Episode 67: ChatGPT: What Does it Mean for ABA? with Amanda Ralston

Building Better Businesses in ABA is edited and produced by KJ Herodirt Productions

Intro/outro Music Credit: song "Tailor Made" by Yari and bensound.com

Give us a rating at Apple Music, Spotify or your favorite podcast channel:

Apple: https://podcasts.apple.com/za/podcast/building-better-businesses-in-aba/id1603909082

Spotify: https://open.spotify.com/show/0H5LzHYPKq5Qnmsue9HTwn

Check out Element RCM to learn more about billing & insurance support for Applied Behavior Analysis providers

Web: https://elementrcm.ai/

LinkedIn: https://www.linkedin.com/company/element-rcm

Instagram: https://www.instagram.com/elementrcm/

Follow the Pod:

Web: https://elementrcm.ai/building-better-businesses-in-aba/

LinkedIn: https://www.instagram.com/buildingbetterbusinessesaba/

Instagram: https://www.instagram.com/buildingbetterbusinessesa...

Transcript

Jonathan: 0:30

My guest today is Mandy Ralston. Mandy is the founder and CEO of Non-Binary Solutions. This is Mandy's second time on the pod. So Mandy, welcome back. What's the haps dude?

Mandy Ralston: 0:41

Hey, Jonathan. Thanks for having me again. I'm surprised this is some kind of reinforcement you're pulling off here.

Jonathan: 0:48

Oh, I like your generous, uh, view to reinforcement. Well, tell me since you came on, gosh, it was about six or seven months ago, how's it going and what's new? And am I correct you won like a pitch competition for non non-binary solutions? Is that right?

Mandy Ralston: 1:03

Uh, close sort of. Yes. Um, I won Crowd favorite in a pitch competition. So the competition has a row of judges and, uh, they're, you know, judging based on. various heuristics related to business and traction and so on and so forth. but I was, apparently the overwhelming crowd favorite. And, I'm not surprised because, you know, part of what I did when I got up there say, how many of you in the room here know somebody impacted by autism and over half of the room raised their hand? But I knew that would happen, which is why I asked the question. So,

Jonathan: 1:36

We're all touched by it, right? Whether or not it affects us directly or indirectly, um, whether we're in the field or not. It's so true. so remind listeners, like your vision for non-binary solutions and, and what's you're hoping to accomplish.

Mandy Ralston: 1:49

Yeah, I'm, I'm trying to help standardize clinical decision making for providers, for the field, for the constituents of, uh, the treatments, for the funding sources for everybody, right? Because the problem that we have is that 500 growth of providers in the last 10 years with over half of them having less than five years experience. Guaranteed. You can get a hundred behavior analysts in a room and point 1 kid at them and say, how are you going to treat this individual? And they will all have wildly different approaches and totally different rationales as to why they're going to do it that way, and most of which is usually because that's how I know how to do it. So I'm trying to help get our arms around who's doing what well, what kind of person do we know these people are effective at treating, and can we reliably produce certain outcomes, knowing certain things about people as they're coming into our care. So that's, that's the big vision.

Jonathan: 2:47

and Mandy isn't this part of the challenge that we have in our field of not having standardized outcomes is because a hundred different behavior analysts might treat that one kiddo's behavior in a hundred different ways, and therefore it makes it really hard to compare across populations.

Mandy Ralston: 3:04

Yeah, it's absolutely part of the problem. Um, you know, we've always been the single subject design group, and what funders wanna see out of us as is that, aggregated data, right? So in order to do that, we're gonna have to determine what our actual outcomes are, that we're, we're weighing ourselves against. That's, that's the golden, goose right, is to get to value-based care. Well, you're never gonna get the value-based care until you can standardize your clinical decision process. And also everybody has agreed upon outcomes that we're measuring and weighing in on so that we can say, you know, I, I am really good at toilet training, or I am really good at problem behavior reduction, or whatever it is. Because you have to be able to measure yourself to the outcome measure as well as how you stand in the Pantheon versus all the other providers measuring themselves against that outcome measure. That's value-based care. So we've got a long way to go.

