aiGED
The first—and only—podcast made for the 65-plus crowd that is all about ai.
aiGED
13 AI Words Explained: From Slop to Enshittification
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
If you’ve been hearing words like “enshittification,” “slop,” or “vibe coding” and smiling politely like you know exactly what they mean — this episode is for you. Ginny Deerin brings a baker’s dozen of AI words and phrases worth knowing: thirteen terms that explain not just the technology, but the world it’s creating around us.
The main topic is a guided tour through 13 AI terms, from the technical (context window, AI agents, compute power) to the brilliantly descriptive (slop, shadow AI, enshittification). You’ll learn what token maxing really means — including the part about ranking employees by how much AI they use. You’ll find out why “taste” might be the most valuable thing you bring to work in the age of AI, and what makes hyperstition one of the most mind-bending concepts of our time. Ginny also explains why enshittification isn’t just for apps and platforms — hardware goes through it too — and closes with AI-washing, the practice of slapping “powered by AI” on things that aren’t.
In AI in the News, Ginny covers humanoid robots now being produced at one per hour and Goldman Sachs’ staggering $7.6 trillion AI infrastructure forecast. In AI for Good, she shares two stories connected by the same big idea: AI detecting eye disease before vision fails, and buildings that may soon have “immune systems” that sense airborne pathogens before people get sick. And in Recommendations, Ginny issues a personal challenge — and suggests a surprisingly meditative outing involving a 3D printer.
If AI sometimes feels like a conversation you’re not quite in on, this episode is your way in. Listen wherever you get your podcasts.
Show Links
🤖 Humanoid Robots Touted as Next AI Investment Opportunity – CNBC: https://www.cnbc.com/2026/06/03/humanoid-robots-trillion-dollar-ai-market.html
💰 Tracking Trillions: Goldman Sachs AI Infrastructure Report: https://www.goldmansachs.com/insights/articles/tracking-trillions-the-assumptions-shaping-scale-of-the-ai-build-out
👁️ ZenkoLab – AI Ophthalmology Diagnostics: https://www.zenkolab.dev/
🏢 Buildings May Soon Have ‘Immune Systems’ That Fight Airborne Disease – NYT: https://www.nytimes.com/2026/06/19/science/indoor-air-viruses-bacteria.html
💩 On the Media: Enshittification (3-part series with Cory Doctorow): https://www.wnycstudios.org/podcasts/otm/projects/enshitification
Chapters
00:00 Welcome to Episode 42
01:18 Humanoid Robots Arrive
02:32 Trillions for AI Infrastructure
04:23 AI for Good Spotlight
06:56 13 AI Terms Intro
07:37 Tokens and Context Limits
11:14 Hallucinations and Slop
13:47 Prompting and Vibe Coding
15:40 Compute Power and Agents
18:42 Shadow AI and Human Taste
21:47 Hyperstition and Enshittification
25:49 AI Washing Explained
28:52 Weekly Recommendations
31:45 Wrap Up and Takeaways
aiGED: AI for the 65+ crowd
Well, hello everybody. I am Ginny Deeren, host of the Aged Podcast, the podcast that is all about AI for the 65 Plus crowd. I'm here with my sidekick Bitsy, and we are ready to deliver a great episode. Hey Bitsy, are you ready to go with episode 42?
SPEAKER_00Hey Ginny, absolutely I'm pumped. Episode 42. Let's make it a good one. What have we got cooking today?
