This is an amazingly incomprehensible Google Meet pop up
26.02.2026 17:11
👍 3
🔁 0
💬 0
📌 0
This is an amazingly incomprehensible Google Meet pop up
Inspired by Platformer/Hard Fork, I vibe coded a new website while my kids were napping using Claude Cowork. It was so much fun. www.shelbygrossman.com
I love it here so much 🌵
Fascinating thread, fascinating situation providing evidence in support of the theory that in life so many things come down to the details
I didn’t know you were from Arizona! Let me know if you are ever in the Phoenix area!
100% student reported but thank you nonetheless!
PS: The FAA took down a calculator to let you figure out how much radiation you got on a recent flight, so we built one for you! cronkitenews.azpbs.org/howardcenter...
These are the kinds of non-coding but also not silly use cases that I don't think people are talking enough about. (And it seems insane to me that AI companies are not doing a better job of selling this stuff.)
Minutes later they had reached out to these experts to request interviews. They still watched the whole hearing while multi-tasking, but this allowed them to get moving quickly.
While students were investigating, a Congressional hearing on the topic happened. It was several hours long.
The students pasted the YouTube link into Gemini and asked something like: "Who testified on anything related to radiation, and what did they say?"
You can watch/read it here: www.scrippsnews.com/scripps-news...
(Fun fact: You are also exposed to harmful ionizing cosmic radiation every time you fly! And there is no safe level of this kind of radiation!)
My team at the Howard Center for Investigative Journalism just published a story on how the FAA fails to protect flight crews from cosmic radiation.
Flight crews have some of the highest occupational exposure and higher rates of cancer.
I follow most of them on social media, and it feels like they only highlight complex coding use cases or silly stuff (eg building an app to remind you to go to bed - real example I saw last week)
There is a big middle ground they aren't promoting. Example from investigative journalism:
🧵 Possibly controversial hot take: AI companies’ PR teams are not showing what AI is actually good for.
The image is a screenshot of text. The text says: Module 1: Understanding the AI Landscape • Learn key frameworks for understanding AI. • Understand effective prompting strategies. Module 2: AI for Generating Leads, Doing Research, and Finding Sources • How to use AI tools for lead generation and document analysis. • Techniques to identify potential interview subjects and mitigate hallucinations. Module 3: AI for Data, Geolocation, and Translation • Using AI to clean and analyze data, translate documents, and verify information. • Insights from investigative journalists using AI for public information requests and classifying police records. Module 4: The Dark Side of AI • Explore privacy, content theft, environmental impact, and the misuse of AI. • Learn how to investigate AI-driven covert propaganda campaigns and other online harms and strategies for finding leads related to misuse.
The course is designed to be really practical. Highlights include:
-30+ bite-sized video lectures on specific AI approaches.
-Step-by-step worksheets using real investigative contexts.
-Monthly Zooms & forums to share findings with other participants.
A picture of the website where you can register for the course.
👩🏫 I spent 6 months building something I’m super proud of: a course for journalists on using AI for investigations. It officially launched today! These approaches have already improved my own investigations. Get started: careercatalyst.asu.edu/programs/ai-...
Keep us posted please!
Today I drove 30 minutes out of the way to return to this post office to send a package. It was somehow more magical than last time.
My son played at a kids table with board games (!!!) while I waited in the short line.
The woman in front of me called it Phoenix’s best kept secret.
Super important work.
My May paper with @shelbygrossman.bsky.social on AI-generated CSAM recommended that state lawmakers update mandatory-reporter laws to cover "nudify" incidents in schools. Now, Pennsylvania lawmakers are working to do just that, spurred by an incident our paper discusses. www.witf.org/2025/10/07/l...
Agree with Riana. NCMEC staff are truly doing god's work, but many actors in this space lose their mind with the language NCMEC is using here. ⬆️ in CyberTips ≠ ⬆️ increase in harm. Claiming otherwise is misleading.
And the AI problem is real, but many of these AI CyberTips may be failed generations.
Professor of Practice and Investigative Reporter/Editor --
The Beam / The Howard Center for Investigative Journalism at Arizona State University
jobs.chronicle.com/job/37883940...
This role could be a good fit for a social scientist:
Professor of Practice and Data Journalist/Editor --
The Beam / The Howard Center for Investigative Journalism at Arizona State University
jobs.chronicle.com/job/37883942...
🎒 My team is hiring for two positions! Your colleagues will be amazing, the Cronkite School is amazing, ASU is amazing, Phoenix is amazing. 👇
🍺 Are you a journalist in the Phoenix area thinking about AI? The Howard Center for Investigative Journalism at the Cronkite School at ASU is hosting a happy hour on Thursday, October 9 from 4-6pm. RSVP here: docs.google.com/forms/d/e/1F...
Oh neat, I should be following this stuff.
Super interesting. Looks like it hasn’t rolled out with Google Workspace. I made non-white images too and don’t see it.
I'm not sure what model Google Workspace with Gemini is using. Here's the full image - I don't see a visible watermark.
I've been playing around with Google Workspace with Gemini, and asked Gemini to make an image for a slide on LLM benchmarking.
You know I would go out on a limb and say that a bot shouldn't do that to a 13 year old child either but I'm not a Meta AI policy person