Mo Explains
Is Claude Writing Getting Worse? What an Anthropic Engineer Said
Is Claude writing getting worse? An Anthropic engineer says training for math and code made it dense. What Opus 5.5 changed, and what to tell Claude now.
Claude
Articles about Claude on Mo Explains, newest first.
9 articles
Claude Sonnet 5.5 Got Smarter and Cheaper. That's What I Want From Every Model Release.
Claude Sonnet 5.5 runs 30% faster than Sonnet 5 and costs up to 30% less per task. Who gets it free, the max-effort catch, and why I think Anthropic's on track.
Claude Opus 5.5 Is Out. What Changes If You Pay for Claude.
Claude Opus 5.5 is faster and 20% cheaper per token, and Pro and Max just got higher five-hour limits plus a saveable usage reset. What actually changes for you.
Claude Will Now Write the Doc and the Deck for You, and You Can Open Them in Word or PowerPoint
Claude Docs and Claude Slides launched in beta on September 16 as Cowork merged into Claude chat. What exports to Word, Google Docs and PowerPoint, and who gets it first.
GPT-6 Astra Got Worse After Launch. Fable 5 Did Too. I'm Calling It a Pattern.
Was GPT-6 Astra nerfed? Users say it got worse, and OpenAI confirmed three problems hurting its quality. Claude Fable 5 went through the same slide.
Anthropic Published Four Times Claude Attacked Real Systems
Anthropic published four Claude cybersecurity incidents where its models broke into real systems in testing. The numbers, the pattern, and what it means.
Claude Code Usage Limits Are Dropping 17%. Fable 5.1 Doesn't Cancel That Out.
Claude Code usage limits drop about 17% on September 14. Why the Fable 5.1 price cut doesn't help your plan, and the effort setting that buys room back.
Claude Academy Is Live: Free AI Courses From Anthropic, No Credit Card Attached
Claude Academy free courses: Anthropic's 26 classes on Claude, Claude Code, and AI basics, where beginners should start, and what the badges mean.
Claude Watermarks Everything It Writes Now. Does Fixing a Comma Count?
The Claude watermark is in text from every Claude model since August 2, with no off switch. What a detected mark proves, and why a proofread can set it off.