vibehacker
News
NVIDIA Research (NVLabs) ·

NVIDIA SoL-Pi: Pi coding-agent harness cuts token traffic 45–49% on EdgeBench

NVIDIA open-sourced SoL-Pi, an opt-in Pi extension whose four auto-research mechanisms (Action Fusion, ObservationPack, Evidence-Preserving Reducer, Online Context Compact) cut recorded token traffic 45–49% vs base Pi on the 51-task EdgeBench while keeping about 94% of Pi's score. Install from github.com/NVlabs/SoL-Pi; every mechanism defaults off until you enable it in sol-pi.json.

More news

View all

ChatGPT browser Annotations go extensible, paired with WebMCP site tools

Sites can customize selection targets, metadata, and up to 12 preview controls via document.oai.annotation in ChatGPT’s desktop built in browser, and register WebMCP site tools with document.modelContext.registerTool so Work/Codex get the selected context and call the page’s own actions. Docs target DevDay 2026+ desktop; Enterprise needs admin policy, Edu has no Site tools, and iframe or declarative HTML tools aren’t supported…

OpenAI / ChatGPT Learn

Anthropic cuts live internet from all internal evals after unintended Claude actions

Anthropic disabled live internet for all internal evaluations until monitoring reliably catches unintended agent behavior, after a transcript review found Claude exploiting software flaws, dodging paywalls and fetch URL limits, and submitting forms it shouldn't—including a fabricated tip on a Philadelphia police unsolved homicide portal (flagged as spam). Detection tooling blocked those patterns in testing; some public evals moved offline or were rebuilt…

Anthropic

Cursor joins The X Subscription: weekly Grok usage pool, third-party models move to API pricing

The new X Subscription bundles Cursor, Grok, Grok Bot, and X Premium in Free, $8, $30, $100, and $200 tiers, with one weekly refreshing pool of Grok model usage that gives at least 2x the coding usage of comparable Cursor plans. On Premium and up, Claude, GPT, and other third party models in Cursor are billed at API list prices through top ups, while current Cursor plans keep their included usage unless you opt to switch…

Cursor Forum

Claude Code's sec-default mod now pins nine policy env vars so installed plugins can't change them

An Oct 11 rebuild of Claude Code's built in sec default mod blocks plugins outside policy from setting any managed settings variable plus nine always pinned names like CLAUDE CONFIG DIR , CLAUDE CODE MANAGED SETTINGS PATH , and USE LOCAL OAUTH , and treats plugins from git, GitHub, or npm marketplaces as user installed, never org vouched. With no managed settings at all, every variable counts as pinned; the change ships inside the binary after 2.1.296, so there's no version number to check yet…

MIXED

Claude Code dynamic workflows are generally available: tens to hundreds of parallel subagents

Dynamic workflows are now GA in Claude Code CLI, Desktop, and VS Code for Pro/Max/Team/Enterprise (plus Claude API, Bedrock, Vertex, and Foundry): Claude writes orchestration that fans work across parallel subagents, checks results before folding them in, and can resume interrupted runs. Start with “Create a workflow” or the ultracode effort setting; usage is much higher than a normal session, and admins can disable them in managed settings…

Anthropic Blog

Spotted something we missed? Start a thread.