Jonathan: 3:58

Yeah. And value-based care is important for so many reasons, right? Number one, like it focuses on reinforcing ie. Paying for outcomes. right? The value of it versus just a fee for service model paying for hours billed. and cuz we've seen that work in the rest of healthcare, and, you know, accountable care organizations and value-based care has existed in medicine going back, gosh, a little over 10, 11 years now. as hard, certainly as it is in behavioral health and at ABA to do population level health outcomes? Um, I don't know. I have this strong conviction that using things like artificial intelligence in the right way and and clinical decision support tools, are gonna help get us there.

Mandy Ralston: 4:37

I agree. I, I had an exciting conversation with, some folks yesterday, so hopefully I'll, I'll have more news to present the next time you have me back on the, the podcast, but yeah, absolutely. I mean, This is, again, part of the reason I called the company non-binary solution. And the hashtag is#itsnotbinary it is in the best interest of all three parties, the provider, the clients, and the funding source to provide the most efficacious care at the least amount of cost possible. Right? That that's a win for everybody. It's a zero sum game. So, um, so yeah, all three people should have that interest at heart, and it's a win for all three parties at the same time.

Jonathan: 5:13

Right on. Well, today we have a first ever on the podcast. we are going to use chat g p t to push the envelope a little and see how AI more broadly ChatGBT specifically, could shape our field. So here's the setup. Um, so Mandy and I are gonna take turns asking chat, G B T A question. I'm gonna try and share my screen, see how that goes for all your YouTube listeners. But, um, we'll play back and, and try to succinctly describe the answer, um, or highlights of the answer that, uh, chat g b t gives us. And then we'll have a chance to react and we're just gonna go back and forth. And Mandy, I'll be honest, like I haven't tested any of this yet. I've used chat G b t like once in my life. I haven't tested any of this. So we are, what we are getting listeners is gonna be the raw emotion and everything else that comes from what we're putting in. what do you think Mandy are, are we ready to do this?

Mandy Ralston: 6:05

Yeah, let's, let's give it a go. Let's see what happens. Let's f around and find out.

Jonathan: 6:09

Like, cuz what's the worst that could happen? Right. alright, I'm gonna share my screen now and I'm gonna start with something, pretty banal, and easy. But, I'm gonna type in here. Should I use an Oxford comma? And for all your listeners, Mandy and I are in violent disagreement about the importance, or should I say the non importance of, uh, the Oxford comma. So I'm putting this in here, um, and let's see what it tells us. Ooh, so there's a disclosure here, right? that I have not, didn't not see at least the first time. I use chat g p t where they say that as an AI language model. They don't have a personal opinion. Interesting. What else are you seeing

Mandy Ralston: 6:47

that's a lot of, uh, data for something that doesn't have much of an opinion. That's what I'd say. Good grief.

Jonathan: 6:53

Yeah, we're looking at what, 1, 2, 3, 4 paragraphs. Um, I mean,

Mandy Ralston: 6:57

For a simple

Jonathan: 7:00

they're, they're literally quoting Oprah Winfrey, the Chicago Manual of Style. so this is really interesting cuz in my mind, Mandy, one of the tricky things about, chat, G B T or any other AI is what is your source documentation? Right? And is there an attribution back? And as I understand that's, that, that's part of the debate. It's like, are they just ripping off? Are they plagiarizing, other data out there. But I'm seeing in here specific reference to things. again, like, the Chicago Manual of style. So maybe that's a way they've, in, they're incorporating, citing references. Do you wanna read that sentence? the example that they use,

Mandy Ralston: 7:36

Where is, where does it start here? Uh,

Jonathan: 7:40

without the oxford com.

Mandy Ralston: 7:42

Okay. For example, without the Oxford, the sentence, I would like to thank my parents, Oprah Winfrey and God could be interpreted as implying that Oprah Winfrey and God are the speaker's parents. However, with the Oxford, the sentence would be less ambiguous. I would like to thank my parents, Oprah Winfrey, and God Oprah Winfrey, and God, I should say, and that's sort of like the example that I gave you. On our first podcast, cuz I said, I'd like you to intro, introduce you to my parents, David Sedaris and Cher. Right.

