OpenShot 4.0 lands with AI masks and a recording dock as platform betas and .NET tooling land

OpenShot 4.0 lands with colour grading, local AI masks and a recording dock

OpenShot 4.0 is out, bringing colour grading, local AI masks and a recording dock. The release pairs traditional non-linear editing features with on-device AI masking, which removes the round-trip to a remote service for that workflow. A recording dock is bundled in, so capture and edit can live in one window. The move signals that creative open-source tooling is investing in capability parity with commercial editors while keeping inference local.

Apple ships visionOS 27.0 beta 8 as the platform cycle closes in

Apple released visionOS 27.0 beta 8 (24M5361a) on August 31, 2026. A beta with that ordinal in the cycle typically signals Apple is tightening regressions before a general-availability build. Developers on the visionOS track should re-run their spatial apps against the new build and re-check any RealityKit and SwiftUI surface they have touched since the prior beta. There is no published general-availability date in the evidence, so treat the next move as qualitative.

Telerik and Kendo UI trim AI token spend and extend Agentic UI Generator support

The Telerik and Kendo UI "What's New" notes list up to 45% lower token consumption across .NET AI experiences, expanded Agentic UI Generator support for ASP.NET Core, and additional AI-assisted product capabilities. For .NET teams wiring agents into existing applications, the headline change is a hard cap on how many tokens each AI-driven UI flow burns, which directly affects cost projections and rate-limit planning on hosted model APIs. ASP.NET Core projects that generate UI from prompts should be re-evaluated against the new Agentic UI Generator surface.

A Rust-based Linux server management tool and OpenAI Codex news

A new Linux server management tool written in Rust has been published alongside broader developer tooling coverage. Rust on the operations side continues to displace higher-overhead runtimes for system utilities where memory safety and predictable latency matter. The same roundup notes OpenAI Codex developments, which keep pressure on competitors to publish cadence and pricing for AI coding assistants. Practitioners evaluating server fleets should weigh a Rust-native management tool against incumbent shell-and-Ansible stacks.

What to check before the next build

Three concrete items belong on a working developer's checklist. 0 in a non-production project and exercise the AI mask workflow with a clip whose mask you can verify visually, since "local" only matters if the model still produces a usable alpha on real footage. 0 beta 8 (24M5361a) and rebuild any app that targets RealityKit, especially anything that depends on hand-tracking or scene-reconstruction APIs that tend to drift between betas.

NET AI flow before and after the update; the headline figure of up to 45% lower consumption is the kind of number that needs an empirical check on your own prompts. 0 in the evidence, and no version number for the Rust server tool beyond the language choice, so do not anchor plans on either. NET tracks in the same week, the common thread is the same: local inference and tighter AI cost profiles are both moving from keynote slides into shipping defaults, and any toolchain that cannot keep up with that pace is now the outlier.

If your build needs to verify visual outputs after an OpenShot export, an Audio Cutter helps trim the sample clips you feed into the mask test.