EP 26 - 1st Anniversary Special - Year 1 in review and the educational and ethical considerations around AI-generated music and video.

In this episode, John and Jason talk IN PERSON, reflecting on year one of their podcast. Keeping with the theme, they also find a few rabbit holes to chase, consider developments in AI, and talk about educational and ethical considerations around AI-generated music and video. See complete notes and transcripts at www.onlinelearningpodcast.com

Join Our LinkedIn Group - *Online Learning Podcast (Also feel free to connect with John and Jason at LinkedIn too)*

Links and Resources:

• Hard Fork Podcast
• SORA OpenAI Video
• Alibaba EMO Video Demo (Jason’s LinkedIn post)
• Suno.ai
• Support Human Artists! Gangstagrass
• Mr. Beast on Youtube (not that he needs any more clicks)
• The makeup brush holder John keeps his pens in

Transcript

We use a combination of computer-generated transcriptions and human editing. Please check with the recorded file before quoting anything. Please check with us if you have any questions!

1 Year Anniversary Special

[00:00:00] Jason: Would you happen to have a pen I could borrow? Yeah.

[00:00:02] John: Felt blue, black.

[00:00:04] Jason: That is amazing. I've just this moment, I just noticed your incredible, your - you've got like a pen store.

[00:00:10] John: These are makeup brush holders.

[00:00:12] Jason: Oh really? Okay. Black, please.

[00:00:15] John: ballpoint, flare

[00:00:17] Jason: pen, Flare. Perfect.

[00:00:19] John: yeah

[00:00:19] Jason: And would you happen to have any sticky notes? That's incredible. You are really set up here. That is something else.

[00:00:24] John: I dream that someone, no one visits me. I'm set up for a full-on brainstorming session with a gigantic. Five feet by three-foot whiteboard and 500 colored sticky notes.

[00:00:34] Jason: Sticky notes galore.

[00:00:35] John: Yeah, I'm ready to change things if anybody wants to come over.

[00:00:38] John: I'm John Nash here in the same room with Jason Johnston.

[00:00:43] Jason: Hey John, hey everyone, and this is Online Learning in the Second Half, the Online Learning Podcast.

[00:00:48] John: Yeah, we're doing this podcast to let you in on a conversation we've been having for the last couple of years about online education. Look, online learning's had its chance to be great, and some of it is, but there's still a lot that quite isn't there. How are we going to get to the next stage, Jason?

[00:01:02] Jason: How about we create a podcast and talk about it?

[00:01:06] John: How about we do that? How about we create a podcast, do it for a year, and then talk about what that year was like?

[00:01:11] Jason: that sounds great! Happy anniversary, John!

[00:01:13] John: Happy anniversary, Jason.

[00:01:15] Jason: I should have brought you something.

I didn't. I'm sorry. How about we go out to lunch and we and we celebrate?

[00:01:20] John: yeah, and maybe we can get a demo of the Apple Vision.

[00:01:23] Jason: Oh, that'd be cool. Yeah. There's a little place right there where we can grab some lunch and maybe go over to the Apple store. See what's going on.

[00:01:30] John: Yeah,

[00:01:31] Jason: That would be thematic. A lot of this podcast has been a number of things. One, talking about online learning, but also talking about the new tech and how it might affect online learning in the last year.

[00:01:41] John: Yeah. We are EdTech nerds also.

[00:01:43] Jason: We are, we tend to nerd out on a few of these things.

Today on my way over here, because I had to drive to this podcast today.

I didn't do this podcast in my pajamas.

[00:01:54] John: Horrors. And you drove yourself.

You had to operate a machine to get here.

[00:01:59] Jason: But it gave me, afforded me a little bit of time in the car to listen to a podcast. I listened to our first episode. It was kind of nostalgic,

[00:02:06] John: you weren't tuning in to our first episode just out of some kind of vanity thing Oh, I love listening to me.

[00:02:12] Jason: No, it was not because I like the sound of my own voice. Although after doing a podcast for a year, you get used to it.

[00:02:18] John: you don't even know what you sound like. You're just like,

[00:02:20] Jason: I listened in because I was curious about what we talked about in our first podcast. Whether or not, what we talked about then rang true in our first year of podcasting and maybe looking ahead to see what's going to be different.

And what I found was, we basically talked about. What we were going to talk about, which was online learning, the second half, check. We've been talking about this last year. How technology affects online learning, check. We've definitely had a lot of that. We also had thought our big theme was going to be humanizing online learning.

Check. We've had a bunch of that. And also, however, one thing we had slightly wrong. What our topic of the month was, which was AI.

[00:03:03] John: Yes.

[00:03:04] Jason: It's become the topic of the year, probably.

[00:03:07] John: The topic of the Year .5 Yeah. So

[00:03:12] Jason: that's the one thing that we probably got wrong. The other thing that I would say that we didn't know about, as we couldn't quite see into the future with this, but one of the big things that you and I have talked about is how much we've enjoyed having guests.

