Results for “Context window”

Why it matters

Hugging Face released funes, an open-source CLI tool providing durable, cross-agent working memory across Claude Code, Codex, pi, and Hermes. It deterministically indexes local session traces into Lance datasets, reducing context reloading overhead up to 8x compared to manual handoffs.

Why it matters

Adding more than 15 tools to an AI agent degrades selection accuracy and spikes token costs. Implement pre-filtering gates and semantic vector retrieval (RAG-MCP) to keep agent operations precise and efficient.

Why it matters

Google has launched agentic video understanding across Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite. The feature scans video segments dynamically, reducing token consumption by up to 88% and API costs by up to 66%.

Why it matters

Alibaba released open weights for Qwen3.8-2.4T-A95B with 2.4 trillion parameters and 95B active per token. Featuring a hybrid linear/full attention architecture and configurable reasoning depth, it serves at 4K tokens/sec/GPU on NVIDIA GB300 systems.

Why it matters

Anthropic introduced inter-session messaging for Claude Code, allowing active terminal instances to communicate directly. Users can ask one Claude session to send a summary to another active session mid-task without sharing file history or raw conversation logs.

Why it matters

NVIDIA Labs released NOOA, an open-source agent framework that defines agents as single Python classes with type annotations. By passing live object references instead of text dumps, NOOA cuts agent token consumption by half on SWE-bench Verified while reaching 82.2% accuracy.

Why it matters

Researchers leveraged GPT-5 to resolve a three-year-old bottleneck in immunology related to T-cell behavior. The model successfully synthesized complex biological datasets to provide a actionable hypothesis.

Why it matters

A developer spent fifteen hundred dollars evaluating whether LLM agents could successfully identify and exploit custom application vulnerabilities. While they solved basic issues, they struggled with complex, multi-step logic flaws. Use structured pentesting suites for automated security evaluation.

Why it matters

Anthropic has raised $65 billion in a Series H funding round, valuing the AI company at $965 billion post-money. This massive capital infusion, one of the largest private rounds ever, signals intense investor confidence in the future of agentic AI and foundational models. It provides the runway to aggressively scale compute, research, and developer tooling.

Why it matters

Exa launched Agent Ultra, a deep research API powered by a subagent swarm designed for exhaustive web entity discovery. The service crawls and aggregates large-scale structured candidate lists for autonomous agent pipelines.

Why it matters

A bug in OpenAI Codex Desktop app serializes local model chat transcripts and routes them to OpenAI backend servers during background memory syncs. This occurs even when telemetry and analytics are completely disabled. Developers running private models can prevent leaks by setting memories feature to false in their config.

Why it matters

Alibaba released Qwen 3.8 27B with vision capabilities under the Apache 2.0 license. Testing by Simon Willison reveals that its default 'xhigh' reasoning parameter causes severe overthinking, requiring users to explicitly configure reasoning levels.