The latest wire

Vercel's case study describes Tailscale's Aperture routing model requests through AI Gateway and running agents in Vercel Sandbox. Access follows tailnet identity instead of issuing each agent provider keys. Tailscale says this let it move from prototype to paying customers in months. The integration exposes request costs and can restrict routing to providers supporting zero data retention.

A three-person engineering team runs Featured's AI-assisted public-relations tools across three brands, according to Vercel's September 11 customer story. Featured migrated 374 Sanity sites from AWS Elastic Beanstalk and uses AI Gateway to route requests across 17 models. The company says its agents deliver more than 100,000 media pitches monthly, with Workflow SDK handling the long-running monitoring and qualification jobs behind them.

Pro and Enterprise teams gained tighter control over Vercel Connect in a September 11 update. Team owners can restrict connector creation and management to owners and members with the Connector Manager extended permission. Connectors give applications and agents access to external services through team-managed credentials; the restriction is enabled in the team's Connector Permissions settings.

In Kubernetes, resource allocation has historically been a static decision made during a Pod's initial scheduling and placement. With the graduation of the core in-Place Pod resize feature to General Availability in v1.35, application developers and cluster operators gained the powerful ability to dynamically adjust CPU and memory allocations of running containers without incurring disruptive restarts or application downtime. However, in-place resizing introduced a unique resource scheduling gap: if a running Pod requested a resource scale-up that exceeded the host node's allocatable headroom, the Kubelet was forced to mark the request as Deferred. The Pod would remain parked in this state indefinitely, waiting for resources on the node to naturally free up.

Vercel added an official GitHub Copilot adapter to the AI SDK harness layer. Applications can invoke Copilot through the same HarnessAgent interface used for other coding agents, reducing the application changes needed to switch agents. The adapter connects through the Agent Client Protocol using the harness ACP package.

Vercel's September 10 FastAPI update moves eligible frontend and StaticFiles content onto its CDN during builds, avoiding a function invocation for those requests. Route declaration order still decides whether an earlier application route takes precedence. Frontends protected by dependencies and static files behind middleware stay on the function by default, preserving checks that the CDN cannot perform.

Google highlights three ways its Search tools can support race preparation: creating a training schedule through AI Mode's Canvas, generating playlists after connecting a YouTube Music account, and comparing running equipment against personal requirements. The article describes shopping comparisons and local stock information as additional aids.

Mistral and Cloudera have announced a partnership to bring Mistral models into Cloudera's hybrid data platform. Mistral says the integration will support deployment in public and private clouds, on premises and in disconnected environments. The companies also aim to help enterprises build models using proprietary data while retaining control of their infrastructure, data and resulting models. The integration is planned.

On 9 September 2026, Check Point released emergency security updates addressing two critical vulnerabilities affecting Check Point Security Gateway, Security Management Server, and Spark Firewall deployments configured to use Remote Access VPN or Site-to-Site VPN. Both vulnerabilities carry a CVSS score of 9.8 and could allow an unauthenticated, remote attacker to execute arbitrary code on affected appliances. CERT-EU strongly recommends applying the available hotfixes as soon as possible, prioritising internet-facing and perimeter appliances.

Vercel introduced an integration for deploying long-running applications built with the OpenAI Agents API. OpenAI manages the agent loop and session state, while Vercel connects each session to an isolated Sandbox for execution and file access. The announcement describes persistent workspaces, signed webhooks and queues for reconnection, plus scaling to zero without a continuously running worker.

Le 06 août 2026, Metabase a publié un avis de sécurité concernant une vulnérabilité critique permettant à un attaquant non authentifié de provoquer une injection SQL (SQLi) dans la base de donnée de l'application Metabase.

Vercel's Tako Search promotion ended September 30; the announcement says subsequent searches are billed at standard rates. The integration lets Gateway models query Tako's curated data and the live web, with domain and date filters. It supports citations and visualizations without a separate Tako account or key, and works across Gateway models without rebuilding the search integration.

Vercel reported on September 10 that grouping CDN routing metadata into indexed files cut its measured 99th-percentile lookup time from 215.8 to 19.1 milliseconds. The roughly 200 KB groups warm multiple paths in one fetch while parsing only the requested record. These are Vercel's production measurements from August 5–12; deployments built after July 17 already use the format, while older ones need redeployment.