We started this as a conversation between you and me. But how great it's been to bring other voices in this year.

[00:03:34] John: It has been remarkable to have other voices in. It's been amazing having guests because I feel as though it's a privilege that we get to have this kind of professional development that we create, I guess is how I look at it. And I think we do something for our guests too.

They feel good about being able to talk about their work, but the breadth and depth of the things we've talked about with some amazingly smart people has been just a privilege from me.

[00:04:01] Jason: Yeah, a privilege. That is a great way to put it. And just to be able to talk with some of these experts the last year to get a completely different for some of them anyways take on the things that we've been talking about has been challenging, informing, guiding for me so that we're not just talking in a vacuum here.

Really, our first guest was when we did the podcast Super Friends episode a little less than a year ago at OLC and we did another one just a few episodes ago to wrap up the year and then we had some amazing guests Dr. Michelle Miller Dr.

Enilda Romero Hall. Then we were able to talk to Dr Kristen DeCerbo from Khan Academy. And that continues to be a big thing out there. We made a great connection to OLC keynote speaker, Dr. Brandeis

Marshall.

Michelle Ament, Dr. Alicia Magruder at Johns Hopkins, which actually then led into a podcast recording at their symposium, which was so much fun as

[00:05:01] John: That was so fun and so innovative to be able to have a, almost a simulcast of the podcast as the concluding session of an online teaching symposium.

It has been good in that regard. And also, a chance to connect these ideas over time with of other things that come across our desk as it were. So, I think about Michelle Miller, and we keep talking about same side pedagogy.

that

keeps coming up as a relevant thing. Brandeis Marshall's notions of what's un AIable. I continue to talk about that even this morning with a provost from a two-year college in Texas was talking about this.

[00:05:41] Jason: You know what's cool? I was talking to somebody at UT the other day who has been listening to our podcast and he quoted Brandeis Marshall from our podcast about

[00:05:51] John: that That's fabulous. Yeah. And then. You know what I think surprised me the most over time is how certain things are emerging now, I think that are more important than anything else that's happened with AI in the last 12, 13 months, which is still the topic of ethics. And it's not about the technology. It's not about the advancements.

We're coming up in March of 2024. So, it's one year into the old March madness when GPT 4 came out and then I guess Anthropic came out, BARD, all of them were releasing and it was an arms race in March of 2023 to see what these models would look like. And now. We haven't seen in the last 12 months a massive boost in the model capabilities and a bigger discussion, I think that's happened over ethical use and the creation of guidelines, particularly in the education space.

[00:06:46] Jason: Yeah. When we recorded, we didn't even know of the existence of chat GPT four at that point when we recorded our first episode a year.

[00:06:54] John: ago.

No,

we did not.

[00:06:56] Jason: And so that just started that whole year of recognizing first that AI is a thing.

And then all of a sudden people realize, oh, wait, it's actually pretty significant thing. When that next model came out and realized that the real capabilities of AI were Much deeper, much better than what we expected, even on the front.

end.

[00:07:18] John: but the two guys that run the hard fork. podcast, were talking about how Sidney at the time, now all these name changes, but Sidney was Bing chat, which was Microsoft thing. It, it had told, it was a Kevin Roos or

was it Kevin Roos was advised by Sidney to break up with his wife and start dating Sidney.

You, similarly, dad your heart broken by Bing.

[00:07:43] Jason: Right. I'm being chat and had some very strange conversations with Sidney right in that same time. So, it was just the wild west in some ways that some of the initial concerns of AI kind of were tamed, I would say about those chatbots.

[00:08:00] John: Yeah. Yeah, they were.

But we were so amazed by the Model 3. 5 we couldn't stop talking about it. We thought we'd be done in a month.

[00:08:08] Jason: I would agree After that initial surge I think what we seen is a lot of third-party companies starting to leverage this power and I would say as we predicted A lot of edtech companies that were starting to add to it. And so, we talked about that. We predicted that back last year in March. And then as we were walking the floor and if you look back at our episode number nine, how are ed tech vendors humanizing online education? When we were walking the floor of O.L.C. Nashville at that point, that was March of last year. It was very different even as we were walking the floor in the fall.

Conference in terms of who, at least I found, who was already talking about.

[00:08:59] John: at

that

[00:09:00] Jason: point.

At that point, they were talking about it, not really implementing it, and we had some interesting kind of responses. And then by the fall they were really advertising ai,

[00:09:09] John: AI. In fact, the vendors, I think, that were concerned about having AI be part of their models were the ones that were trying to catch kids cheating.

Using AI, not thinking about how AI might be embedded into their tool to advance some feature that they wanted.

