AI “Slop”
📰 AI IN THE NEWS
This week I discuss the issues with "AI Slop," and why Mark Cuban's take on AI is about as wrong as a person can be.
🖊 ON MY WRITING DESK
My acknowledgments section is about done. Scroll down to see a list of scholars I wanted to call out, by name, for their mentorship and advice over the years.
📖 AI SCHOLARSHIP
Kids and teens are adopting the use of generative AI at a pace that exceeds their adoption of social media. Should we be worried about it? (Probably.)
🎨 ON MY PAINT DESK
Fans of the X-Men? No? Well, you should be, because they're awesome. This week I started a new project. You might recognize him as a villain, but maybe he's just misunderstood?
🎥 YOUTUBE
Priorities! We all have them. When it comes to research and work it's important to be strategic about what's prioritized. This week I discuss some strategies.
Until next week!
Stephen J. Aguilar
📰 AI IN THE NEWS
AI Slop
It'll surprise absolutely no one that generative AI is being used to create a lot of garbage. Medium, which I briefly tried to use and then gave up on, is suffering from around 47% of it's content being written by generative AI. One might ask: does it really matter if no one reads it? I'm going to chime in and say: maybe.
My less flippant take is that of course it matters—although "to whom?" and "in what way?" are more important questions. Regardless, the reason we have to pose such questions is that written human expression has been turned into a cheap commodity "written" by machines. We get what we pay for.
Speaking of garbage words:
Mark Cuban is the billionaire I dislike the least, as his opinions are typically coherent. This take, however, is either in bad faith or simply dumb. First, there is no such thing as zero education. The only distinction the matters is formal education vs. informal education.
If I were to be less literal, "no education" means "no formal education," so let's start there. How might one learn to "ask AI models the right questions," as Mark put it? Could it, perhaps, be some sort of, oh I dunno, formalized set of practices that teach people things? Maybe all these "learners" will gather together to learn stuff taught by "educators."
We'll call the process: Schmeducation.
His point about performance is also just...annoying—and wrong. Did you know that if I drive a car I can outpace someone who chooses to walk? It's true. I've seen it! Why even bother with these weak-ass legs of mine? Everyone should be in a car at all times. VROOOM.
In all seriousness, we shouldn't fetishize optimizing performance. Work is done by people. Who the worker is, and what they're expected to do, matter just as much as how quickly they can do it. Of course, this assumes that you see workers as fellow humans, and not simply cogs in a corporate your machine. Only a technocrat would frame a world where humans are mere "AI pilots" as something to strive for rather than a dystopia to avoid.
🖊 ON MY WRITING DESK
It's been nice reflecting on the experiences that have shaped who I am as a thinker and as a writer. Those experiences have led me to write Authenticating Intelligence. The book may land as a work akin to an old man who yells at an AI-cloud, but it's been fun so far, so why stop?
My path has been directed by some incredible people. Below is a list of professors who I'm thanking in my acknowledgements. In some ways this book will be a final synthesis of the ideas I've been playing with since I was and undergrad over 20 years ago. (I'm not old, you're old!) You'll notice that most of the list is made up of philosophers!
Alisa Carse (Georgetown)
Barry Fishman (University of Michigan)
Maggie Little (Georgetown)
Martha Nussbaum (University of Chicago)
Stuart Karabenick (University of Michigan)
William Blattner (Georgetown)
W. Gerrod Parrot (Georgetown)
I see the folks above as individuals who served as inflection points in my intellectual development. Spending time in their classrooms and offices changed how I thought about the world.
📖 AI SCHOLARSHIP
AI is here—whether we like it or not.
Let this sink in: "A 2023 report by UK regulator found that two of five children (age 7–12 years) and four of five adolescents (age 13–17 years) are now using generative AI tools and services." This pace of adoption is also faster than social media adoption was. Manefield and colleagues make the case that if we are to understand the impact of generative AI on kids, we need to get our act together pretty quickly.
I agree! On campus faculty often talk about catching students using AI to do all of their assignments, and how it's both annoying and worrisome. What's the point of going to college if AI does most of the work? Our gut—and eyes—tell us that many students are forsaking learning to optimize "performance," but that road leads to ruin. Ok maybe not ruin, but it definitely leads to dumber people.
The fact of the matter is that if we are to understand the impact of AI on learning, motivation, and intelligence, then we need to collaborate with industry. Otherwise we will be forced to do post-mortem analysis on the effects of a new technology on the learning process. And, while saying I told you so is fun sometimes, it won't lead anywhere productive if all I have to look forward to is feeling smug.
🎨 ON MY PAINT DESK
This week I started a new piece: Juggernaut. He's a game piece for Marvel Crisis Protocol, but he's chunky enough to really get in and play with some of the muscle volumes as as well as try to tell a story about him being a dick and breaking everything in his path. To start, I've blocked in the main colors. I'm going for an old school paint scheme that emphasizes browns, reds, and more subdued tones.
Over the next few weeks I'll start adding shadows and highlights. My goal is to finish him in time to enter him into the Path of The Worthy painting competition in a little over a month.
🎥 YOUTUBE
My YouTube channel focus on helping graduate students and researchers stay productive. This week I tackled prioritization. When should you say no? When should you say yes? How do those things change based on where you are in your career? Research presents endless questions, but learning how to focus on the right ones is essential.