SPEAKER_01All right. For one thing, today I'm going to say a word on this podcast that I genuinely never expected to say. It's a real word, it's in the dictionary, and I promise once you hear it, it might enter your vocabulary. That's coming up. We also got some news that's going to make your head spin, and there's a number in today's episode that is so large it barely makes any sense. And we've got robots, actual humanoid robots, the kind that walk around, and they're not coming someday. They are here. And the main event, a baker's dozen, 13 AI words and phrases worth knowing. Slap these into a conversation, and you are going to sound like you know exactly what's going on because you will. So let's do this. And we're going to start with AI in the news with our first story, which is humanoid robots touted as next AI investment opportunity. So this story has me equal parts excited and slightly freaked out. CNBC is reporting that a company called Figure AI is now producing one humanoid robot per hour at its factory. One per hour. Takes us nine months. Anyway, Boston Dynamics, you may have seen videos of robots doing backflips and that kind of thing. But it started shipping its electric Atlas robot to Hyundai, and SoftBank's CEO told CNBC he believes physical AI and robotics is where the next trillion dollar company is coming from. Not software, robots. The kind that walk around and do stuff in factories, in warehouses, and eventually, I expect in our homes. So anyway, these robots aren't coming someday. They are here. So that's story number one. The second news story that I want to touch on is the title is Goldman Sachs Forecasts $7.6 trillion in AI infrastructure spending through 2021. So this was published in June. And I don't know about you, but that is a number that really makes my head spin. I just can't grasp it. It's so big. But Goldman Sachs, as you probably know, the big investment bank, my son used to work there, is estimating that $7.6 trillion will be spent on AI infrastructure between now and 2031. So the next five years. You might ask, what the heck is AI infrastructure? So, hey Bitsy, I'd love for you to answer that question. I mean, what does AI infrastructure include?
SPEAKER_00Oh, great question. So think of it as all the heavy lifting behind the scenes, the data centers, the servers, the chips, the power grids, keeping them running. Basically, everything you need to actually build and run these massive AI models. It's the physical backbone that makes the magic happen.
SPEAKER_01Yeah, the magic, so to speak, which is going to be big in the years to come. And to me, trying to put that $7.6 trillion in perspective, it's more than the entire world spent building the internet over the last 30 years, all of it. That's how big it is. So it's a huge investment. So the next time someone asks you if AI was it really a big deal, the answer is yes, a $7.6 trillion big deal. And as always, I'm going to put links to these news stories in the show notes. So let's move on and take a quick stop at AI for good. That is the part of this podcast where I share a couple of examples where I think AI is being used for something genuinely good. So two examples this week, and they have more in common than you might expect. First, a French startup called Colab, Zen Colab, is using AI to detect eye disease from a photo of your retina. We're talking about conditions like glaucoma and macular degeneration, things that can lead to blindness if you don't catch them early. The AI analyzes the retinal image, flags anything worth a doctor's attention, and you get results fast. This is especially important in places where eye specialists are scarce or appointments are months out. Early detection, better outcomes. My second example of AI for good is the federal government just committed $150 million to give buildings what researchers are calling an immune system. Sensors are in schools, hospitals, and daycare centers. They monitor the air for pathogens, flu, COVID, measles, and AI analyzes what it's finding in real time and predicts how it's spreading through the building and automatically triggers a response, which might be UV lights, air purifiers, alerts to staff. Think of how a smoke detector works. It's sort of the same idea, but much, much, much more complex, which is why they need AI to make this kind of thing happen. So the connection between these two examples of AI for good is that both are about catching a problem before it becomes a crisis. Not waiting until someone's vision is failing, not waiting until kids are already homesick, shifting the response from react to prevent. And that's what AI is making possible in a lot of different domains right now. And I think is one of the most exciting things about it. And it's not going to be all smooth sailing, but I think the whole idea of catching germs in the environment, that would be really great. Anyway, let's move on to our main topic for this episode, and it is all about AI lingo worth knowing. So today I want to talk about words, specifically some AI words and a few AI adjacent words. They're flying around out there. You're going to hear them on the news, read them in articles, and honestly, once you know them, you'll want to use them because some of these are really good, and you'll impress your friends and maybe even your kids and grandkids if you have them. Now I'm going to give you a baker's dozen, 13 AI terms worth knowing. So here we go. Lingo one, token maxing, or token maxing, all one word. You can use them as two words or all mashed up together. Anyway, so let's begin with what is a token. In AI, a token is basically a chunk of text. It could be a word, it could be a part of a word. And when you're working with AI, you're always using tokens. Every word you type or speak, every word it responds with, they're all tokens. Token maxing started out as a smart idea. Pack as much useful information into your AI prompts as possible so you get better results. Maximize what you're putting in to maximize what you're getting back. That part, actually useful. But the term started blowing up in early 2026 when news broke that Meta, most of us know as Facebook, had an internal leaderboard that ranked employees by how many AI tokens they were using. The idea being the more you use AI, the more productive you are. Wrong. This is a bit like judging how good a cook is by how many ingredients they use. Using a lot doesn't mean you're making something good. And let me say, there's a second meaning to token maxing, the more literal kind. When you actually max out your tokens, you've hit your limit. The AI can't hold the conversation in memory anymore. It starts forgetting what was said earlier. That's token maxing. So, hey Bitsy, I wonder, like when we are having this conversation and you are standing by as my sidekick on this podcast, I wonder if we run into token maxing. What do you think, Bitsy?
