So… is ChatGPT making us dumb? A viral MIT study, what it actually found, and the delicious irony of how we consume research. (Steffi Kieffer, Substack)

People expect that AI will make news cheaper, more current, and easier to understand, one of the findings of the 2025 Digital News Report. (Nic Newman, Reuters Institute)

OpenAI is launching its own podcast where Sam Altman claims that coding with o3 gives people their next big “wow moment.” He suggests that while his AI won’t exactly cure cancer, it will make researchers more productive. Of course, critical questions are missing.

LLMs have a “lost in the middle” problem – they focus on the start and end of documents but miss key info in between. (Adam Zewe, MIT News)

30 People to Follow at the Intersection of AI and Academic Publishing. (Helen King, LinkedIn)

“A.I. Is Poised to Rewrite History. Literally.” is a love letter to NotebookLM and how historians are using it to analyze data and structure books. (Bill Wasik, New York Times Magazine)

SEO agency analyzes 75k AI Overviews, finds that one-fifth cited media as sources. 31 percent came from BBC, NYT, and CNN.

This research paper argues against “reasoning/thinking” hype: intermediate tokens often lack substance, despite appearances.

After 6 years in hiding, Venture Capitalist Mary Meeker is back with one of her famous trend reports. This time on AI. 340 pages of statistics.

But AI isn’t creative, isn’t funny, isn’t surprising! Always just average, everything’s been done before. That’s kind of true – but here comes a paper saying: You can actually train LLMs to be creative too. It’s called Creative Preference Optimization.