Pre-Prompting AI with Tim Desmond

Download MP3

AVFL Tim Desmond
===

[00:00:00] Welcome to a Vision for Learning, a proud member of the Be Podcast Network. The best podcasts out there. I am excited to have, oh, I'm Jethro the host. Welcome. Glad to have you here. Uh, you can find me at Jethro Jones on all the socials.

I. And I'm excited to have on the show today, Tim Desmond, who's a career educator and entrepreneur and writer with more than 25 years of experience educ innovating in education. He's the owner of Thorn Publishing LLC, and the founder of New Educator ai, which is dedicated to preparing educational leaders for the AI era, ensuring future generations are adept in using AI safely and responsibly.

Tim, welcome to a Vision for Learning.

Thanks, Jethro. Excited to be here.

Great to have you. Um, so what is most valuable from our conversation today?

I think the most valuable thing if you're listening, is that, uh, we are all responsible for being proactive right now. Uh, whatever we think of generative AI [00:01:00] in education in the classroom, I think a lot of people will agree things are gonna change. Uh, some for better and some for worse. But, uh, the only way we can help steward our kids through this is to be actively involved in it.

Yeah, I, I agree. I think that was, was so valuable. Uh, one of the things that I would say is a good takeaway from our conversation is talking about. Uh, the skills you actually need to work with AI effectively right now. And we didn't like list out here's A, B, C, and D, but if you're listening, you'll hear these are the things that you need to be able to do, and I think that is incredibly valuable.

So, uh, thank you Tim. Appreciate you being here. Uh, we'll get to our interview with Tim in just a moment.

​ ( ad here)

All All right, Tim, I think a good place to start is talking about, uh, new educator AI and what you're doing there now, uh, to set the stage, that's a substack. People can get it in their email. They go to new [00:02:00] educator.ai, and then they subscribe. Most of what you've got on there are, uh, AI generated summaries of interviews and, and things like that.

Uh, tell me your thought process behind that and, and why you started that and what, what you're doing with that.

Yeah, absolutely. Um, thrilled, thrilled to talk about it. One of the benefits of getting old is you start to see a lot of stuff

Yeah.

to, and you start to see things multiple times, and I. What's, what's been happening in the world of generative ai and especially as it pertains to education is sort of another go in the cycle.

And I can remember back in the first cycle I was exposed to in the early to mid nineties was the internet. uh, you know, if we think back to that time, there was a lot of consternation, a lot of, uh, concern. In addition to excitement and, uh, and, and that kind of energy is kind of back with ai. And what I realized was that I had a [00:03:00] perspective that could be helpful to educational leadership. So I decided that I was gonna start a newsletter because I'm, uh, I'm a writer by profession and that that's sort of my, my sweet spot. And I decided I was gonna start writing. I think there is one interview. Mostly though they're sort of long form articles and research pieces in, uh, pertaining to how AI is potentially going to affect education. And I know enough about education to know that classroom teachers are not always the ones, uh, who are first looking for this information because they are just overwhelmed with the responsibilities of being a teacher. So I'm sort of targeting more educational leadership, whether that be. Principles, board members, division heads, uh, and sort of just setting the stage for, you know, here are some of the things that are coming down the line.

Here's what's possible. And I would say that the underlying assumption for new educator AI and my own perspective is that it [00:04:00] isn't something we can ignore. Uh, we can be afraid of it, we can be excited by it. Uh, we can choose to do or not to do things with it, but I don't think it's something we can ignore right now.

I, I agree with you, and I think that it's, it's easy to. Ignore it and it's easy to think this isn't really gonna affect me. Uh, Aaron Mackey, who I've had on the show and is, is doing a lot of cool stuff. You should definitely interview him for your substack. He's got a lot of interesting ideas. One of the, the key things that he says is that if you're not using it on a regular basis, you're basically just gonna get left behind and.

And, and when you talk about this cycle and how like the energy is around this thing, um, that's definitely what I have felt as well. And, and I have, I have not been excited about technology in education for quite some time because there just hasn't been anything really. [00:05:00] Intriguing happening. Now, let me, let me share another perspective on this.

Here's a stance that I take that if you are, if you are not going to take. Take the time to, uh, write or create something yourself and you're gonna leave it all in the hands of ai, then I'm probably not going to take the time to read or observe or view or whatever it is that you are creating. And so how do you, how do you square with that, with what you're doing with new educator ai?

