AI Coding Live News — Real-Time 𝕏/Twitter Updates & Breakthroughs

👀50

🔥89

Moonshot AI Kimi k1.5 launches hybrid code retrieval API cutting semantic search latency across 10M-line codebases to 80ms

Moonshot AI released a specialized Hybrid Code Search API powered by Kimi k1.5. Combining dense vector semantic embeddings with sparse AST inverted indexes, it resolves cross-file call hierarchies across 10-million-line monolithic codebases with search latency under 80ms.

Key Takeaways

Combines dense vector embeddings with AST syntax indexes to disambiguate overloaded symbols.
Delivers sub-80ms semantic retrieval across code repositories exceeding 10 million lines.
Exposes standard REST and gRPC endpoints for easy integration with enterprise code review tools.

Read full story

Linked Models & Agents:

⚡ Kimi k1.5 (Reasoning)🌐 Moonshot AI (月之暗面)

#Moonshot AI#Kimi k1 5#Code search#Semantic retrieval#Enterprise API

Sourcegraph@Sourcegraph·Collected 64d ago

👀29

🔥85

View on 𝕏 ↗

Sourcegraph Code Finder Reaches General Availability for Enterprise-Scale Semantic Search

Sourcegraph announced the General Availability of Code Finder, offering sub-second semantic code search and dependency mapping across enterprise multi-repo architectures.

Key Takeaways

Officially reaches General Availability (GA), delivering enterprise-grade search across massive monorepos;
Combines deep syntactic parsing with vector embeddings to map complex cross-service dependency graphs;
Integrates seamlessly with Cody AI agents and enterprise continuous integration developer pipelines.

Read full story

Linked Models & Agents:

Cursor@cursor_run·Collected 64d ago

👀45

🔥95

View on 𝕏 ↗

Cursor Launches Background Indexing 2.0: Sub-Millisecond Semantic Code Search on 10M+ Line Monorepos

Cursor rolled out Background Indexing 2.0 with incremental AST optimization and lightweight embeddings for 10M+ line codebases.

Key Takeaways

Incremental AST indexing recomputes only changed nodes during Git branch switches, saving 75% CPU
Multi-file codebase reference accuracy rises to 94.8%, eliminating cross-module agent hallucinations
Supports local offline embedding mode keeping sensitive enterprise code 100% on-device

Read full story

Linked Models & Agents:

⚡ Claude 3.7 Sonnet (Hybrid Reasoning)⚡ OpenAI o3-mini (High Reasoning Coding)🤖 Cursor Composer (Agent Mode)🌐 Cursor (Anysphere)🌐 Anthropic

#Cursor#Indexing 2.0#Monorepo#AST Indexing#AI IDE