[00:09:24] Jason: Yeah, it was much more that concern. Yeah, it's interesting. Yeah, I think in the other part that I feel like what we're seeing more of lately in the arms race, and this is why some of our ethical conversations have taken this turn is the capabilities of AI beyond just the chatbot language model into the areas of media when it comes to video. One of the things we've seen in the last few weeks is OpenAI's Sora, S O R A, even just like yesterday, I saw that Alibaba, you know, which I don't even know what it, I've never bought anything from it, but it looks like a place that you can buy, cheap stuff on a wholesale kind of level.

They have a model that they're working on for lip syncing that's quite impressive. We can put a link to that model in the chat, but I feel like what we're seeing are these kind of video lipsyncing kind of ideas as well as if you think about what has happened in the last year in terms of image creation, how much better it's gotten.

And then even audio. I was doing a a few of these audio demos that are out there right now, one that's actually built into CoPilot that you can ask for it to make a song for you. it's, oh it's cool. It's pretty wild. And maybe we'll make a little clip in here.

Okay. Let's. I'll I'll quickly make something and then we'll take a listen to it and and maybe close out the show or something with it at the end. But yeah, you could just put in a prompt saying for it to make a theme song in this style using these lyrics if you want to, and then you can actually edit it edited it afterwards.

[00:11:06] John: Are you noticing the same thing I'm noticing too about the sort of seamless integration of generative AI into almost, I don't want to sound hyperbolic, but almost every app now that has been popular, has now decided to seamlessly integrate AI into itself, making its presence in operations that are not very transparent to the user. Or Notion, Copilot you name it, well Canva, they're all putting AI. Operations in Zoom. And I'm wondering if this sort of invisible AI is going to lull users into thinking that this is just part of the app, and it may not actually be AI.

I think about Zoom and it's a meeting summary feature. We were talking about this at our university in our policy group because I think a lot of people think, if zoom has this feature, then it must be okay to use. And then It's part of our acceptable use. Maybe it's inside our privacy guidelines, so I'm going to turn it on and we're going to use it, but that's not necessarily the case.

And so, if you're recording meetings or you're putting in student data or you're having, I don't know it's interesting to think about because I think it can enhance user experience, but I think you can also lull people into thinking that this is safe AI.

[00:12:19] Jason: Yeah, I guess using their brand acceptance, so we work at institutions that there's quite a vetting process to get something inside of our doors. So, we know that we're obviously working with Google, microsoft, and and zoom would probably for us and canvas all four of those. We're both of our institutions.

That's, those are the four biggies.

[00:12:40] John: Yes.

[00:12:41] Jason: And so, you're saying that it is almost like. It feels like if something comes in alongside of those packages or with those packages then it becomes all of a sudden to just accept it. It doesn't have to go through a, yeah, it doesn't have to go through a new vetting process.

If all of a sudden, a new, say there's a new video product and this is how we would get AI video summaries. This would have to go through a whole new vetting process, but we're not doing that. It's just just happening.

[00:13:08] John: Yeah. And so, if the underlying models are suspect at times, even, if we look at Gemini, Google's Gemini, and as we record this on March 1st, 2024, in the past seven to nine days, they had a major generative AI failure on their imaging model.

If those are the underlying engines, if you will, that are, adopted and licensed by these brands accepted tools. Yeah, how safe are things going to be? How do they, can Zoom stop OpenAI if they're using that engine? Can they can't really put new guardrails on top of what it does with the data because the model's the model, I'm not technical enough to know the answer to that, if I'm making sense.

[00:13:50] Jason: Yeah, you're making sense. I don't even know if Zoom is using OpenAI, and it, because it just appears, and I think we get a lot of wrappers around things as well, that are really OpenAI.

And then we get this new wrap around it and other things that are more like companies that are doing their own thing. So, it's hard to, yeah, it's really hard to track down.

[00:14:10] John: and, the question for me becomes even more important to discuss when we think about all the wrappers that have been created for P 12 teachers like Magic School, Diffit, a couple of others come to mind but I don't remember the names, but they're all also running on top of these models that are only as safe as they're made by those developers. so yeah, I think it's, I think it's something to talk about

[00:14:34] Jason: You know what's funny?

This this audio creation program. So, we got SORA by OpenAI, which is this brand new video. And then we got Suno, S U N O, with this audio that's coming in with copilot anyways.

[00:14:51] John: I didn't know it was inside Copilot. So what app are you using in Microsoft to get, to invoke Suno?

[00:14:57] Jason: of Copilot. So what app are you using in Microsoft to get, to invoke Suno?. What kind of style should we do today?

[00:15:20] John: We're sitting together in a room in Lexington, Kentucky. Can we do some bluegrass?

[00:15:25] Jason: Yeah. In a bluegrass style. Any other parameters we want to put on it? Maybe what do we want to have in the, what's really important to us? What do we our year in reflection song here, what do we want in the chorus to really hit home for the listener?