Tell, talk to me about that.

Yeah, I, I rely heavily on both, uh, Claude Opus and Chat GT four as the time of the, of this recording, uh, in the writing process. And I think you're. There, there's something called push button creativity that I found myself saying when I, when I speak in front of groups, which is if you can sit down and push a button and [00:06:00] AI will pump out an, an essay, a song, a video, I'm not interested in that. Uh, I think if there, if there's sort of no creative direction, if there's no ownership of the idea itself, then I agree with you. I, I'm, I'm not gonna bother reading it. And so the way I've been, uh. Trying to frame it for folks mostly in the publishing industry, because I do a lot of author workshops and conferences, and as you can imagine, those people are very sensitive to protecting the writing process, uh, you know, for for good reason.

And they have, you know, serious, uh, you know, reservations about it. And what I explained to them is have to think of yourself more as a director versus, um. Um, a wordsmith, right? Um, we are now to the point where generating the words themselves, that doesn't hold any value. Whether that's an ai, whether that's a, a ghost writer, an assistant, like you could, you could hire anyone to write you an essay.

You could, you [00:07:00] know, hire someone to write your, you know, AP history essay, and it, it doesn't mean it's gonna be any good. So I think as the creator, as the author, and what I try to model with. With new educator AI is I try to bring a set of ideas, a perspective, and then I direct the AI in helping me craft that essay.

So I'll give you an example of one that's coming up. had this, this connection. Again, I think this is, this comes with age where I recognize a lot of the same arguments that were levied against Wikipedia in the early two thousands are being against AI in education right now. What I realized was at the time, educators were concerned about, um, the validity of Wikipedia.

Could you trust it? Who's writing this stuff? Is it going to eliminate the need for kids to do research? Like these are all things that are being said about generative AI right now. And I had this moment where I was like, oh yeah, I [00:08:00] remember those arguments. I remember those concerns. And not that they've all gone away, but I think for the most part. use Wikipedia for, for a lot of stuff, and they, they don't necessarily question the veracity of it. Like, I think it's well accepted. This sort of decentralized approach to a knowledge base is, is common now. So

And, and beneficial.

Right.

Yeah.

Yeah. So, so I, I, I think my point is if, if I, if I would sit down and. At, at say chat GPT and say, write me an essay about generative and AI education. I'm gonna get a very bland, vanilla aggregate response based on its own dataset. That's very different than if I come and say, you know what? I want to, I wanna write an article comparing Wikipedia to generative ai, pros and cons, differences, similarities, and I start there.

I think it's a very different approach.

Yeah. One of the things that I've found is that if I infuse it with my opinion first [00:09:00] and then have it do the writing generation, I, I love what you said about being a director rather than, um, I. Whatever a com a comparison could be. I don't remember what you said, but you said being a director.

Yeah.

Yeah, so being, being a director I think is really powerful because when I do that with anything that I'm writing and I say, here are the points I wanna make, here are the things I wanna say, then.

The writing is in, in my opinion, exponentially better than what it would've been otherwise. And so the, the thing that I'm getting is I, I feel like it is really easy to get AI to do anything, but it's really difficult to get it. To do something very specific, to be in my voice and talk the way that I would talk and that kind of stuff is, is much more challenging.

And even with all the content that I have out there that I've tried training it on, I still find that it is, it is not that good or [00:10:00] maybe I'm not that good at training yet, but it, it still leaves a lot to be desired. Uh, what are your thoughts on that?

I, I would agree with you. I think how I, I would put a caveat on that, and, and I know you know this, but I think a lot of, a lot of people don't really understand. The nature of exponential growth, this isn't like, this isn't like a new Microsoft Office product. It's not a new version of Word where Microsoft ships the application and then it remains static for a year until they upgrade it. This, this thing is changing and learning and growing and getting, getting better all the time. So I would agree with you. I think right now I can, I can preload, I can pre-Pro, you want to call it, with my own writing. And what I get out of it might be 85, 90% close to what I would've written on my own, but it's not a hundred percent. Uh, I think that's gonna change, and I think it's gonna change faster than we realize.

