
Cursor is in the X bundle. Claude usage is a separate bill.
The subscription includes the editor, but its shared allowance covers eligible first-party models. Team plans have another charge to check.
Contributor
Reporting and context from the SInt newsroom.

The subscription includes the editor, but its shared allowance covers eligible first-party models. Team plans have another charge to check.

Meta describes an assistant that carries context across conversations and stays with a task. A remembered preference is only the beginning.

Connecting an assistant is the beginning. Permissions, missing fields and a clear record of what changed decide whether it can finish the job.

An urban planner used Claude to build a playable city. The interesting part is how it lets you question the assumptions behind the result.

ElevenLabs and Google let you direct a performance. Suno adds music. The useful test is what happens when you change one line.

A public finance task shows the gap between a working workbook and a financial conclusion that follows the evidence.

Researchers used GPT-6 Astra to repair a smaller model's repetitive thinking. A failed follow-up helps explain what the result does, and doesn't, show.

An outside assessment estimates roughly doubled progress on well-defined tasks. Its conclusions are public, but important parts of the evidence behind them aren't.

Cloud workspaces become read-only a week earlier. Conductor offers a transfer to your Mac and an account export, but the local app is also heading toward a wind-down.

The company says it needed more time to protect its model files. Its filing also describes failures in tests that were supposed to stay contained.

OpenAI’s new model is cheaper to run through its API. For an invoice or a coding job, the saving depends on how much work it takes to get a correct result.

Project of the week: Morluto and contributors turn familiar software into something you can investigate, one feature at a time.