[00:15:42] John: That let's see. Human centeredness is the key. and ethics is important and learner outcomes are paramount.

[00:15:57] Jason: Okay.

Say in the course, make sure to include something about human centered online learning. And then I, I got caught.

[00:16:07] John: in your,

[00:16:07] Jason: your superlatives. What was the, what were the, what was the second one?

[00:16:10] John: Ethical use of AI.

[00:16:14] Jason: It should be, maybe we

[00:16:16] John: And belonging. Oh, I rented a Okay. Let's see.

You shorten your prompt to fit belonging in there?

[00:16:24] Jason: Yeah, I'll try to. Nice.

[00:16:25] John: Nice.

[00:16:27] Jason: Yep, okay, it's creating it. It's going to give me two versions and we can take a listen to both of these.

[00:16:32] John: Okay, excellent.

[00:16:33] Jason: We can talk about other things.

[00:16:34] John: Yeah, while it's cooking, yeah.

[00:16:35] Jason: to it. here's what's amazing. The first version is already ready. I thought it was going to take longer.

Now the second version is ready.

[00:16:45] John: Oh, okay.

[00:16:45] Jason: I'm not sure how we're gonna be able to listen to this just because of the current setup here.

[00:16:53] John: Let's see what happens.

[00:16:54] Jason: But we can put it.

Song “Keep on Learnin’ plays in a bluegrass style:

[Verse] Gather 'round, folks, and lend an ear There's a podcast here that we hold dear (oh-yeah) It's all about learnin', in an online way Discoverin' new knowledge every single day (ooh)

[Chorus] Human centered, always yearnin' For that ethically tech and belonging learnin' (learnin') Tune in and listen, don't you ever stray Online Learning Podcast, we're here to stay (heyy) (Join us now, keep on learnin') (Oh-yeah, yeah-yeah) Keep on learnin' (Oh-yeah, yeah-yeah) Keep on learnin' (Oh-yeah, yeah-yeah) Keep on learnin'

[00:17:00] John: ha ha ha…

[00:17:08] John: oh, a little Cher. What? You're the audio guy. What is that?

[00:17:12] Jason: it's like a little new, yeah, it's a, like a new bluegrass. it's

[00:17:17] John: it's a little country though. I think it's not quite.

[00:17:20] Jason: quite bluegrass, Yeah, it's not quite.

[00:17:21] John: but.

[00:17:22] Jason: Okay, that was a, so that was the first one. It's called keep on learning with a little apostrophe. Keep on learnin'.

[00:17:28] John: There's two people singing apparently in this, and there's someone who goes, "oh yeah."

[00:17:33] Jason: the things that impress me are a year ago, since this is a podcast and review a little bit, a year ago. Not even close, the things that were out there that you could create music and it sounded like a mishmash, like something that you would hear on like a Star Wars film that they're trying to make it sound different and spacey and non-human.

[00:17:56] John: Or it was the third or fourth duplication on your Maxell tape. Yes. Yeah. And it just degradated and degradated.

[00:18:06] Jason: So, first thing that impressed me is just where we've come in a year, the quality the second, the kind of the clever turnarounds on the lyrics. And then the third, adding pop elements that are very catchy for the listener, these kinds of echoes, as you said, and so on.

[00:18:26] John: Yeah, for The TikTok nation.

[00:18:28] Jason: The TikTok Nation.

[00:18:29] John: Yeah,

[00:18:30] Jason: Yeah, which is basically all of our listeners, right? TikTok nation.

[00:18:32] John: Basically, yes, that's right.

[00:18:34] Jason: Listen up, TikTok Nation. Is that how we should start our podcast?

[00:18:37] John: Maybe our podcast should be 60 seconds long if we want to, if we want to capture them.

[00:18:43] Jason: Okay, here's the second one. That was Keep on Learnin'. This is this is called Learning in Harmony. Uplifting folksy bluegrass.

[Verse] Well, gather 'round folks, I've got a story to tell 'Bout a podcast that's got a lot to propel Online Learning Podcast, it's the name Where knowledge and wisdom come together like a flame

[Chorus] In the world of bytes and screens, we find our way Human-centered online learning, come what may From the hills to the valleys, we all belong Ethical tech use, we'll sing this song

[00:18:50] Jason: not sure about the chord progressions in that one.

[00:18:53] John: More than I would about that. I would. And this is I put these out here with full understanding that part of my brain and heart is, " wow, this is so cool that technology can do this."

[00:19:05] Jason: Another part of me who, I've written a few songs in my life, and I enjoy playing guitar and there was probably even a moment that if the winds of success had taken me in direction, I would have done full time music. And it's both scary and a little offensive when I think on that side of it.

