What grephn's AI compression actually looks like.
These are hand-curated demonstrations of how grephn compresses long Hacker News threads into a 30-second read. Each one mirrors the structure you'll see in the live app: a TL;DR, structured arguments for and against, what's still unknown, what it means going forward, and notable quotes. Open the app to run compression on any current thread.
Postgres 18 adds asynchronous I/O
Postgres 18 ships async I/O via io_uring on Linux, finally letting checkpoints, sequential scans, and WAL writes run without blocking on disk. The thread splits between cloud-Postg…
Why I switched from VS Code to Zed after 8 years
A long-form switch story from a senior dev. Hacker News splits three ways: Zed enthusiasts confirming the speed gains, VS Code defenders pointing at extension ecosystem gaps, and a…
Anthropic releases Claude 4.5 with 1M context window
Anthropic ships Claude 4.5 with a 1M token context, beating GPT-5 on coding benchmarks and matching Gemini 3 on multimodal. The thread is unusually concrete: real users posting rea…
The state of async Rust in 2026
A reflective post on what async Rust got right, what it got wrong, and what's still unresolved. The HN response is almost entirely from people who ship async Rust in production — r…