Claude is a family of AI assistants developed by Anthropic. Named versions (Claude 3.5 Sonnet, Claude 3 Opus) are known for long context windows, coding ability, and safety-focused design.
Anthropic's Claude formalized Fermat's Last Theorem in 11 days using 6 billion output tokens, a task that Kevin Buzzard's £1 million project aims to complete over five years. The achievement demonstra…
Anthropic, OpenAI, Google, and Microsoft have introduced AI 'Skills' that turn recorded demonstrations of work into reusable, portable files, marking a shift from files as outputs to files containing …
A developer's guide details how to build a production-ready RAG pipeline using Claude and the Model Context Protocol (MCP), emphasizing that MCP servers enable AI to read files directly from disk, eli…
Evals Coach, a Claude plugin by justshipai, guides product managers through designing AI evals step-by-step, producing importable test cases, judge prompts, and calibration guides. In a six-task blind…
A tutorial by Rachel Goldstein demonstrates how to regression-test LLM prompts using promptfoo 0.122.2 in GitHub Actions, catching prompt and model regressions on every pull request. The setup include…
A developer named Heppu built an ESP32-POE-based Bluetooth-over-Ethernet bridge to eliminate latency in remote gaming setups, using an optical HDMI connection and a Claude subscription to solve Steam …
A new analysis argues that many complaints about AI in test automation stem from user error rather than genuine tool limitations, citing the most dangerous pitfall as AI models silently rewriting asse…
CodexBar, a free and open-source macOS menu bar app, launched to help developers track usage limits, credit balances, and reset countdowns across AI coding tools such as OpenAI's Codex and Anthropic's…
A developer ran the same single-period input through three different AI agent runtimes and observed starkly different behaviors: a bare model greeted the user, OpenAI's Codex asked for direction, and …
Chess5.ai, an open-source web app by developer Wangxing-Ye, lets users play chess, Go, Xiangqi, Gomoku, and Othello against large language models including GPT, Claude, Gemini, Grok, Mistral, DeepSeek…
A developer has created an open-source agentic skill called 'prompt-me' that flips the typical AI writing workflow: instead of the user prompting an AI to write text, the AI prompts the user to write,…
Salesforce made Anthropic's Claude the default reasoning model across Agentforce, Slack, and its engineering org on 26 August, but kept the deal non-exclusive and Claude selectable in Agent Builder, p…
Anthropic released a blueprint enabling retail companies to integrate AI shopping agents built on Claude into their e-commerce platforms, citing a 35% increase in online basket size and a 60% increase…
A developer has built Orivox MCP, a Model Context Protocol server that connects Claude to the Orivox website platform, enabling the AI to create, edit, and publish real websites through natural conver…
An engineer has developed PII Guardrail Studio, a local, air-gapped reverse privacy proxy and encrypted token vault designed to run inside a private VPC, enabling enterprises to use external LLM APIs …
A developer built BandwidthGuard, a free browser-based tool that analyzes JSON payloads to identify unnecessary data bloat in API responses. The tool provides field-level breakdowns, compression compa…
Anthropic's AI model Claude produced the first complete computer-checked proof of Fermat's Last Theorem in 11 days, writing 13 million lines of Lean code and proving 29,500 intermediate theorems. The …
A Wikipedia talk page discussion proposes adding a 'TermWatch' template to pages of companies listed at termwatch.io, which would link to AI-generated summaries of terms-of-service changes. The list i…
SupXH (api.supxh.xin), a third-party AI API relay service for OpenAI-compatible LLM access, has been rated high risk by the directory Uprouter, which cites limited or no official documentation, an opa…
An AI developer using a VOICEVOX-compatible text-to-speech engine found that its pitch-recalculation endpoints returned 0.0 for all mora values, because the engine is end-to-end neural and does not ex…