AI Infrastructure news and analysis on Web Pulse: 32418 curated articles tracking the latest AI Infrastructure developments, tools, and research, updated continuously from vetted sources.
South Korean Science Minister Bae Kyung-hoon announced May 29 the government will "double down" on support for the domestic AI sector to become a "full stack" supplier, including development of a frontier foundation mode…
A development team cut its AI coding costs by 65% after discovering that 60-70% of API calls did not require expensive frontier models. By implementing automatic task-level routing that assigns simpler models to routine …
Researchers have developed a conditional scaling law that predicts optimal large language model architectures for inference efficiency, training over 200 models from 80 million to 3 billion parameters. The study found th…
Claude Code versions 2.1.154 through 2.1.158 contain a regression that corrupts the streaming delivery of tool results back to the model, causing results to arrive empty-then-late in bursts, duplicate inline blocks up to…
Deputy Prime Minister and Science and ICT Minister Bae Kyung-hoon on Friday touted expanded AI infrastructure, record R&D spending of 35.5 trillion won, and regulatory reforms as key first-year achievements, urging more …
A developer has created a set of 12 reusable configuration files for Cursor and Claude Code that encode consistent AI behavior patterns across projects. The files include specialized subagents for code review, security a…
OpenAI and Anthropic have launched multi-agent autonomous features for enterprise use, enabling AI systems to coordinate and execute complex business workflows without human intervention. Anthropic released Claude Manage…
The apcore project has introduced "Thought Spans," a new OpenTelemetry-based tracing system that captures AI reasoning steps alongside traditional system performance data. By wrapping each AI agent execution in a rich Op…
Sentry engineers encountered a rate-limiting bug when Cursor Automations exhausted a human-sized bucket of 60 requests per 60 seconds in seconds, highlighting the challenge of gating MCP servers for agent traffic. Three …
A developer successfully ran Google's Gemma-4 E2B-it large language model on a Raspberry Pi 4 using llama.cpp, achieving text generation speeds of 1.5 to 1.8 tokens per second. The project involved converting the model t…
A developer building a team of AI agents to generate reports from transcript data spent a month rewriting jobs as durable executions after cascading errors from failed API calls and memory issues broke the entire generat…
Litterbox, a new Linux sandbox environment for developers, launched to provide partial isolation between containerized development environments and host systems by restricting file system access to a single folder and us…
Trajectory, in collaboration with UC Berkeley Sky Lab and Anyscale, released a concurrent multi-LoRA training stack for continual learning that achieves a 2.81× end-to-end experiment-throughput gain over single-tenant re…
A developer in Ouagadougou, Burkina Faso, built a vision-language model that detects blockchain attacks invisible to text-based security systems. The fine-tuned model, named Imina-Na, identified a Drain Star attack that …
Data Workers, an open-source autonomous agent swarm for data engineering, bet on the Model Context Protocol (MCP) to connect its AI agents to dozens of tools in the modern data stack, rather than building custom integrat…
The W3C draft specification WebNN defines a graph-based neural network API that enables JavaScript to submit static model graphs to native OS runtimes, according to a November 21, 2025 blog post on ziade.org. Browsers ma…
A hardware startup founder's guide to manufacturing in China reveals that most crowdfunding campaigns fail not during fundraising but in the four-to-eight-month transition from prototype to factory-ready design. The guid…
A developer has outlined a five-layer security stack to give AI agents like Claude or ChatGPT access to Snowflake data without exposing personally identifiable information (PII). The approach uses role-based access contr…
Taiwan has launched the "Healthy Taiwan" initiative, placing digital health care at its core by integrating big data, artificial intelligence, and cloud technologies to improve care quality and efficiency. The government…
Canonical has released Workshop, an open-source tool that creates reproducible development environments from a single command using YAML files. The tool uses unprivileged LXD system containers with strict sandbox control…