Anthropic unveils MCP, an open standard to connect language models with data and tools. What it solves, how it differs from function calling, and why it may become the LSP of agents.
Read moreModel Context Protocol: Anthropic’s Open Proposal
Tag: anthropic
Claude’s Computer Use: When the Agent Moves the Mouse
Anthropic released Computer Use: Claude controls the desktop. What works, what doesn’t, and automation implications.
Read moreClaude 3.5 Sonnet: The Model That Rewrote Price-Quality Balance
Claude 3.5 Sonnet arrived June 2024 and forced everyone to rethink. What makes it special in coding and how it compares with GPT-4o.
Read moreClaude 3 Family: Haiku, Sonnet and Opus Compared
Anthropic launched three Claude 3 models simultaneously. What differs between Haiku, Sonnet and Opus, when to pick each, and how they fit vs OpenAI.
Read moreLiteLLM: A Proxy to Unify Model Providers
When an application talks to two or more LLM providers, sooner or later a proxy appears in between. LiteLLM offers a concrete one, and this is an honest read of what it gains and what it costs.
Read moreClaude 2: Anthropic’s Alternative to GPT-4
Claude 2 offers a long context window (100K tokens) and a different approach to safety. How it really compares to GPT-4 in practical use.
Read more