Robots and autonomous systems generate massive multimodal data streams that existing data infrastructure cannot efficiently handle. The physical world lacks standardized formats, edge processing, and cross-modal synchronization capabilities that the current cloud data stack assumes, creating an opportunity for new infrastructure companies to build the equivalent of Snowflake for physical AI.
A new Postgres extension called chdb leverages ClickHouse's in-process engine to enable fast imports and exports from cloud object storage in multiple data formats. Benchmarks show chdb outperforms competing extensions like pg_duckdb and pg_lake by 2-3x on CSV, JSON, and Parquet imports, while supporting a broader range of formats through ClickHouse's native capabilities.
Let's Encrypt replaced broken log processing scripts with ClickHouse, a self-hosted data warehouse offering cost-efficient storage and fast analytics. The deployment on three PowerEdge servers reduced stats computation from hours to under 10 seconds and enables rapid incident investigation without impacting production systems.
ClickHouse has a permission bypass vulnerability where unprivileged users can read restricted data by disabling query validation in UPDATE mutations, which execute with full database privileges. The issue affects ClickHouse 26 and was rejected by the vendor's disclosure program.
A technical analysis of three internet datasets using ClickHouse: identifying underrated LLM posts by comparing bookmarks to likes across small accounts, measuring LessWrong's drift on AI alignment over time using embeddings, and tracking adoption of 'vibe coding' terminology on Hacker News.
Data Science Weekly Issue 669 covers practical ML and data engineering topics including maze-solving neural networks, data backup strategies, statistical sampling for inference, ClickHouse cluster operations at scale, RAG system implementation in fintech with lessons learned, open-source dbt Charts for dashboard governance, climate risk disclosure in Miami real estate, and limitations of MSE as a model evaluation metric.
WalShadow is an open-source engine that replicates Postgres data to ClickHouse directly from physical WAL, achieving sub-second latency (around 200 ms) and sustaining 289K rows/sec throughput. It eliminates logical replication slots and operational overhead by consuming the same physical WAL stream used by Postgres replicas, decoding it externally, and writing ClickHouse-native blocks directly. ClickHouse is also launching WalShadow in private preview for its managed Postgres service.