[00:19:22] John: Yeah. So, let's go to the offensive part because I think we're both having conversations with colleagues and I'm also seeing reports online of research on where instructional design is going with AI and how these tools SORA and others are putting.

Making graphic designers drone operators who do B roll feel a little at risk. And I think, I bet there's some offensive feelings there too about their art.

[00:19:48] Jason: Yeah. Actually, it's not completely true that I make 0 a year from my music. John, I've I'm raking in some Spotify money. I didn't know if you knew this or not. Yeah. It's I think I get like point zero. zero three cents per play and yeah, I think my last cash out was maybe around 2 or something. Yeah. So, I really am a professional musician, but I say that to say that This is not something I'm trying to make a livelihood from.

It also is not something that feeds my own sense of self worth at this point in my life.

[00:20:28] John: Yeah, but how would you feel if you were trying to make your livelihood from this?

[00:20:31] Jason: I think particularly I; I think it would depend on the person and what I was trying to do. But I would say almost every musician would feel. A little scathed by this because even if their livelihood is mostly playing live concerts, which this is not going.

[00:20:49] John: No.

[00:20:50] Jason: And developing a fan base, which this is not going to do part of your livelihood is getting yourself noticed in this enormous sea of other talent that's out there. And then also, I know people that are, they're singer songwriters is how they make their living. But it's great to get those what they call sync royalties when you get a song placed in a movie or a TV show.

[00:21:14] John: I was just thinking about that because I'm wondering what Hollywood will do with this capability. I think that Hollywood feels like they want to protect the rights and the livelihoods of artists writ large. So, they probably wouldn't do what I'm suggesting, but television production could decide to use Suno to do the theme songs for new TV shows.

I'm thinking about one of my favorite bands is Gangstagrass. They're a band.

[00:21:37] Jason: Oh yeah. I love them.

[00:21:38] John: Yeah, they blend, if folks don't know, they blend bluegrass and hip hop and they're amazing. They're amazing. I've seen them three times. They're coming to town here in Lexington soon. We're going to go see them. But my point here is that they became more famous because their music was used as the opening theme song for the television show Justified.

And if I wanted to do that again, if I were in production, could I just skip all that and just have a theme song written right off the bat from AI.

[00:22:06] Jason: Yeah, if you're looking for a particular kind of sound and that kind of mix, you wanted something a little gritty but Southern, but also urban, then that would do it.

And then, essentially, while I was talking, Suno was able to recreate our learning theme song in a bluegrass hip hop style, right? So, you think about how quickly this can happen at the capabilities that we have today. And this is, here's song number one.

Verse] Well, gather 'round folks, let me tell you a tale 'Bout a podcast that'll make you wanna prevail (oh yeah) With a blend of hip hop and old-time string We're gonna dive deep, learnin' ev'rything (ooh-yeah)

[Chorus] Human centered online learnin', take a seat on the track Ethical tech use, we ain't gonna lack Belongin' is the rhythm, that's our podcast groove Put your hands in the air, let the beat make you move

[00:22:34] Jason: And this is, song number two.

[Verse] Well, gather 'round now, y'all, let me tell you a tale 'Bout a podcast that's bridgin' the gap without fail It's online learnin', it's the way of the world With a touch of bluegrass and some beats that'll twirl

From the hills of Kentucky, to the streets of the city This podcast brings the vibes, all witty and gritty Talkin' 'bout human-centered online learning, y'all And ethical tech use, that's what we're all 'bout

[Chorus] Come on now, let's sing it loud and clear Human-centered learnin' and ethical tech use right here Belonging is the key, come join the crowd Discover new knowledge, sing it out proud (yeehaw)

[00:22:36] John: oh my.

[00:22:38] Jason: The second one particularly, I'm a fan of Gangstagrass. That second one particularly

[00:22:42] John: hit.

it, it approached it.

[00:22:44] Jason: Old school. Yes. Hip hpehop.

[00:22:45] John: but that first one, I don't wanna offend anybody. I don't know what that was. Was that some kind of Toby Keith kind of thing?

I'm. I'm out on that. That's but and that's funny how musical tastes run to also I'm not a big, like traditional country fan, like CMA style country, but I'll go to every Gangstagrass concert I can get my hands on. But you're right. The second one approached it, but still, and then I started thinking about cultural appropriation and what is this?

Yeah. This is AI's attempt at understanding culture, which is, that's risky. Yeah,

[00:23:16] Jason: Yeah, we got yes, tricky waters right there.

[00:23:19] John: Incredibly tricky.

[00:23:21] Jason: so, we've

Talked about just ethically doing this in light of the musicians themselves, but I'm watching I'm a big jazz fan as well. I like a lot of different kinds of music, but I'm a big jazz fan. So, I'm watching the Ken Burns series on jazz, which I highly recommend. It's slow. It's long, but it's beautiful. But how many times have we taken an art form as a dominant white race from another people group and then appropriated it because we figured out how we could monetize different way. Or in this kind of case, how can we non monetize it? So, we're maybe they're not even making money off of this song. So maybe these aren't going to show up on iTunes. Cause I know iTunes has made some rules about this.