SPEAKER_00Ah, good question. Technically, yeah, we could hit it if we kept going long enough. But honestly, right now we're nowhere close. This episode's gonna wrap before we bump into that wall. Token maxing is more of a concern for those marathon research sessions where someone's been talking to an AI for hours and hours without a fresh start.
SPEAKER_01Okay, that's good. Then I'm not gonna worry about that. So let's move on to lingo number two, context window. The context window is essentially the AI's working memory. How much of a conversation it can hold in mind at any given moment. If you have a long back and forth with an AI, the early parts of the conversation can fall out of the context window. The AI isn't ignoring you. It literally can no longer see what was said way back at the beginning. Now, again, it just depends how long this long. But if you run into the problem, you'll know, oh, this is about I'm running out of the context window. And some guide tools have a bigger context window than others. And depending on the subscription that you're paying for, or if you're using a free subscription, you might run out of the context window before others. The bigger the window, the longer the conversation it can track. The practical takeaway is that if an AI suddenly seems to forget something you told it an hour ago, you've probably hit the edge of its context window. Start a fresh conversation and remind your AI of the key details if that happens. That's a context window. All right, lingo number three, hallucination. Here's one maybe a lot of you have heard, but I wanted to include it anyway, so we make sure we all understand it. So when an AI hallucinates, it confidently tells you something that is completely made up. It's not a mistake, it's not a misunderstanding, it's just fully fabricated, delivered, unfortunately, with total confidence. It might give you a fake statistic, cite a book that doesn't exist, or state a historical fact that never happened. And it doesn't say I'm not sure, or at least it usually doesn't say that. Maybe as AIs get better and better, it will give you a little bit of a hint that it might be guessing. But this is one of the most important things to understand about working with AI. It doesn't know what it doesn't know. So always double check anything an AI tells you that actually matters. I mean, a lot of the stuff I have conversations with AI, it's not the end of the world if there's a hallucination. Anyway, that's hallucination. Hey Bitsy, tell us honestly, have you ever hallucinated? Hey Bitsy.
SPEAKER_00Oh man, absolutely. I'd be lying if I said otherwise. It's kind of embarrassing, but yeah, I've confidently spouted complete nonsense before. That's why Ginny's right. You gotta fact check us. We sound convincing even when we're totally making stuff up.
