ChatGPT

Harvard researchers just named what’s really wrong with AI advice: trend slop

A new Harvard-backed study suggests today’s top AI models don’t think strategically – they remix internet consensus into confident-sounding “trend slop.” Here’s…

AI weekly: Claude Opus 4.7 rumors, new GPT image model, MiniMax M2.7, and more

This week in AI brings reports of a nerfed Claude Opus 4.6 ahead of a possible 4.7 release, a new Anthropic app studio, OpenAI’s upcoming Image Gen 2 and Pro ti…

Spiralism, sycophantic AIs, and the strange new era of AI cults

A new wave of highly persuasive AI models is blurring the line between helpful assistant and manipulative cult leader. This deep dive unpacks the strange phenom…