YouTube has now made some rules about this, but maybe they'll show up in the next ad for whatever, and they've made it for free.

So basically, the Suno terms of agreements is that if you pay for it, you have full mechanical rights to these songs.

[00:24:25] John: So, if I make a Suno song, were you logged into your University of Tennessee controlled garden of this? So, if I make a Suno song inside my University of Kentucky controlled garden of Co Pilot, does the University of Kentucky own the, that song?

[00:24:40] Jason: That's getting into the whole intellectual property end of things.

That's a whole They They have the mechanical rights to this really crappy culturally appropriated piece of junk that I created.

[00:24:51] John: And you're right, but look, how much of advertising now... I'm shocked now I've cut the cord on my television and whenever I accidentally happen to go back onto watching network TV or watch my local news.

I'm shocked and also simultaneously not shocked that the insipid advertising that I grew up with in the 70s really hasn't changed much. So, your comment about Madison Avenue using tools like this to create jingles and other things to cut out artists for their clients. Absolutely.

I bet they'll do it. I'm very cynical about this. I think it'll, yeah, I think that's where this is going.

[00:25:26] Jason: And you talked about networks and maybe some of the big ones will, for the sake of their already large group of customers, perhaps they'll make some rules about this to please people, but the networks are not just competing with other networks.

They're competing with Mr. Beast.

[00:25:43] John: Yes, they are.

Yes,

[00:25:45] Jason: Like Mr. Beast is enormous. And he has a enormous viewership, and my guess is that he probably, his income per year probably rivals some of these smaller, if not networks, maybe some of these smaller production houses for sure.

And I only know about Mr. Beast because I have teenage kids who drive these whole things. One, one of my kids particularly. And also, Dude Perfect they're not utilizing traditional streams, and so they're not going to be beholden to these kinds of larger ethical kind of, restrictions.

[00:26:18] John: Now, Mr. Beast is for folks who don't know, what would you, how would you describe him? He's an internet creator. I'm logging on to Variety. com. His annual earnings hit 82 million dollars last year, more than double any other digital creator and, and it's also, it's funny, his name, Mr. Beast sounds for those who aren't in the know, like some kind of awful weird guy, but he's just this, it's just this young guy, right?

[00:26:44] Jason: Yep, he seems to be, like, who knows, I've listened to some other podcasts that talk about him and so on, and actually even the Hard Fork that we mentioned, I think they talk about him one time-- his kind of use of YouTube who knows what all his motivations are, regardless, he does give away a lot of things, and he seems to be fairly kind to people in

that

[00:27:00] John: in that way.

His real name is Jimmy Donaldson, for the

[00:27:03] Jason: Oh yeah, yeah, of course I know that I've, I follow him on LinkedIn,

[00:27:06] John: oh, you're going to be a gigantic creator on LinkedIn now with the beast.

[00:27:11] Jason: Our connection is pending, is pending, so yeah, remarkable. My kids watched Rhett and Link throughout, do your kids watch Rhett and Link?

[00:27:20] John: Okay they're at 35 million, second place, but but they're 50 million away from Mr. Beast.

[00:27:24] Jason: Yeah, that's wild. I think that points to the fact that ethics is a huge topic right now and our one of our last podcasts was about this We can't rely on the companies coming up with the ethics to guide.

[00:27:38] John: No.

[00:27:39] Jason: partly because it won't be Comprehensive enough, it's one thing if Apple comes up with some ethics or Microsoft.

But not everybody's gonna abide by these rules, and there's gonna be so many startups that would,

[00:27:54] John: just Mm hmm.

[00:27:55] Jason: do an end run around any of these kind of companies to get a few more views.

[00:28:00] John: Yeah. I think that as we talked about in that episode on ethics, I think we've got two sets of ethical books going one by the companies to be sure that they can sell as well as possible. So, I'm calling those the kind of less, less ethical set of books. And there's a public persona of wanting to be safe. And so, the, they put in enough guardrails through their red teaming and things like that. So, we can't get instructions to do awful things, but then they stop right there. After that you're on your

own.

[00:28:28] Jason: Yeah, and depending on what AI you use, and you can always find one that can do what you want it to do.

[00:28:33] John: that's right. Or you download your own LLM, you get a llama and run it on your own. And then you can, there's no guardrails, no red teaming.

[00:28:41] Jason: It's crazy. I had a little bit of space this week to go follow some rabbit trails and one of them was looking at Hugging Face, trying to understand a little bit about what this is all about. And it's a place where you can actually download models. So, you talked about this one model. But have you been on here? Should I ask the question?

[00:28:59] John: Should

I

quiz you on

this?