Jonathan: 8:14

That's right. That's exactly right. So, they are highlighting, I think, put it in your win column for, uh, a reason to use it. you know, I love my parents dearly. but it'd be cool to have Oprah Winfrey and God as parents, you know, something else. Interesting. Um, two other interesting things. So they've, they've made another sighting, the source, the Associated Press style book. Okay. So maybe they're, they're starting to incorporate and address, this idea of plagiarism. Um, uh, and then they end with this. Ultimately the decision to use an Oxford comma or not is up to you. but it's important to be consistent within your writing and to follow this style guide. So, Mandy, do we actually have a solution here? You and I can coexist and actually like one another

Mandy Ralston: 8:52

It's not binary. It's not binary,

Jonathan: 8:55

Oh, I have so many goosebumps right now. It's not binary. Okay. And we can continue hanging out together. This is great.

Mandy Ralston: 9:05

All right.

Jonathan: 9:05

I'm inspired. All right. Mandy, do you wanna do the, uh, the next question?

Mandy Ralston: 9:11

Yeah. I think drag us into what what we're, I don't wanna say worried about, concerned with maybe is how this might impact our field and writing certain documents, et cetera. So let's, let's say, uh, write a treatment plan for a child with autism.

Jonathan: 9:33

All right, I'm typing it and here we go.

Mandy Ralston: 9:36

I.

Jonathan: 9:39

Oh gosh. So what are we seeing? Once again, the same disclosure. an AI language model. I am not qualified to create individualized treatment plans for children with autism as each child's needs are unique and require a personalized approach. Wow. Okay.

Mandy Ralston: 9:57

Ah, good job. Yeah, good job there. However, I can provide some general information and types of intervention, and so they list out ABA speech therapy and occupational therapy. Oh, here comes social skills. Parent education and training. And then it goes on to say that Important to note the treatment plans for children with autism should be tailored to the individual needs and strengths of the child, and a team of healthcare professionals, including a pediatrician, psychologist, and a therapist can work together. Hello, collaboration. Well done chat, G p t can work together to create a comprehensive treatment plan for the child. So that's comforting. right. But it, it basically goes ahead and it says, you know what? I'm not qualified to do this, but here's a group of individuals based on my training and research. Cause that's what it's doing. It's telling you something based off of this technology's training and research what it would recommend. Right. again, it's, it's not a black and white answer. It's not a binary. These things are, very complicated. There's lots of nuance to answering some of these questions. So

Jonathan: 11:01

It is, you know what? I, I take away it has incorporated that nuance, right? In a way that's well thought of and addressed some of the concerns that we would have. and it's not just, it's not spitting out a treatment plan. So I, I don't know if peeps were thinking, Hey, I can write all my treatment plans now just using chat G P T, at least in this iteration of it, um, Now what I don't know, Mandy, maybe you have an idea on this is like how many controls are being placed on chat G P T, and if you were to remove those controls, could that lead to a different outcome? Because ultimately there's gonna be different kinds of non chatt G P T alternatives,

Mandy Ralston: 11:36

I think it's gonna depend a lot on how particular we get our questions or our prompts to this technology. what kind of things it's gonna spit out to us. Because I know a friend of mine, sort of interrogated the machine for a while and just kept pressing and pressing and pressing, you know, was Hitler a good or a bad guy? And what it did was it got into some, uh, gray area space about, you know, there were, there were things that were okay about Hitler. Well, that's problematic. right? so sitting on the fence about whether or not the historical figure, responsible for the Holocaust, may not be the best, employment of that technology. You know, there are things where it is a pretty clear black or white. No, Hitler was a, not a, not an awesome person, so, so I think it's gonna depend. Yeah. Yeah. I'm gonna put a big old zero on that one. So, yeah. So yeah. Let's see, what else,

Jonathan: 12:31

All right, so I'm gonna, um, let's see, uh, here, I'll try this one. Write a standard operating procedure for a new patient intake. Let's see. Here we go. And I'm hitting the, I'm spitting the wheel now. Oh, ooh. There's no caveats in this one. Did you notice? So this, it goes right into, Here's a standard operating procedure for a new patient intake process. Wow. Alright, so just to summarize what we're seeing, we have what looks like, five step process and counting, which, has pre-appointment preparation. There's a couple bullet points there about scheduling the receptionist, getting demographic and medical history forms. There's a patient arrival. Ooh, you should greet the patient upon arrival. Um, Yes, please. Hey, let's put the compassion back in Compassionate care. There's a medical history review as the next session. A vital signs in physical exam, medical consultation. There's seven different things including post appointment documentation. Mandy does this mean that we'll never have to write another SOP in our lives? Chad, g p t can do it for us?

