Dell Expands AI Data Platform Capabilities for 2027
Dell Technologies plans to integrate a semantic layer and knowledge graph into its AI platform by 2027 to improve data context and processing efficiency.

Dell Technologies plans to integrate a semantic layer and knowledge graph into its AI platform by 2027 to improve data context and processing efficiency.

Datalab has launched OmniParseBench, an open-source evaluation harness featuring 16,288 tests to audit the performance of OCR document parsers.

Astrophysicist Brice Menard and Claude Science have created the first complete ultraviolet map of the sky by using AI agents to process 38,000 observations and fill data gaps.

Yandex has open-sourced YTsaurus Flow, a real-time data processing technology that reduces latency for advertising models and analytics from over 10 hours to near-instant speeds.

Jev is a hybrid tool designed to bridge the gap between structured and unstructured data by providing confidence-scored assessments for business processes.

Microsoft is testing new Excel features that enable cells to store multiple structured values, including lists and nested arrays, to improve data management.

Microsoft is testing a multi-value cell feature in Excel, enabling users to store lists within single cells and filter individual items without splitting data.

A content creator has compiled a decade of professional articles into a single dataset to enable AI-powered information retrieval and personalized insights.
A comparison of log shippers reveals that only OpenTelemetry and Vector offer native deduplication, while others require custom scripts or backend-side filtering.

OpenAI utilizes a human-led initiative called Project Lily to review ChatGPT conversation logs for model refinement, with reviewers earning over $50 per hour.

Delta Lake 4.3 introduces new selective overwrite APIs and shifts table operations to the Unity Catalog REST API to ensure cross-engine consistency.

A developer built a custom .NET application to process 14.5 million stock bars and filter 13,420 active tickers using a local SQLite database.

Modus has launched version 1.6.14 of its ETL platform, featuring improved load balancing for 1C:Fresh data collection and optimized processing for large JSON files.