Yeah. Yep. [00:11:00] So talk about that pre prompting piece and the importance of prompt generation. Uh. During 2023, that's what everybody was talking about. And now it seems like that has died off because people have made, uh, custom, uh, agents and apps that you can, that you can use and they do that stuff for you.

Um, where, where are you at on the prompt engineering part of it now?

Yeah, I, I agree. I don't think that's a, I don't think that's a long-term, uh, career strategy.

Yeah.

you know, so I think if you're a college kid right now and you want to go into prompt engineering, I don't think that field will be there by the time you graduate. Uh, because the AI is getting so good, it's gonna start to intuit.

I. What you want, like you, you know, so this idea of prompting it, uh, I think it's a short term skillset, but it is, it is a valuable one. And what I've found for myself, uh, you know, I, I've been a, a writer and publisher for the past seven years, and I also have a lot of material that I've written and, [00:12:00] uh, what I find is that I'm still better off doing a custom. Pre-Pro, for lack of a better term, for every piece, because something I might like, if I'm gonna write fiction is Jay Thorn. That tone and that voice and that writing is gonna be very different than what I use for new educator ai. So some of the pro, some of the platforms like chat, GPT you the ability to put a, uh, a pre-pro in front of every prompt automatically.

And I, I don't. I don't use that, uh, because I want it, I want it specialized. And even if I'm writing nonfiction, if I'm writing a new educator AI article versus say Creative AI Digest, which was another newsletter that I hadn't since sold, um, those are very different even though they're in the same industry.

And so I would, uh. I would, again, start with a, I would start with a sample of whatever the style or type of writing that I want to do a sample of my own, not, not of anyone else's. And I, that would become my pre-pro. And I would [00:13:00] simply say something like, this is the, this is what the, the output should look like in terms of author, voice, and style.

So I, I do think that's a, that's a very valuable thing. I think it's the number one thing that who are new to AI overlook, they come to it. And they use the LLMs or the, the large language models, like a search engine, it's not a search engine. I, I think the more you can give it upfront, the better you're gonna get on, on, on the output. I.

Yeah, that that's so true and it's so fascinating to think about it like a search engine and. How, yeah, you can get information from it, but it's not, that's not really what it's there for. And, and that's, that's interesting. I've been using the ARC browser to do, to browse for me on, uh, when I want to search things.

Oh my goodness. That is such a better experience than Google. I mean, Google has just destroyed itself over the past [00:14:00] few years by ranking all these things that are just ridiculous and pointless. And I, I pretty much can't trust Google to go to, to, to get the answer that I want. I can trust Google to give me, uh, clickbait search results that, you know, is 3000 words of nothing.

To get to the answer that I really want, the arc browser, I just type in something that I'm interested in and then boom, it, uh, it gives me the answer and, and browses the websites, gives me the references. Like it's just phenomenal. And that kind of what you mentioned before, this idea of intuiting what you want, that is exactly what it's doing so well, and I'm sure at some point they're gonna have to figure out how to monetize and bring ads in and, and.

You know, add the complexity there that makes it stink in the end. But for this time period, man, that is just, it's a delight to use and it feels so [00:15:00] good to do it. And, uh, if you, if you haven't downloaded that app on your phone and use that, uh, to just use as your Google search when you wanna look information up, man, it's, it's powerful.

Yeah, I haven't, I'm, I'm gonna, I'm gonna have to do that when, when we're done here.

Yeah, it's, it's cool.

more point on the, the search engine mentality. is, it's, it's a bit of a paradigm shift and I think it's something that when I share this in with a room full of authors or people, I can see light bulbs going on. And here's the example I give. I said, instead of asking, instead of your first, uh, prompt. Asking the, uh, you know, chat GPT to do something. it what it needs to accomplish the task that you want done. So for example, instead of saying, write me an essay on the causes of the American Civil War. You know, I know it, it's gonna pump out, you know, that pretty vanilla of the road response, right?

Instead you ask it, you, you say, I need to write an [00:16:00] essay on the causes of the US Civil War. What information do you need to help me write that?

Yeah.

you just do that little thing, whether you're a student, the teacher, whether you're looking to. Plan a birthday party, you know, and instead of plan me a birthday party, like what information do you need to plan a birthday party for a 9-year-old girl?

You know, something like that. The, the, the results are exponentially better.

Oh yeah, for sure. I ran into that accidentally one time and I was like, I, I asked it to do something and then it said, okay, I need all this information. And I was like, oh.