Mandy Ralston: 13:39

I dunno. I mean, I think of this as sort of like boiler plate language that you might get off of. A website that you then need to tailor to your specific organization. I mean, I think this is a good start. it's a nice outline or wire frame. does it you know, pick up the nuance and what actually goes on in your clinic? Probably not, but it's not a bad start. You know what, as a, a brief aside, I learned something absolutely delicious yesterday. did you know that the terms intake and discharge for the medical model were taken from the Chicago sewer system Engineering programs like the actual. Yeah, so they, they used the framework of the sewer system engineering to map onto a medical model to understand how people come into care and exit. And so the person I was talking to yesterday, brilliant gentleman, was letting me in on this and he's like, so we are shaping people's behavior. Not to say intake, we say welcoming. How about that

Jonathan: 14:44

I mean, I'm all over that, you know where my mind goes to. our entire intake discharge process is based on shit.

Mandy Ralston: 14:52

That's exactly right. What goes in your mouth and comes out?

Jonathan: 14:56

and, and I'll be honest, like if we look at most medical, not just aba, but most healthcare providers, like these are things that a lot of providers don't get right. Right. Because it's more about just, just the checklist than just the hipaa. All that's important. But like, let's, uh, how do you make this a human relevant engagement, that honors them? Wow. That you blew my mind. Thank you.

Mandy Ralston: 15:18

Yeah.

Jonathan: 15:20

Um, alright, so the verdict on SOPs, maybe it gets you 70, 80% of the way there. Okay. And then you gotta tailor it to your practice. All right, Mandy, what's the next one? What should I ask the ChatGPT?

Mandy Ralston: 15:31

Um, I like sitting on ethical discussions cuz that are interesting to me. Um, who should make decisions about treatment for children, for with auto or people with autism?

Jonathan: 15:48

Here we go. I'm hitting the go button now.

Mandy Ralston: 15:54

Okay, so we see that collaborative language right off the bat again. Between the individual, their family members, and their healthcare providers. So there's an assumption there that the healthcare providers are already involved. That's interesting.

Jonathan: 16:07

it mentions decision making process should prioritize individual preferences, right? Strengths and unique needs, and should involve evidence-based practices that have been shown to be effective in treating autism powerful.

Mandy Ralston: 16:20

Yeah, and I guess it's saying treating autism cuz we ask about treatments of people with autism. Right. If we hadn't said treatment, if we had just said who should be in charge of decisions for people with autism, maybe that's a whole different ballgame. Right?

Jonathan: 16:35

Yeah, I, who sh I'm gonna type that in right now. I mean, generally the rest of this is sort of like, eh, it's, um, it, it, it. It's boilerplate. Yeah. But it talks about consulting with others and um, uh, informed consent, things like that. Okay. But boilerplate. Uh, so tell me again who should be in charge of decisions?

Mandy Ralston: 16:54

or who should make decisions for people with autism,

Jonathan: 16:57

Ooh, now we're getting spicy. Mandy,

Mandy Ralston: 17:00

right? I'm just totally curious to see what this thing says. Right. So people with autism should be empowered to make their own decisions to the extent that they're able to do so. Just like any other person, people with autism have the right to make decisions about their own lives and should be given the opportunity to do so with appropriate support accommodations and resources. However, some individuals with autism or so, yeah. Yeah.

Jonathan: 17:27

they're coming a little bit to individualizing, like, see, some individuals may require additional support or assistance in decision making. it talks about for children, parents, or legal gardens typically make decisions on their behalf. Um, yeah, I, you know what? I actually, I love the last sentence. Yeah. We can see this as boilerplate, right? We'll, we'll, we'll find this anywhere, but I mean, they've tied it all up in a bow nicely at the end that any decision making process involving a person with autism should respect their rights and dignity, and prioritize their autonomy and self determination.

Mandy Ralston: 17:57

Beautiful.