Vercel expanded Sandbox from four to all 20 compute regions in its September 10 announcement. Every plan can choose a region, helping teams place workloads near their data or within approved geographies. Ordered failover regions are limited to Pro and Enterprise. Existing sandboxes keep their original location, the default remains iad1, and Vercel says CPU and memory prices vary by region.

EFF highlights reporting that some police agencies use vague descriptions when documenting searches involving surveillance tools such as automated license plate readers. The organization argues that this can obstruct public oversight and scrutiny of evidence in court. Its article calls for clearer disclosure of surveillance practices and connects the issue to earlier disputes over investigative technologies.

Kubernetes has many ways to describe what is happening on a Node. Readiness, taints, Pod state, labels, annotations, and provider-specific APIs each expose part of the picture. What has been missing is a shared, Kubernetes-owned way to say that a Node is draining, undergoing maintenance, or undergoing Graceful Node Shutdown. Kubernetes v1.37 introduces five well-known Node conditions that provide that description: DrainInProgress Drained MaintenancePlanned MaintenanceInProgress GracefulNodeShutdownInProgress The new Node lifecycle conditions Condition What it reports DrainInProgress The Node is actively being drained according to the administrator's chosen drain criteria. Drained The Node has reached the drain criteria selected by the administrator. MaintenancePlanned The Node is expected to undergo a change in the future.

EFF examines the competing meanings of digital sovereignty, from government control of infrastructure to individuals' ability to choose and maintain their own tools. The authors argue that reducing dependence on dominant vendors requires more than moving data between countries. Their discussion emphasizes open systems, interoperability and local technical capacity as foundations for meaningful user control.

Google announced football upgrades for Search on September 9, combining live game timelines, highlights and AI insights with richer player statistics. The initial professional-football feed launched on mobile in U.S. English. Fans can also connect Yahoo Fantasy or Sleeper accounts to AI Mode for roster-specific suggestions, including lineup and waiver decisions; that account-linking feature also launched in U.S. English.

A September 9 account from Google DeepMind explains how the short documentary Love, Rendered recreated an elderly couple's first meeting, which had never been recorded. Engineers combined restored photographs with performance capture based on the couple's present-day mannerisms, while Ethelle Shatz guided visual details. The result is a filmmaker-directed reconstruction of their recollections, with AI used to create younger likenesses and movement.

On 8 September 2026, as part of its September Security Patch Day, SAP released Security Notes addressing two critical vulnerabilities affecting a broad range of SAP products[3]. The most severe, CVE-2026-44756 (CVSS 10.0), is a memory corruption vulnerability in SAP Extended Passport (EPP) processing, nicknamed "OVERPASS" by the Onapsis Research Labs (ORL), which discovered and responsibly disclosed it[3]. The second, CVE-2026-58240 (CVSS 9.8), nicknamed "S4GET", is a missing authentication check in the SAP NetWeaver Message Server[6]. Both are remotely exploitable without authentication. According to the reporting researchers, successful exploitation of either can result in arbitrary operating system command execution under the account that owns the SAP installation, leading to full compromise of the affected system and the business data it holds[6].

Mistral reports using AI agents to migrate 40,000 lines of a European energy operator's Fortran reservoir simulator into C++. The work combined numerical comparison tests, reconstructed documentation and human review of agent-generated changes. Mistral says fully autonomous attempts produced weaker results. The case study covers an initial portion of a 300,000-line system and cautions that projects without a working original program could be harder.

Vercel's September 9 update extends free Vercel Authentication protection to production deployments on every plan. Visitors must sign in with an account authorized for the project. Teams can apply protection to all deployments and make that setting the default for new projects. Deployment Protection Exceptions are also free, allowing selected preview domains to remain publicly accessible.

Vercel's September 9 pricing update lets Pro teams buy Password Protection for individual projects at $20 per project each month. Visitors must enter the team's chosen password before viewing protected deployments. The option previously required a $150 monthly team-wide add-on. Teams enable it in Deployment Protection settings and can disable it to stop future charges for that project.

Vercel added persistent memory to eve agents, using named slots that define storage providers and sharing scopes, such as individual authenticated users. Relevant memories enter the model context before each turn; update behavior depends on the provider. On Vercel, the built-in file provider uses private Blob storage to preserve memories across restarts and deployments, with additional and custom providers supported.