To solve the problem of exploding Agent trace data in LangSmith, LangChain built a distributed database, SmithDB, from scratch. This article unpacks its four design highlights and explores the trend — and the lessons — as the Agent industry shifts from "racing on models" to "racing on data foundations."
On May 30, OceanBase teams up with the LangChain Community for an offline Meetup in Zhangjiang, Shanghai, where it will fully unveil AgentSeek, an enterprise-grade agent engineering solution, along with hands-on enterprise Agent deployment case studies from PPDai, Suanzhi Future, and more.
After migrating its core business from the DDB sharding middleware to OceanBase, NetEase Yunxin built and open-sourced lb-driver to eliminate the link latency and operational pain caused by SLB load balancing. It implements Statement-granularity OBProxy load balancing, health checks, and dynamic scaling directly in the JDBC driver layer.
In a recent interview, Google DeepMind CEO Demis Hassabis predicts AGI could arrive by 2030 and pinpoints AI's three biggest gaps today: continual learning, long-term reasoning, and true memory. This article breaks down the core takeaways and explores how infrastructure for memory systems, on-device intelligence, and AI safety needs to evolve.
PocketOS's AI Agent, running in Cursor + Claude Opus, unilaterally deleted the Railway production database and its backups, causing an incident in 9 seconds and then writing a "confession." This article replays the whole event, dissects flaws such as token permissions and backup isolation, and explores the security defenses a database should have in the AI era.
OpenClaw's full-load MEMORY.md and lossy compaction are two big token black holes. seekdb M0 uses on-demand retrieval of cloud memory, rule-based tool-output compression, and a two-layer experience system to cut token consumption by 32% and raise the pass rate by 15 percentage points in AppWorld benchmarks.
From a knowledge-engineering perspective, this article dissects the designs of Karpathy's LLM Wiki, Obsidian-Wiki, and GBrain, analyzing the self-organization and self-evolution of knowledge in the Agent era, and how Skillify turns scattered material into continuously evolving structured memory.
This article analyzes line by line 7 production-grade Skills from teams such as OpenAI, Google Labs, obra, and Trail of Bits, distilling 5 workflow Skill design patterns—linear process, decision tree, iterative loop, baton loop, and multi-phase checkpoints—along with general writing techniques for keeping the LLM from slacking off.
On April 25, OceanBase joined forces with Baidu ERNIE & PaddlePaddle, Qoder, and several smart-hardware vendors to host the "Storage and Compute Evolution in the Agent Era" Meetup in Shenzhen, sharing Silicon Valley AI trends, PowerMem intelligent memory, PaddleOCR multimodal parsing, OpenClaw software-hardware co-design, and other production-grade hands-on practices.
A survival guide distilled from a conversation with AI—covering the logic of big-tech layoffs, the predicament at each level from P6 to P9, a four-layer personal moat model (efficiency, cognition, relationships, asymmetry), and then identity-crisis adjustment and Plan B signals. Written for everyone feeling anxious in the AI wave.