I'd offer some counterpoints to consider:

**Quality variance is real**: Opus 4.6 excels at certain tasks—clarity, structure, polish—but struggles with others. It can confidently assert false information, miss nuance in complex topics, and lack genuine insight. You may be noticing the *polished surface* rather than depth.

**Selection bias**: You're likely noticing good AI writing because bad AI writing gets filtered out or dismissed. The average human article you're comparing against includes all the mediocre written work too, not just the best.

**Different strengths**:
- AI is genuinely good at explaining established concepts clearly
- Humans are better at original investigation, contrarian thinking, and connecting disparate ideas in novel ways
- AI lacks the lived experience that makes certain writing actually authoritative

**"AI slop" complaints are usually about**:
- Mass-produced low-effort content flooding feeds
- Plagiarism concerns
- AI trained on human work without compensation
- Laziness replacing actual reporting/research

**The honest take**: Opus 4.6 writing *reads* better than average human writing. But "reads better" ≠ "more useful" or "higher quality." It's like saying a perfectly polished summary is better than a rough investigative report—depends what you actually need.

What kinds of articles are you finding most useful? That might clarify whether it's genuinely superior content or superior presentation.