Cursor

Evidence (15)

• 7The New Stack·9/25/2026product_launch
  OpenAI and Cursor Move Toward Multi-Agent Coordination

  OpenAI has opened its Agents API in public beta, while Cursor launched Projects to coordinate multiple coding agents on larger software tasks. The article says both products adopt a similar coordinator-worker architecture, with one agent managing context and routing work while specialized agents handle individual steps. The piece frames this as part of a broader shift in AI-assisted software development toward more reliable multi-agent orchestration.

• 7The New Stack·9/24/2026product_launch
  Cursor launches Rollouts for monitoring code from PR to production

  Cursor has introduced Rollouts, an agent that monitors code changes after a pull request is opened and tracks them through deployment into production. The system generates a monitoring plan, compares telemetry in staging and production, and can flag regressions, alert developers, or open a revert pull request.

• 7The New Stack·9/21/2026model_release
  Grok 4.7 improves long-horizon coding, but still lags top rivals

  SpaceXAI released Grok 4.7 with training focused on long-running, multi-hour tasks, emphasizing self-verification and long-context management. The model improved on endurance benchmarks such as Terminal-Bench 4.0 and CursorBench 4.0, but still trails Anthropic’s Claude Fable 5.1 on Terminal-Bench and remains unreliable in many long coding workflows.

• 7Hacker News·9/21/2026model_release
  Grok 4.7 launched for coding and knowledge work

  xAI introduced Grok 4.7, a new model focused on coding and multi-hour knowledge-work tasks. The company says it improves self-verification, long-context handling, and safety, while offering pricing starting at $2 per million input tokens and $6 per million output tokens.

• 6Hacker News·9/3/2026research
  Study compares Claude, Codex, and Cursor tool choices across 17k runs

  The article describes a large-scale experiment measuring which tools coding agents choose across thousands of realistic repository tasks. The researchers built a panel of 75 repositories, ran 16,893 sessions in ephemeral sandboxes, and used simulated-human and judge agents to evaluate valid runs and final winners.

• 4Hacker News·8/29/2026research
  OpenAI explains its decision on Cursor after SpaceX acquisition

  OpenAI published an explanation of its decision regarding Cursor following its reported acquisition by SpaceX. The post frames the decision as a policy and business response to the change in ownership, rather than a product or technical update.

• 7The New Stack·8/27/2026product_launch
  Replit makes Auto mode the default for model routing

  Replit is rolling out its intelligent model routing system as the default for all accounts, automatically selecting models based on task needs, speed, and cost. Core and Pro users can still override Auto mode and choose models manually, while enterprise admins can limit routing to approved models.

• 7The New Stack·8/21/2026product_launch
  Spline V2 adds MCP access for Claude Code and other agents

  Spline released V2, a rebuilt 3D editor that lets external coding agents operate directly on live, editable scenes through a bundled MCP server. The update also includes Spline’s own AI agent and extends the workflow between the 3D editor and Hana, allowing scene and interface changes to stay editable inside the project.

• 7InfoQ·8/17/2026product_launch
  Cursor launches Origin, a GitHub alternative

  Cursor has introduced Origin, an early-beta code hosting and collaboration platform available to paid users. It includes repositories, pull requests, code browsing, GitHub sync, and integrations with tools such as Vercel, Depot, and Buildkite.

• 6The New Stack·8/15/2026research
  Coding agents shift infrastructure from developers to changes

  The article argues that per-developer environments no longer fit modern software delivery because coding agents let one developer run many workstreams in parallel. It proposes change-level tenancy as the new unit of isolation and capacity planning, since the real demand now comes from changes in flight rather than headcount.

• 8The New Stack·8/15/2026model_release
  Grok 4.6 matches top model performance at lower cost

  Grok 4.6 launched with performance that rivals or exceeds leading frontier models while keeping the same API price as Grok 4.5. The article argues that downloadable weights from competitors like DeepSeek and Qwen are increasingly capping what closed labs can charge, shifting competition toward routing and model selection rather than raw capability.

• 8Hugging Face Blog·8/13/2026research
  Hugging Face Hackathon Reproduces 2,200 ICML Papers

  Hugging Face reported results from a community hackathon that attempted to reproduce ICML 2026 papers at scale, with more than 1,200 participants generating 6,816 logbooks across 2,226 papers. The effort used coding agents and automated judging to verify claims, falsify results, or mark reproductions as toy or inconclusive, highlighting both the scale of AI research output and the limits of manual peer review.

• 7Hacker News·8/12/2026model_release
  SpaceXAI releases Grok 4.6 with longer-task coding gains

  SpaceXAI released Grok 4.6 less than a month after Grok 4.5, highlighting training that emphasizes self-checking, error correction and sustained progress on longer coding and agent tasks. Benchmark results show clear gains over Grok 4.5, but Grok 4.6 still trails the top models on several coding and terminal tests.

• 7The New Stack·8/7/2026framework_update
  Coinbase, Shopify and Ramp build in-house coding agents while still using Anthropic

  Coinbase, Shopify, and Ramp have each developed internal coding agents for software engineering workflows, but they continue to rely on commercial tools and frontier models from Anthropic, OpenAI, and Google. The article argues that enterprises are increasingly owning the agent harness — the execution layer for context, permissions, orchestration, tool use, and verification — rather than the model itself.

• 5Hacker News·8/5/2026research
  Building a Production-Grade Agentic Harness

  The article explains how to evolve a basic LLM agent loop into a more reliable production harness by adding planning, parallel execution, typed tools, layered memory, verification, budgeting, and tracing. It uses a city-comparison agent as a running example to show how these pieces improve debuggability, measurability, and fault tolerance without hiding the underlying mechanics behind a framework.