30 candidatos extraídos · 0 já revisados · 30 pendentes · 0.5% de cobertura de vocabulário
“It helps that Cursor is used by more than half of the 500 largest tech companies on the Fortune 500.”
“A low-latency sync engine passes encrypted context to the server, which runs inference.”
“A Merkle tree is a tree whose every leaf is the cryptographic hash of the underlying file (e.g.”
“The leaves of the tree are files.”
“Also, the epic effort of moving to turbopuffer in hours, during a large indexing outage.”
“The way it manages this is through indexes of the codebase.”
“Also, this is not code itself, but embeddings, as Cursor doesn’t store code in its databases.”
“This meant we either needed to build a brand new IDE, or fork an existing editor.”
“This means Cursor needs to generate autocomplete suggestions, which the Cursor team calls tab suggestions.”
“To complicate things, I defined a similar method called addTodo() in index.html, inline.”
“Engineering challenges.”
“How the autocomplete works.”
“TypeScript and Rust, cloud providers, turbopuffer, Datadog, PagerDuty, and others.”
“It tries to locate the embeddings that are best matches for the context using vector search.”
“The team uses Cursor to build Cursor.”
“This means both the client and server store their respective Merkle trees.”
“20 NVIDIA H100 GPUs in a Microsoft Azure data center.”
“Database migrations out of necessity.”
“Here’s how it does this: Create code chunks.”
“Forking let us focus on the experience, not the editor.” Backend: TypeScript : most business logic written in this.”
“They also operate tens of thousands of NVIDIA H100 GPUs.”
“Sualeh says: “We are very much a ‘cloud shop.’ We mostly rely on AWS and then Azure for inference.”