Claude 5

1. Show HN: A Claude usage menu bar small enough to read before you run it (github.com via hn)

   +2 6w

   Tokease A lightweight macOS menu bar app showing your Claude 5-hour and weekly limits in real time The Claude limit tracker that never touches your token or Keychain. It reads only what official Claude apps already publish on your Mac.

2. Stop micromanaging Fable/Claude 5 (shumer.dev via hn)

   +2 6w

   Every time I show something I built with Fable, I get the same questions. "I can't get results like this." "How did you do that?" "There's no way that's real." And people assume each one came out of some insanely complicated prompt, or tha…

3. Claude 5 models are visibly smarter and better even if verbose - they just need steering. (www.reddit.com via reddit)

   6w sonnetopus

   It's now been enough time to get a proper read on Claude Sonnet and Opus 5 compared to the previous families and though the verbosity and flip flopping kind of language takes time to get used to, they are much much better at getting real w…

4. Do you believe there's an important difference between the "agentic loop" and the "reasoning loop" that's perhaps at play when it comes to people's perception of how good a model is? (www.reddit.com via reddit)

   6w agentic

   For example, Claude 5 is designed to rely on greater amounts of agentic-loop-based self-correction to arrive at a result and has a fine-tuned primary "instinct" to unblock execution and power through. It's good for coding, clear goals, and…

5. Claude 5 family's hallucinations look a lot like internal Anthropic emails (github.com via hn)

   +11 7w anthropic

   The script used to gather, and the outputs related to, this blog post.

← all threads