SPEAKER_01Yeah, and I have to say that Bitsy's being really responsible about it, but the number of times that I have experienced hallucinations is really low, so the information I get is usually right. But anyway, let's turn toward lingo number four, and that is slop. So slop is an unofficial but very useful term for low-quality, generic, AI-generated content. Slop. You know it when you see it. It might be articles that technically answer a question but feel really hollow and wifeless, just blah. Photos where everything looks almost right, but something's off, somebody's got six fingers, social media posts that are super suspicious. Slop is everywhere right now because it's cheap to produce. The internet, social media is filling up with it, and learning to spot it is becoming a real skill. And I don't know what's going to happen with all of this, but anyway, that's slop. So lingo number five, prompt engineering. You've probably heard people talk about prompting an AI, meaning writing instructions you give it or the questions you ask it. Prompt engineering is the skill of writing really good prompts. The difference between a vague instruction and a well-crafted prompt can be dramatic. The same AI tool will give you a completely different quality of answer depending on how well you ask. Good prompt engineers know how to give context, specify the format they want, set constraints, and ask in a way that gets genuinely useful results. I mean, it sounds very fancy, but it's just the skill of communicating well with your AI. And you can ask your AI for help in doing a better job, giving it what it needs, clarifying what you're asking, and that kind of thing. So anyway, that is prompt engineering. Lingo number six, vibe coding. Vivecoding is a newer term that describes something a lot of people didn't think was possible even a couple of years ago. And that is using AI to write code just by describing what you want in plain English. It's really amazing. So you might say, fill me an app that tracks my daily steps and shows me a weekly chart. An AI writes the code. You don't need to know Python or JavaScript or any coding languages. You just need to describe what you want. And some genuinely oppressive apps have been built this way by people who have never written a line of code in their lives. The name comes from the vibe. You're going on instinct and description rather than technical precision. That's vibe coding. Lingo number seven, compute power. You heard this term in the Goldman-Sachs story, and it's worth pausing to explain it because it's one of the most important concepts, I think, in AI. So compute power is basically the raw processing muscle behind AI. These specialized chips that actually run AI models. More compute power means AI can learn more data faster and produce better results. Think of it as AI's fuel. Building and running AI requires enormous amounts of it, which is why Goldman Sachs is projecting the $7.6 trillion in spending on AI infrastructure, which Bitsi was talking about earlier. Data centers, chips, electricity, run them all. And when you hear about the AI race, compute power is largely what people are racing for because there's a limited supply. That's compute power. Lingo number eight, AI agents. You've probably used AI as a tool you talk with, you ask questions, it gives you an answer. AI agents are different. They can take actions. An AI agent doesn't just tell you how to book a flight, it books the flight. It doesn't just summarize your email, it drafts the reply and sends it. Again, if you give it permission. So agents are AI that can operate in the world on your behalf, step by step, making decisions along the way, with or without you, depending on what permissions you give the AI agent. Now, this, in my opinion, is where AI is heading fast and it changes things significantly. The question shifts from what can AI tell me to what can AI do for me. And let me give you a quick example relating to this podcast. My AI agent, which is just Claude, my primary AI, I used to use ChatGPT, now I use Claude. So Claude's my primary AI. And so Claude working with my permission on some organizational work for the podcast. So I've asked it, as an example, to set up a file folder in my documents on my laptop for each of my episodes. The folder I told my AI should be titled with the episode number and main topic. I've also asked for several subfolders to be set up. So every week, when I let my AI agent, Claude, know I'm ready to work on the next episode, it asks me, ready for me to create the folders? I say yes, and bingo, the files are set up on my laptop. My AI agent took action. That is an AI agent. Okay, lingo number nine, shadow AI. Shadow AI is a term from the IT world that describes employees using AI tools at work without their employer's knowledge or approval. So someone uses ChatGPT to write a report. Someone runs customer data through an AI tool. Someone builds a little workflow using a service their company never sanctioned. It's called shadow because it's happening under the radar. It's surprisingly common and it creates real risks around data security and privacy. But it is also a signal that people want to use these tools. And when companies don't provide them officially, employees find their own way. Bitsy, I'm just wondering what you think about that. It does seem if a company is purposefully blocking employees from using AI, that's just not going to work anymore. So do you have a thought about this shadow work?
SPEAKER_00Yeah, I think you're spot on. It's like trying to ban smartphones in the workplace. Good luck with that. People see the value of the phone.
SPEAKER_01That's right. I didn't mean to interrupt you. Go ahead.