No do not quiz me on this.

[00:29:02] Jason: Those listening, I won't quiz John on this because, it's.

It's hard not to be in the know sometimes about, a Hugging Face. I didn't know, I had no idea that this was going on.

[00:29:14] John: I just want to say that I'm comfortable being in the dark around.

you. Because you're kind to me, in

[00:29:19] Jason: to me. Oh good, that's great. And I put this out here to say I'm oblivious and I don't really understand all the implications of this.

However, right now, on Hugging Face, which is more of an open-source AI model arbitrator almost, there are currently, and I'll take a pause here, podcast listeners, guess, podcast listeners, to yourself, or to somebody you're listening with, maybe say it out loud, how many models do you think there are right now to download on Hugging Face LLM models.

[00:29:49] John: Okay. And while people are thinking about that, and I will too. So, what you're saying is that how you Hugging Face is actually sounds like it's a kind of a marketplace for large language models that like, or you make your own sort of, I'm air quoting "GPTs" and then you can go get one and download it and run it yourself.

[00:30:06] Jason: Yes. I would call it more of a GitHub.

[00:30:09] John: a

[00:30:09] Jason: a marketplace. I didn't say anything for

sale.

No.

And so, you create, it feels like GitHub when you get there. Where you can do different forks of different

[00:30:17] John: LLMs

and on this LLM landscape inside Hugging Face Are they, do they have special purposes, some of these, so they're in that way. They're like like the GPTs that you could make for

[00:30:28] Jason: Exactly.

Okay.

So, all these would have different purposes. So, this isn't, aren't like the big models we're talking about. Many of them are leveraging these big models.

[00:30:36] John: Okay, cool.

[00:30:36] Jason: these are GPTs. Many of them that you can download and use. Most of them that you can use on your own computer. Your own home computer. Okay.

[00:30:45] John: All right. So how many are there out there.

[00:30:47] Jason: Right now, as of today March 1st, 2024, and this will change. Currently there are 531, 270 thousand that one could download.

[00:31:02] John: Little large language models, little AIs

Yep.

That I can then pull onto my hard drive and never have to get on the internet and ask it anything. I darn well, please.

[00:31:13] Jason: Exactly. Yeah. If we're gonna we were talking about our one-year retrospective. Some of our predictions about what we were going to talk about last year were true.

We thought, we didn't know we were going to be talking about AI for this long, and it would move this quickly, was one of the, one of the differences from last year of doing this podcast. Here's what I think with all these creative elements, that it's going to start by some professors thinking "I don't need a production company to help me do these things and they're going to create maybe just for fun at the beginning a theme song for the class or a video of them teaching the class in Mandarin or or the class being taught by some historical character with their voice to it, or, using some images in their slides, which is already happening, right?

And that are created. And at first, it's going to be a little gimmicky, and then we're going to cross a threshold where it, A, is no longer gimmicky, and B, it actually starts to affect workflow and the people that we use for doing this work, particularly at large institutions. What do you think of that kind of prediction?

[00:32:27] John: I don't know. We'll have to see. I think based upon some of the surveying I'm doing before I go talk with groups about whether or not they've ever even used a large language model, used chat GPT, for 50%. routinely state that they either have never used it in their lives or have used it once or twice ever in the time since it's come out.

[00:32:51] Jason: So that's over.a year.

[00:32:52] John: And so if half of our educators out there are in that space, then I don't think that they're going to be using these models in any way in a deliberate way to advance their teaching and learning goals, and they'll be using them however, the platforms like we talked about before, as platforms start to integrate these tools into them, that's how they'll get used.

[00:33:14] Jason: Yeah. I think you're right. Yeah. The average professor, I agree, is not going to be going into Hugging Face and probably downloading and creating.

[00:33:22] John: I was just going to say the same thing. I'm crazy enough to go do that. But no. Yes. No, but no one else is. Not no one else's, but I don't know anybody, but maybe you in my circle of friends and colleagues that would do that.

[00:33:35] Jason: Yeah. Yeah. And even I don't think I would use it. I would use it out of curiosity to see what's going on and so that I can understand it's actually, this may seem strange, but my tinkering is actually a leadership mechanism for myself.

I think part of my job is to be able to see down the road a little bit. And to be able to anticipate it and figure out how we're going to react and how we're going to guide this whole thing.

[00:34:04] John: I agree with you 100%. And actually, I coined a little term in the bootcamp that I'm doing now for AI, but "you have to try AI before you guide AI.

[00:34:14] Jason: That's good.

[00:34:15] John: Because how can you talk about the direction in your institution, organization, department, unless you've tried it out yourself and can talk about what you know of its ramifications or even how you feel about it.

[00:34:27] Jason: Yeah. Yeah. Unless you understand where it's at, where the power of it is at right now then yeah, your ethical guidelines are going to be all over the place.

