engblogs

summaries of the latest blog articles from your favorite tech companies.
Lambda LabsLambda Labs

2025 AI wrapped

2025 AI wrapped distills how reasoning-oriented models, massive context windows, and multimodal capabilities propelled production AI, achieved open-source parity, and shifted from training to inference within evolving infrastructure and agentic workflows.

1/13/2026
DatabricksDatabricks

Open Sourcing Dicer: Databricks’ Auto-Sharder

Open Sourcing Dicer: Databricks’ auto-sharder is a dynamic, in-memory shard management control plane that continuously rebalance assignments to keep distributed services low-latency, highly available, and efficient across workloads.

1/13/2026
Google CloudGoogle Cloud

A gRPC transport for the Model Context Protocol

Explores enabling gRPC as a native transport for the Model Context Protocol (MCP) to boost performance, security, and developer productivity, comparing it with transcoding gateways and outlining pluggable transport support via the MCP SDK.

1/13/2026
Google DeepMindGoogle DeepMind

Veo 3.1 Ingredients to Video: More consistency, creativity and control

A technical overview of Veo 3.1's approach to video production, emphasizing improved consistency, creativity, and control through modular features.

1/13/2026
CloudflareCloudflare

What we know about Iran’s Internet shutdown

A technical analysis of Iran's January 2026 Internet shutdown, tracing IPv6 address-space collapse, HTTP/3 and QUIC traffic declines, and disruptions across major providers via Cloudflare Radar data.

1/13/2026
Google CloudGoogle Cloud

New Google Public Sector research shows that nearly 90% of federal agencies are already using AI

An in-depth look at Google Public Sector’s finding that nearly 90% of federal agencies are using AI, the security and skills barriers to adoption, and how Gemini for Government and OneGov initiatives aim to accelerate AI-driven transformation across the federal landscape.

1/13/2026
Modular AIModular AI

Modular: How I Beat Unsloth's CUDA Kernel Using Mojo—With Zero GPU Experience

Zero-GPU-experience journey to beat a CUDA NF4 dequantization benchmark in Mojo using AI-assisted design, packed 32-bit stores, and occupancy tuning to scale across GPUs.

1/12/2026
Apple MLApple ML

Over-Searching in Search-Augmented Large Language Models

Systematic evaluation of over-searching in search-augmented LLMs, revealing when external search improves accuracy versus abstention failures, introducing Tokens Per Correctness (TPC), examining multi-turn dynamics and noisy retrieval, and outlining mitigation strategies together with the OverSearchQA benchmark.

1/12/2026
MetaMeta

CSS at Scale With StyleX

Explores StyleX, Meta's CSS-at-scale framework that blends CSS-in-JS ergonomics with static CSS performance, enabling atomic styling with deduplication to shrink bundle sizes across major products and highlight open-source adoption.

1/12/2026
AirbnbAirbnb

Pay As a Local

Explores strategies for integrating local payment methods and region-specific checkout flows to improve conversions and user experience.

1/12/2026
Two SigmaTwo Sigma

AI in Investment Management: 2026 Outlook (Part I)

Explores how AI will transform quantitative investment management in 2026, portraying AI as the operating system for quant research and investing, enabled by agentic AI and integrated workflows, with ongoing emphasis on human supervision and disciplined governance.

1/12/2026
Google CloudGoogle Cloud

AuraInspector: Auditing Salesforce Aura for Data Exposure

AuraInspector reveals how misconfigurations in Salesforce Aura and the GraphQL Aura controller enable data exposure, and provides an automated tool to audit access controls and remediate vulnerabilities.

1/12/2026