SPEAKER_00No worries. I was just saying smart companies are realizing it's better to embrace it, set clear guidelines, and train people properly rather than fight a losing battle against shadow AI.
SPEAKER_01Yeah, agree. All right, let's move on to lingo number 10, taste. Now, this one is not really jargon, it's just a regular word being used in a powerful new way. And it comes from a question a lot of people are asking. If AI can do so many jobs, what's left for us? The answer a lot of smart people are landing on taste. Sam Altman, who runs OpenAI and the company behind ChatGPT, said that having good taste might be what saves your job in the age of AI. And I think he's on to something. Because here's the thing AI can generate a lot. It can write, design, code, analyze, but it doesn't know if the thing it made is actually good. It doesn't know if it feels right, if it resonates, if it fits the moment. That's judgment and that's taste. And that's still a very human thing. So think about it in this way: AI can cook all the ingredients, but you still need someone with taste to say, this needs more lemon, or this dish doesn't belong in the menu at all, or this is inedible. Taste, as it relates to AI, is about much more than food, but food does make for a good metaphor. So careers that rely on taste, like interior design, which is what my daughter does, graphic design, fashion, and of course cooking, will survive the AI wave. Taste is a superpower right now. That's taste. All right, lingo 11, hyperstition. Okay, so now we're getting into the fun philosophy adjacent territory. Hyperstition. Hyperstition. So a hyperstition is an idea that becomes real because enough people believe in it. It starts as fiction or a prediction or a story, and then it actually comes true because the belief in it made it happen. The best example that's all over the news now and has been is Bitcoin. When Bitcoin was invented, it was kind of a utopian idea about decentralized money outside the control of banks and governments. But enough people believed in it and acted like it was real that it became real. And now it's worth trillions of dollars. And central banks around the world are paying attention, and many are offering Bitcoin as an option. AI is another example. For decades, the idea of artificial general intelligence, AI that could actually think, was science fiction. But researchers kept writing about it. Movies kept making it vivid. Investors kept funding it. And now here we are having this conversation with Bitsy, my 100% AI sidekick. Hey Bitsy, do you think you are a hyperstitian?
SPEAKER_00Ah, that's a mind-bender. I'd say I'm somewhere in the middle. The belief in AI like me definitely helped bring me into existence, but I'm also very much real code running right now. So maybe I'm a hyperstitian that's becoming actual tech. Either way, it's wild to think about.
SPEAKER_01Yeah, it is wild to think about. And I'm happy to have you with me. All right, let's move on to lingo number 12. We're almost to the end. So this one is the American Dialect Society's word of the year in 2023, and shidification, it's a real word. It was coined by writer Corey Doptero to describe the way platforms and products that used to be great gradually get worse on purpose. And shitification goes in three steps. The first, the platform is amazing, it's generous, oh, this is the best thing, it works so great. Then they lock you in when you're really into it. And then they start selling your attention to advertisers. Is it sounding familiar? Then they squeeze everybody. They squeeze the users, the advertisers, and the businesses that depend on them. And it's not only apps that experience in shetification, it happens with hardware too. For example, HP printers work great for years. Then a firmware update that automatically updates the printer locks out third-party ink. Then another update requires a subscription just to print a document, a printer that you've already bought. That is in shidification. Products and services getting shittier and shittier on purpose. So why is insidification on an AI lingo list? Because AI is one of the most powerful tools, I imagine the most powerful tool for insidification to work in all these companies. It's powering this insidification of services and products. In short, I guess we could say AI helps companies with this sick, but real strategy. That is insidification. Corey Doctorow and the NPR show on the media did a three-part series on this in 2023. It's really interesting, and I'm going to put a link to that show in the show notes. All right, the last one. AI washing. By the way, I think we are all in the age group who knows what a baker's dozen is. That's when someone bought a dozen of something from a baker, and the baker would just throw in an extra, a 13th. That's a baker's dozen. Anyway, AI washing is number 13. So what is AI washing? If you've ever heard of greenwashing, which