We describe our internal benchmark CursorBench which represents a more realistic sampling of coding problems. We discuss why we think it is important to include the complex problems software engineers see everyday.
SOFTWARE
-
Composer 2 Uses Continued Pretraining and RL for Coding
By
–
Composer 2 had three main efforts: continued pretraining, reinforcement learning, and benchmark development. The goal of each was to closely emulate the Cursor environment to produce a highly intelligent coding model.
-
SambaStack: Hardware-Software Stack for AI Inference
By
–
Simplify AI with SambaStack, the top hardware & software stack built for AI inference. Deploy on-prem or in the cloud to accelerate innovation with dedicated SambaNova infrastructure. Learn more: https://
sambanova.ai/products/samba
stack?utm_source=x&utm_medium=organic&utm_campaign=enterprise
… -
Replit Developer Relations: Meet Matt, a Builder and Creator
By
–
Meet @mattyp from the Replit Developer Relations team
— Replit ⠕ (@Replit) 24 mars 2026
He talks to builders, creates cool stuff, and somehow makes it look too easy
If you like shipping fast, learning faster, and working with people like Matt, we should probably talk pic.twitter.com/MfcmYXFuSbMeet @mattyp from the Replit Developer Relations team He talks to builders, creates cool stuff, and somehow makes it look too easy If you like shipping fast, learning faster, and working with people like Matt, we should probably talk
-
Package Registry Update Delays in Ubuntu Debian Systems
By
–
The registries that DO make those promises are things like the Ubuntu/Debian operating system package repos – there's a reason it can take weeks, months or even years for updates to show up there
-
Claude AI Gains Full Computer Use Capability for Task Automation
By
–
Full computer use. This is huge Claude (@claudeai) You can now enable Claude to use your computer to complete tasks. It opens your apps, navigates your browser, fills in spreadsheets—anything you'd do sitting at your desk. Research preview in Claude Cowork and Claude Code, macOS only. — https://nitter.net/claudeai/status/2036195789601374705#m
→ View original post on X — @jiquanngiam, 2026-03-24 19:31 UTC
-
1T Parameter Kimi Model Runs at 1.7 Tokens per Second M4 Max
By
–
Here's 1T parameter Kimi running at 1.7 tok/s on an M4 Max
-
Claude Code Launches Auto Mode for Smarter Permission Decisions
By
–
New in Claude Code: auto mode. Instead of approving every file write and bash command, or skipping permissions entirely, auto mode lets Claude make permission decisions on your behalf. Safeguards check each action before it runs.
-
Apple Rebuilds Siri into System-Wide AI Agent for iOS 27
By
–
Apples Siri gets a huge AI rework this year, biggest rework ever. The summary tl;dr •Apple is rebuilding Siri into a system-wide AI agent, not just a voice assistant •iOS 27 introduces a conversational, chat-like Siri (text + voice) •New standalone Siri app with chat