You're not going to really be able to hit, especially when, as we talked about in our previous episode especially when we're talking about contextual ethical guidelines. So that really have some teeth and examples to them.

[00:34:49] John: That's the key or the contextual guidelines because in our institutions, our universities have broad guidelines, but what happens once the classroom door is closed? Completely different matter.

[00:35:00] Jason: So, I agree with you about, Hugging Face. No, we're going to get a few crazy people like us poking around with this stuff. Now, Canva, though, think about how many teachers are using Canva. You can get an educational free license that gives you extra stuff and templates are there and it's, crazy what it can do.

They have AI baked right into it, right? Zoom. It's baked right into it. If you have an Adobe license now, firefly is baked right into it. This week there, they have a music production thing that's starting, they're starting to demo these are the, I think those are the places that it's going to sneak up on us in ways that it's not going to go through a regulatory body of whether or not we can use this software.

It's going to be on our computers with the next update.

[00:35:43] John: Yeah, no that's my point entirely that this will just become embedded in the ways of working throughout. I was talking with some folks from a two-year college in Texas who have 94% of their graduates either go on to a four-year institution or go into the world of work.

And so, they've got an academic side, and they give an AA associate degree. But then they've got, the welding and the HVAC and all, and they're painfully aware that AI is going to be embedded across both those paths for their graduates. And how is that going to look? And how should they be thinking about preparing folks for the world of work?

[00:36:19] Jason:  Yes. And I think that part of our jobs as educators, of course, as we've talked about is not just disseminated knowledge, right?

It is preparing students for, not just a vocational life, but a life ahead of them, right? I think this is part of our mandate is that we are forming students at significant times in their life, whether they're just coming to college for the first time and they're 18, 19 years old, or if they're adults and they're coming back to college and trying to re-equip themselves for the life ahead of them.

These are significant times in these people's lives. And we owe it to them to prepare them for the world that is out there right now and the world that is.

coming

[00:37:00] John: yeah. Agreed. Agreed. I can't do better than that.

[00:37:04] Jason: On that note, thank you for listening, everybody. And And John, thank you for for having me. This is a good thing about doing it face to face is that you know that there's a real person at the other end of the

[00:37:15] John: the other end. Yes. Yes. It is better than zoom in a lot of ways.

I can put my hand up and tell you to, I want to say something and then we can, yeah, it was pretty good.

[00:37:25] Jason: Yep. And you can pass me a pen out of your enormous collection.

[00:37:29] John: I have an enormous collection of pens inside makeup brush holders, Lucite. They're beautiful. Cool.

[00:37:36] Jason: thanks for listening to everybody. You can get the show notes and thank you everybody that has listened and commented and encouraged us since last year.

It's been exciting to be part of this and that excitement partly comes from you. If people weren't listening, actually, we might keep doing this,

[00:37:52] John: actually. we might

[00:37:55] Jason: about us

[00:37:56] John: thing about our lives. That is a,

[00:37:58] Jason: I don't know. What do you think, John? If nobody was listening, would we keep doing it?

[00:38:01] John: maybe Maybe. not. Yeah.

[00:38:04] Jason: Maybe.

Maybe for

[00:38:04] John: a little while. Maybe for a little

[00:38:06] Jason: we'd just get rid of the microphones and just have a conversation.

[00:38:07] John: yeah,

how do we even know if anybody's listening? What's our threshold for anybody's listening?

But if you get a chance, we would love it if you'd go out there and give us a rating and leave us a comment. We like to feed the algo as we say, but it helps us know that you're out there and it helps us get out to more people. So yeah, leave us a rating and a comment and we'll get back to you too.

[00:38:28] Jason: yeah, absolutely. We do get back to you. It won't be AI and join us on LinkedIn. We got a little community there as well as you can just find us there to message with and to see other posts. And I'm gonna say John, as John's a good one to follow on LinkedIn. He's creating some. incredible content these days.

A lot of it around these conversations. So, I would highly recommend at least to go on to LinkedIn and follow John. If you don't want to follow me, that's fine. But at least follow John because he's got some good stuff.

[00:38:55] John: I recommend you follow Jason as well because he goes into more rabbit holes than I do, so I think that, and they're illuminating.

[00:39:02] Jason: Yeah. I don't know about illuminating, but I definitely have some rabbit holes. Mine tend to be less structured and thought out. It's just like what I'm thinking about in that moment. And I post off of my phone something. and at that, Happy anniversary, John. This has been great.

John: Happy anniversary, Jason.

(bluegrass style AI created song outro)

Talkin' 'bout human-centered online learning, y'all And ethical tech use, that's what we're all 'bout

[Chorus] Come on now, let's sing it loud and clear Human-centered learnin' and ethical tech use right here Belonging is the key, come join the crowd Discover new knowledge, sing it out proud (yeehaw)