Fifteen minutes. That’s the time I take to scan, curate, and post here. Each topic a light in the night sky, connect the dots, see the Constellation, tell the story. 🚀☄️
Uncover more stars in the sky, greg report Ai 2027
OpenAI’s Computer History turns clicks and keystrokes into a timeline ChatGPT and Codex use to pick up unfinished work, while House legislative counsel is spending extra time repairing Ai-generated bill language arriving from congressional offices and outside groups. Minnesota is defending its nudification ban against xAI, and John Nosta’s Psychology Today essay separates the words appearing on a page from the judgment and reasoning behind them, a distinction getting harder to ignore when fluent prose arrives in seconds.
greg report ChatGPT’s Computer History tracks your clicks and keystrokes: The Verge reports that OpenAI’s opt-in Computer History records activity events on a Mac so ChatGPT and Codex pick up unfinished work, with OpenAI’s Ari Weinstein writing that it lets ChatGPT “learn from everything you do on your computer,” and the forgotten browser tab now has a better memory than the person who opened it.
greg report Ai slop is swamping a House office that drafts US laws: POLITICO reports that the House Office of Legislative Counsel is spending time repairing Ai-generated legislative language riddled with errors, with one House staffer calling the situation “absolutely terrifying,” and Legislative Counsel has acquired a new congressional chore: proofreading the robot before proofreading Congress.
greg report Minnesota defends Ai ‘nudification’ ban after lawsuit from Musk’s xAI: Reuters reports that Minnesota Attorney General Keith Ellison is defending the state’s nudification ban against xAI’s First Amendment challenge, while xAI’s complaint calls the law a “clumsy attempt to prohibit ‘nudification,’” and a federal courtroom is now doing content moderation with constitutional law books open.
greg report Ai and the Two Forms of Authorship: Psychology Today’s John Nosta separates linguistic authorship, the sentences themselves, from cognitive authorship, the judgment and reasoning behind them, writing “Textual fluency stopped being evidence of anything,” and anyone working alongside an LLM now has an uncomfortable question hanging over the page: who actually did the thinking?
Tuesday’s THX 1138 has been reassigned to the typing pool, where every keystroke is remembered and the bill still comes back covered in red ink.
g🚀☄️
The free Morning Tap stays right where it belongs: daily, fast, and open.
Paid subscribers get the extra layer: sharper breakdowns, extended takes, behind-the-post notes, field observations, subscriber-only essays, and the occasional strange little artifact pulled from the Ai attic before the rest of the neighborhood sees it.
You also get a seat closer to the editors. Paid subscribers can comment, push back, suggest topics, pose questions, share observations, and trade notes with other readers who see the same strange signals blinking across the morning sky.
Think of it as the back room of Morning Tap. Same coffee. Fewer tourists. Better arguments.
Subscribe here for the paid Morning Tap edition.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.