Michel Estefan

Nothing like starting out my day with a bite of chocolate cake to make me feel like an accomplished adult. 🤦

when the number of jobs per category is likely quite varied.” (Thanks @midge_the_merciless for pointing this out)

#AvoidSnark #Ai #LaborEconomy #AiEconomy

Another issue with radials someone mentioned was that, “by combining, [radial charts] can accidentally imply that each spoke is equivalent in size, making it seem like you're seeing an overall view of the actual vs potential vs not replaceable with AI,

My comment was inspired by research showing pie charts are easily misinterpreted by people. Radials have some similar features. So it depends on whether you’re going for elegance or making sure folks who aren’t deep into the data/non-experts won’t misinterpret.

Got lots of pushback on this in another platform. In hindsight, I wish I had avoided the snark in my comment. #Sorry Mainly folks were arguing that the bar chart doesn’t add any info compared to the radial chart, but it does add a lot of visual clutter.

Anthropic should have just created a bar chart. Apparently none of their LLMs told them that.

Here you go. Directly from Anthropic: how much work in different sectors of the economy is currently handled by #Ai vs how much could be theoretically handled by it. #TheFutureofWork #Ai #labor #AiEconomy

But still, there’s a difference. Anyway, I didn’t know this part of the story; all I had heard were the criticisms and moral judgements.

#MyColleaguesAreAwesome #LoveMyJob #AcademicLife

another is that he was naive and did something that turned out to be a failed political strategy. Many will say he should have seen it coming.

This puts his administration in a slightly different light because one thing is to think he was a hypocrite or didn’t abide by his purported values across a series of policy domains;

we won’t discuss anything until you deport the people that “shouldn’t" be here. So he did that and then of course no negotiation happened.

Apparently, the reason his admin ramped up deportations was that he kept reaching out to Republican members of Congress to open negotiations for comprehensive immigration reform and he kept getting the same answer:

The second was a discussion about deportations during the Obama era. Obama is often famously referenced as the “deporter in chief” and it’s an issue often highlighted to criticize his administration.

The impact is more visible (and more real) when organizations retool their production and managerial processes around the new technology.

employment numbers, doesn’t happen in the initial stages of individual adoption.

In a conversation around Ai, a colleague mentioned that the reason it’s hard to see (in the data) the impact of Ai in the economy at this point is because historically speaking, the impact of technological advancements in macro indicators like GDP, productivity, or

Here are two things I learned recently from these conversations; one a quick insight, the other filled a gap in my historical knowledge.

I learn a ton from listening to them. Sometimes I learn new stuff I just didn’t know, sometimes they have the perfect way of articulating something I had been thinking about, other times they have really interesting questions or ideas.

One of my favorite parts of my job is being surrounded by a lot of people doing fascinating intellectual work. Recently, I’ve been cultivating the habit of asking my colleagues what has been on their minds recently, what questions or topics have drawn their intellectual attention.

#PewResearchCenter

The distribution of AI "cheating" by income isn't primarily a story about ethics. It's a story about inequality.

#HigherEd #AI #Sociology #AcademicIntegrity #WeKeepItCivilYall

under what conditions are students more or less likely to feel compelled to resort to Ai?

To be clear: wealthier students cheat too, and plenty of them use AI. The point is we have to think about cheating sociologically, not in terms of individual moral judgement, especially from that standpoint of schools trying to better serve students:

But if you're working a closing shift, managing financial stress, maybe caring for a sibling, and you've got a participation quiz due in 15 minutes –the "choice" to use AI looks less like a character flaw and more like triage. Students like this are structurally cornered.

Here's the thing: some keep framing AI use in school as individual moral failure. Laziness. Shortcuts. Bad character.

Then there's new Pew data showing that one in five teens in households earning under $30K do all or most of their schoolwork with AI chatbots–nearly three times the rate of teens in $75K+ households.

www.pewresearch.org/internet/202...

The classroom was never a firewall, but now the cheating infrastructure might literally be on your student’s face.

Let that last one sink in. Because smart glasses mean the distinction between "physically present" and "plugged into the internet" has basically collapsed.

Yup, in-person. Students gaming digital clickers, using LLMs to complete participation assignments, and now smart glasses.

In-Person Classes Aren’t Safe From AI Cheating Boom One professor was surprised to learn how easy digital technology makes it for students to cheat on class participation assignments and some exams.

Folks, two things I've been sitting with recently:

@insidehighered.com has a piece out today by Kathryn Palmer about AI cheating in in-person classes.

www.insidehighered.com/news/student...