This paper identifies that long-running AI research agents get stuck in "idea basins"—localized regions of solution space—causing independent runs to plateau at different performance levels. The authors propose fork-and-flush, a method that periodically branches the agent into parallel trajectories with fresh chat contexts while preserving accumulated work, then continues from the best performer, achieving 66% relative improvement over single-run baselines across 13 complex research tasks.

• Independent runs of the same agent on identical tasks plateau at substantially different scores, with performance gaps persisting even after significant additional compute, suggesting agents explore only localized regions of the solution space rather than broadly searching.
• Fork-and-flush works by forking the agent into K parallel trajectories at episode boundaries, clearing the chat history while keeping the workspace intact, running each for a fixed horizon, then continuing from the highest-scoring branch—this simple intervention requires no architectural changes.
• Across 13 long-horizon tasks (individual runs up to 107 hours), fork-and-flush achieved 66% relative improvement over single-run and 44.4% over best-of-N baselines under equal compute budgets, with artifact embeddings confirming broader exploration through solution space.

Docglow replaces built-in dbt docs with an interactive single-page app that works out of the box with any compiled dbt Core project. It generates a static site featuring unlimited models, column-level lineage, project health scores, full-text search, AI-powered chat, and dark mode with zero configuration.

• Two commands (pip install + generate) turn compiled dbt artifacts into a static docs site with no dbt Cloud or backend required.
• Includes column-level lineage tracing with transformation labels (direct/derived/aggregated) and a health score for docs/test/description coverage.
• Can gate CI pipelines with --fail-under and cut large-project payloads by up to 60% by skipping column lineage or slimming output.
• AI chat and an MCP server let editors query project metadata directly, with chat streaming via a user-supplied Anthropic key rather than shipping it.

Yelp explains how it turned a two-week prototype into a scalable, production-ready AI assistant for business pages. They built near-real-time indices for reviews, photos, and structured data in an EAV schema, combined keyword-first retrieval with LLM prompts, and added query classification and trust-and-safety filters. The system streams answers with citations, logs metrics, and balances freshness, performance, and reliability.

• Yelp built its AI business-page assistant from two-week prototype to production in nine months by swapping Redis warehouse snapshots for streaming ingestion, keeping reviews, photos, and core business data fresh within 10 minutes (menus/community posts still update weekly).
• An EAV schema in Cassandra lets them add new structured fields on the fly without schema migrations.
• Combining keyword-first retrieval with LLM-based query expansion hit quality targets faster than pure embedding-based search.
• Batching only the slow-moving data sources (rather than everything) kept the real-time streaming pipeline manageable.

The Brave Search API lets developers access real-time web search data for applications like chatbots and AI tools. It offers features like summarized answers, high query capacity, and a focus on privacy with zero data retention. Plans include a monthly credit allowance for usage.

• Independent index of 30B+ pages (not reliant on Google/Bing) with millions of daily updates, plus $5/month free credits
• Handles up to 50 queries/sec standard, 2 queries/sec for AI-summarized answers
• Zero data retention policy keeps user queries private
• Custom enterprise plans available for larger-scale needs