I.

That's great. It knows what it needs and will ask for it if it needs it. Whereas previous to that, all my response or all my prompts had ended up with it.

Just making up stuff and saying, here's, you know, here's what you actually need for this, and inserting things that weren't there. And when I discovered that, I was like, oh. Duh. Like, [00:17:00] why don't I start with a conversation? And, and that piece of having a conversation with it is really powerful also, because that, um, that takes away the, the idea of you just copying and pasting whatever it creates and using it more as a brainstorming tool as well.

Um, so, uh. I think the last question that I want to ask, I can't believe how fast time goes by when we have these conversations. This is so good. Um, so what, what do you see as the future for this in, in the short term and in the longer term? Like, how is this going to make a real change in how we learn? Uh, not just individually, but also in schools as well?

I.

Yeah, Jethro, I, I need to be completely honest with you and, and your listeners in response to this question, I think. It's gonna be a world of pain in the short term. [00:18:00] Uh, I really do. Um, uh, you know, I think there, uh, many, many schools, many, uh, of the people in leadership positions have just tried to lock it down or ban it for legitimate reasons. Um, and, and I think. That's caused some pain. I think the bigger pain is when it starts to replace some of those quote unquote unreplaceable skills that teachers have honed for, for decades. Uh, I, you know, I, we, that's all what those are is a whole nother conversation, but in the short term, I. I think there are gonna be aspects of teaching that will, that will disappear because the AI will be, will be better at it term. I have a much more, uh, optimistic outlook. I think what we're gonna end up, uh, and I have no idea how long this will take or what it will look like, but I can see every student. their own sort of personalized learning tutor, uh, whatever that means in whatever subject that happens to be and what whatever discipline or project that happens to [00:19:00] be. I think it's also gonna democratize education in a way that we've never seen before, uh, across race and socioeconomic, uh, status. Um, you know, I think we're gonna empower kids. the most powerful learning tool that's ever existed. Uh, but to get to that point, I think we're gonna have to go through some pain.

(ad here) Yeah, so here's the really amazing thing you can take. You can essentially use this to learn anything. Very simply, you find a video on YouTube. I. You watch somebody teach you something directly and there's billions of videos out there where you can learn stuff, and then when you get to something you don't understand, you go to an AI tool and you ask it to help you explain that so that you can understand that piece.

And so one of the, the, the short term and the long term things that I think are going to be damaging, [00:20:00] um, and challenging is that. People, well, Aaron, who I mentioned before, he calls it a binge box, that we should have a binge box where we put all of our stuff that we have created and, and things that tell people about us, like job interviews and stuff like that, like direct people to your binge box.

And, and I love that idea because I. What I think the challenging part is going to be is how are you going to keep track of what you know and are able to do yourself when, um, when you don't need. A class or a a, a certificate to show that you can do that because the need for those things. I mean, we know already that grades, certificates, all that stuff, they're just made up anyway, uh, uh, accreditation.

I. Uh, by organizations for schools, it's all made up. It's just somebody saying, you're accredited because we say you're accredited. That's all there is to it. And you jump through [00:21:00] our hoops and then we say you're accredited. And so like those things I think are going to, I. Go away and it's gonna be more on what can you show us that you can do yourself.

That's where we're going to find the value. And if you are not using these AI tools to help you with that, you're just going to get beat by other people across the board. And the, the list of things that you can do without AI are continually shrinking. And you know, that that's a real challenge. So, um, I think we need to do another show and talk again about the, uh, the aspects of teaching that will disappear.

Yeah.

sounds really fascinating and I am definitely looking forward to talking to you more about that and other things. So, are you, are you cool to come on again?

I would love it. Absolutely. I.

All right. So, uh, once again, uh, you can find him a new educator.ai and you can check out his creative AI digest, which you sold. It's not [00:22:00] yours anymore.

Right.

not mine anymore. Yeah,

Okay. Still good stuff, but, but not yours. Alright, well thanks so much, Tim for being here. Any other way you want people to connect with you.

No, that's great. Uh, and I'm on LinkedIn, uh, but if they go through new educator ai, that's, that's where all the valuable stuff is. So

Cool. Alright.

uh, the conversation.

Yeah, me too.

Pre-Prompting AI with Tim Desmond