRC networking specification for large AI clusters
Multipath Reliable Connection spreads GPU transfers across redundant paths and uses source routing to reduce congestion and failure recovery delays.
SOURCE / OpenAI
OpenAI publishes Symphony specification for issue-driven Codex orchestration
The open-source specification maps issue-tracker work to isolated Codex agents, with explicit concurrency, retries, state, and human review.
SOURCE / OpenAI
No news matches your search.