The Who's Who of AI

“Claude users are creating public share links, but probably don't realize that means their chats are now ending up in Google searches where anyone can dig through them.”

“Tons of peoples' Claude chats are exposed on Google. Cryptocurrency keys, API keys, login credentials, legal discussions, more. This happened with ChatGPT last year. And Claude too. There's clearly an ongoing issue with search engines indexing peoples' AI c…”

“This Is Donald Trump’s AI Brain Trust www.wired.com/story/this-i...”

“These are the admin power players in Trump's ear about AI regulation (and they're often telling him different things):”

“Just as A.I.-generated books are flooding Amazon, vibe-coded apps are flooding the App store, via @kalleyhuang.bsky.social www.nytimes.com/2026/07/20/t...”

“I wonder if the concept of apps will fade a bit in time. I think vibecoding will eventually take a turn towards more niche applications instead of large market appeal. I vibecoded a chore/allowance app for my son, and it works well, but I'll never put it up…”

“Welcome to the AI-generated poster hall of shame.”

“Thank you everyone who sent me horrible AI generated flyers from around the world at my request. I was fully overwhelmed with hundreds and hundreds of these. These are the worst I saw. The hall of shame. Please enjoy and/or numb your brain: www.404media.co/…”

“Nearly 94% of previously correct MedQA answers fail when medical LLMs face dynamic red-teaming. This new Nature Health paper exposes a massive benchmarking gap where models passing static exams fail under stress. They also leaked privacy data in 86% of clin…”

“Bluesky’s AI assistant Attie expands into an open social research tool techcrunch.com/2026/07/24/b...”

“Google’s Gemini nears billion-user milestone techcrunch.com/2026/07/23/g...”

“Hemispheric “has raised $52 million in funding after gathering data on 100,000 people’s brains to train deep learning models to examine the brain without the need for invasive procedures.””

“🧪 A quarter of a million hours of brain data from 100k people is powering a new AI effort to decode cognitive health. Hemispheric, led by FaceID co-inventor Gidi Littwin, raised $52M to train deep learning models on EEG signals. They're heading to the FDA n…”

“Placing a clinician in the loop doesn't guarantee medical AI safety if they lack time or authority to disagree. The authors here outline four system requirements for real oversight: • Epistemic capacity • Cognitive space • Decisional authority • Effective i…”

“This short piece is lining up with where I think we might see science and AI head. I envision AI use evolving more as a customized toolsets, instead of a one size fits all approachl. Give the AI work they tend to avoid, but reserve the elements they relish …”

“A survey of 272 AI researchers, experts estimate a 12% chance that AI causes a catastrophic outcome by 2030, even with active mitigation. The primary fear among industry leaders isn't generic software failure, but AI expediting the design and synthesis of n…”

“Interesting perspective here. Clinical AI assumes training data influence can persist indefinitely. With evidence evolving and patients withdrawing consent, machine unlearning needs to become standard infrastructure to safely erase targeted data without a m…”

“🧪 Nearly 30% of a CRISPR protein's sequence was altered in new AI-guided designs, a huge leap from the usual 1% to 2% tweak. What's interesting is that these synthetic enzymes still successfully edit human and plant cells, sometimes even outperforming their…”

“Google is rebranding NotebookLM as Gemini Notebook. The big change is shifting its focus from basic text summaries to more complex data analysis.”