A critical security flaw in LangChain Core (CVE-2025-68664) allows attackers to exploit serialization injection to steal secrets and manipulate LLM responses, prompting urgent updates to affected versions to mitigate risks.
LangChain, an AI infrastructure startup known for its tools to build and monitor LLM-powered applications, is raising a new funding round at an estimated $1 billion valuation, driven by its popularity and expansion into LLM observability with its product LangSmith, despite increased competition and evolving APIs from major providers.
LangChain, an open-source AI software maker, has launched its first paid product, LangSmith, which helps programmers building on large language models like OpenAI’s GPT-4 to better track and test their underlying code. The startup, led by CEO Harrison Chase, has confirmed a $20 million funding round led by Sequoia and already has a waitlist of 80,000 for its new tools. LangSmith has been adopted by companies like Elastic, Rakuten, and Moody’s, with 5,000 users working with it monthly. LangChain, which started as a tool for Chase's personal use, has quickly gained investor interest and aims to provide value in the rapidly evolving AI ecosystem.
Andrew Ng has launched a new generative AI course on LangChain, in collaboration with LangChain founder Harrison Chase. The course, titled "Functions, Tools, and Agents with LangChain," aims to update developers on the latest advancements in language models and how to use LangChain to work with their models. It covers topics such as function calling, structured data extraction, and building conversational agents. Additionally, Ng also announced a livestreamed course on building Computer Vision models.