Every story we have kept, newest first.
Looking for the daily editions? → Past editions
Page 91 of 1373
Barclays warns that bipartisan voter backlash against the rapid buildout of AI infrastructure could create political risk for the sector. The concern is especially relevant ahead of the US midterm elections, potentially affecting the market's enthusiasm for AI-related investments.
Bloomberg Technology · 23d ago
Free ChatGPT accounts can now use the upgraded task scheduling tool. Paid users can also configure prompts to trigger when events occur in Gmail, Slack, or GitHub.
Engadget · 23d ago
This guide explains how to use LangSmith for dataset management, fine-tuning, and evaluation of LLMs. It includes practical examples covering LLaMA2 and GPT-3.5 fine-tuning workflows.
LangChain Blog · 23d ago
This article explores how to build and benchmark question-answering systems over CSV data using LangChain agents, retrieval, and LLM evaluation. It also shares debugging insights and open-source code.
LangChain Blog · 23d ago
Timescale Vector introduces an integration for LangChain that enables AI developers to use PostgreSQL for vector search. The update highlights 243% faster similarity search, time-based RAG, and a 90-day free trial.
LangChain Blog · 23d ago
NVIDIA announced the Jetson Orin Nano 2, a new AI module for embedded devices. Compared with the current 8GB Jetson Orin Nano, it delivers twice the inference performance while using 40% less power at equivalent performance.
ITmedia AI+ (日本) · 23d ago
LangChain has raised a $10 million seed round led by Benchmark. The funding will support developers building data-aware, agentic LLM applications with LangChain's open-source framework.
LangChain Blog · 23d ago
LangChain introduces an Airbyte destination integration designed to automate data ingestion for production retrieval applications. It supports scheduling, text splitting, and access to more than 50 embedding options.
LangChain Blog · 23d ago
LangServe provides a playground UI for deploying and experimenting with LangChain applications. Developers can configure parameters, test different models, share the interface with teams, and stream results in real time.
LangChain Blog · 23d ago
LangChain introduces flexible retrieval methods for building AI applications tailored to a user’s data. Developers can use semantic or hybrid retrievers to create personalized ChatGPT-style experiences.
LangChain Blog · 23d ago
LangSmith introduces Pytest and Vitest/Jest integrations for running evaluations. The integrations let developers incorporate LangSmith evals into familiar testing workflows.
LangChain Blog · 23d ago
Cube and LangChain are presented as a combined solution for building AI-powered data experiences. Cube’s semantic layer can help ground natural-language queries, reduce hallucinations, and support conversational interfaces.
LangChain Blog · 23d ago
LangSmith introduces Role Based Access Control to help enterprises manage access to resources. The feature includes custom roles and API keys for more structured permission management.
LangChain Blog · 23d ago
NVIDIA Dynamo’s preview Shadow Engine Recovery feature is designed to restore LLM inference capacity within seconds after an engine process failure. It avoids the lengthy cold-restart process that requires reloading model weights, compiling kernels, and capturing CUDA graphs.
NVIDIA Developer Blog · 23d ago
FlyBuys has revealed Nex, a virtual assistant designed to support analysts. The company remains cautious about expanding the assistant across the wider business.
iTNews Australia · 23d ago
SpaceX reportedly plans to use NVIDIA GPUs for its Starmind project. Elon Musk’s stated goal is to harness solar energy for AI, raising questions about how GPU systems will operate in the space environment.
Engadget · 23d ago
OpenAI confirmed that Chris Malone, the executive leading its data centre buildout, has left the company. The responsibilities of his former role have been divided as OpenAI pivots toward leasing data centre capacity.
The Next Web (TNW) · 23d ago
Chinese AI models are becoming a major hidden component of AI services used by US companies. Chinese-origin models account for 8 of the top 20 models in Arena AI’s coding rankings, while they represent 58% of tokens routed by US companies through OpenRouter.
ITmedia AI+ (日本) · 23d ago
LangChain’s retriever enables automated web research by running parallel searches, scraping pages, and synthesizing findings with LLMs. The workflow can run locally or in the cloud.
LangChain Blog · 23d ago
LangChain’s Plan-and-Execute framework separates task planning from execution for complex AI agent workflows. The approach is designed to reduce errors and improve reliability.
LangChain Blog · 23d ago
Chris Malone, an OpenAI executive involved in overseeing the company's data center buildout, has left the startup. The departure affects a leadership role connected to OpenAI's infrastructure expansion.
Bloomberg Technology · 23d ago
Anthropic has merged Claude’s chat and Cowork memory systems, with the unified memory enabled by default. Information shared in Claude chats is now available to Cowork, while knowledge learned by Cowork can flow back into Claude.
The Next Web (TNW) · 23d ago
Vercel launched Run SDK, a sandboxed runtime for executing untrusted JavaScript and type-stripped TypeScript without direct access to application secrets, Node.js, or the network. Developers can expose narrowly scoped host functions and pause execution for authentication or human approval, then resume without repeating completed work.
Vercel News · 23d ago
Stability AI, the company behind the image generator Stable Diffusion, has raised $76 million in fresh funding. The latest round brings the company’s total fundraising to $232 million.
TechCrunch AI · 23d ago
Taiwanese prosecutors charged nine people with illegally exporting high-end B300 AI servers to mainland China. The defendants include one Nvidia employee and two Super Micro employees, highlighting escalating enforcement around AI hardware restrictions.
The Guardian Technology · 23d ago
Amazon OpenSearch Service now supports MCP Apps, enabling AI agents to return interactive visualizations alongside text responses. A locally run MCP server can guide developers from alerts to traces and logs, helping identify root causes within a single IDE conversation.
AWS Machine Learning Blog · 23d ago
Klarna's AI assistant supports customer service and personal shopping for 85 million active users. Using LangGraph and LangSmith, Klarna achieved 80% faster customer resolution times.
LangChain Blog · 23d ago
LangSmith Workspaces enables enterprises to separate resources across teams, business units, or deployment environments. The feature improves collaboration and organization for teams managing AI development and operations.
LangChain Blog · 23d ago
Anthropic is reportedly preparing to tell IPO investors that its potential revenue opportunity exceeds $30 trillion. The figure would surpass the $28.5 trillion opportunity SpaceX presented before its record listing.
The Next Web (TNW) · 23d ago
This LangChain guide explains how to query SQL databases using natural language and LLMs. It focuses on techniques for reducing hallucinations and improving the reliability of text-to-SQL systems.
LangChain Blog · 23d ago