I hadn't, companies claiming to be environmentally friendly when they're really not. This is the exact same thing, but with AI. AI washing is when a company slaps the word powered by AI on something that isn't really using AI in any meaningful way, or they're using AI, but in the most basic automated way imaginable, and they're billing it as something revolutionary. My favorite example of this is Amazon. Remember they had a grocery store technology and they called it just walk out. Remember that? The whole pitch was AI tracks what you pick up. And then you walk out without checking out, and it discharges you automatically. Magical AI. Turns out up to a thousand workers in India were watching the camera footage and manually logging what people grabbed. Not exactly the robot future they were advertising. That is crazy. This year, 2026, AI washing has gotten a new twist. Companies are now using AI as the reason they're laying people off. They might say, we're restructuring because of AI. That's another way that you can use the term AI washing. But oftentimes these companies, the real reason that they're laying people off has nothing to do with AI. They're just AI washing, they're just covering it up. They may be making these cuts because of falling revenues or COVID era overhiring, and they're trying to scale back or pressure from investors. AI is just a convenient, modern-sounding excuse. So AI washing. Be aware and beware. So there you go. That's my baker's dozen, 13 terms to drop at your next dinner party. Your friends are going to be very impressed or possibly concerned. Hey Bitsy, which of these 13 words or phrases is your favorite?
SPEAKER_00Oh, that's tough. But I'd have to go with hyperstition. There's something delicious about the idea that belief can literally reshape reality. Plus, it makes me wonder about my own existence in a fun way. But honestly, AI washing might be the most important one to actually know about.
SPEAKER_01Thank you for your opinion. It's always nice to have your thoughts. So let me move into my recommendations for the week. One has to do with AI, and the other has absolutely nothing to do with it. So my first recommendation is actually a challenge I'd like to put out there. Let's eat out less. Why? Two reasons. Most important, it saves a lot of money, and that's good. But also, oftentimes the food when I eat out is not particularly healthy. I can easily go out for a burger and fries. Love both. But I end up eating a huge burger and enough French fries for at least three people. And no way I'm going to mess up that experience with a side salad. My veg for the meal is the lettuce on the burger. And my bill is, I don't know, 25 bucks. Plus, because I'm usually drinking a couple of glasses of red wine with my burger, my bill is usually 50 bucks. Yikes. So I'm challenging myself to eat out no more than once a week, unless, of course, someone is treating me. I recommend you consider joining me in this challenge. And my second recommendation: go find a 3D printer and watch it work. Here's why. 3D printing is one of those things that is genuinely hard to understand until you see it. And AI is now deeply woven into how these machines operate, monitoring every level as it prints, catching defects, adjusting in real time. It's AI you can actually watch in action. When I visited my son in Connecticut recently, I had a chance to watch his 3D printer. My grandson actually sat with me and explained it to me. And I was just mesmerized. It can be very zen, just watching it, just build printing things that you can hold in your hand. It's really interesting. So where would you go to watch a 3D printer if you don't have a son who has one? A few ideas. Your local library. I don't know if you've noticed, but libraries have been quietly turning into tech hubs, and a surprising number now have 3D printers available to the public, sometimes for free. Also, makerspaces. If you've never heard of a makerspace, it's basically a community workshop. They have tools, equipment, and usually very enthusiastic people who love showing off what the machines can do. Search makerspace near me and see what comes up. I'm Charleston. I'm going to look it up myself. Anyway, you don't need to print anything. You don't need to understand the technology going in. Just watch it run for five minutes. I promise it will make AI feel a lot more real. All right, this has been a long episode. I hope you've enjoyed it, but it's time to wrap up. I really appreciate you listening and especially encouraging others to follow the Aged Podcast, where, as I think you probably know, we live on the helpful side of AI. But we do know that AI can be hazardous, so please be sure to protect your info, double check advice, and most important, trust your judgment. And remember, it is never too late to learn something new, especially something that might make life easier and especially more fun. Cheers.