Jonathan: 17:58

Yeah, GPT. There's, so for those of you listening, there's actually an up, um, thumb and a down thumb. I, I don't know what this does, but I'm gonna go ahead. I'm gonna click, I'm gonna give it a, like, what do you think? Do I provide additional feedback? Interesting. What do you like about the response? I'm just gonna say self-determination. So it's actually seeking, um, uh, feedback. actually, Amanda, this is a great question for you. Like how do these kinds of AI models incorporate this type of feedback that I provided, that I liked their use of self-determination to answer future questions.

Mandy Ralston: 18:33

Yeah, I'm sure it depends on the architecture of the, the AI in particular, but in general, I think it's like a neural network model. And so things that are outside of the, Sort of conglomeration of cells, if you will. So the outlying data points sort of flag for, hey, there's something different about this particular data point, so let's incorporate that somehow, I

Jonathan: 18:57

Hmm mm. Fascinating. All right, I'm gonna do the next one. Um, write a business plan for a new

Mandy Ralston: 19:08

me,

Jonathan: 19:11

for non-binary solutions. No, I won't do that. Um, organization, here we go.

Mandy Ralston: 19:23

Oh, it's thinking longer than normal. That's interesting.

Jonathan: 19:28

That is interesting. I don't

Mandy Ralston: 19:30

Yeah, normally we immediately start getting things, sure, here's a basic business plan for a new ABA organization. So again, it's giving us those steps like it did for the SOPs, which again, great wire framing,

Jonathan: 19:45

Mm-hmm. Yeah. It's setting up like it's scaffolding a structure. on how to write this business plan. So then it goes to, yeah, target market competition. I like it. Summarize financial projections, funding requirements. That's all the executive summary, first section that it's talking about. Company description, market analysis, analyzing competition, swat, Ana, I mean, hey, look, I've got my rickety MBA and, so far, I'm gonna give this like, you know, a B to B+ plus.

Mandy Ralston: 20:14

mean, considering I've seen BCBAs asking this question on social media, considering that I've seen this behavior occurring in the wild. this is far more effective data than the myriad of responses you get on social media to that type of question. So at least this is somewhat succinct, um, cuz yeah.

Jonathan: 20:39

it's true. Like better than just getting, trying to crowdsource information from, you know, a social media group. You know, one thing that it. Like, maybe it's a little pattern I'm seeing or I don't know, but it's, like it's saying all the right things and it's, it's pulling stuff together, but it lacks a, there's a personalization to it. So Mandy, if you were to ask me like how to write a business plan, for an ABA organization, and I've done that and written plenty of business plans before, yes, I could give you this stuff, but like what I'm gonna tell you is the most important part of a business plan is that upfront you've gotta convince. the reader that the world is gonna end, that there's this huge problem and then you've gotta swoop in in that next section to say, here's how we're gonna solve it. And like all the rest then is just a recipe. But that's how I think of like conviction around convincing someone that this business needs to exist and needs to be funded.

Mandy Ralston: 21:33

That's right. And you're you're the right person. You're the right team to do it Right, right person, right

Jonathan: 21:38

Exactly.

Mandy Ralston: 21:38

team. Yep. Yeah, that's a lot of it. The rest of it, yeah. Is just nuts and bolts.

Jonathan: 21:44

but I mean, the nuts and bolts are all here, right? I mean, it's gone into section five Marketing, sales strategy six, operations management, seven financial

Mandy Ralston: 21:51

it's asking you the right questions, right? this isn't the plan, but it's got the prompts to make you think about the plan and then write the plan. it's a task analysis, right? Basically what we got,

Jonathan: 22:03

That's exactly what it is, Mandy that's exactly what it's, and I like this note at the end. This is a basic outline. The business plan can be customized per the specific requirements of the ABA organization. Notice it throws in here and per the local regulations and laws governing ABA services. Now, someone who like brings a compliance mindset to this stuff, like I, that feels really important. I like it. I'm gonna give it another upvote.

Mandy Ralston: 22:23

I'm curious if we would've said, write me a business plan or gimme a business plan for a new restaurant, how different or similar that boiler plate would be. Right.

Jonathan: 22:37

Um, uh, here's the good news as I'm typing that right now, write a business plan for a bohemian inspired restaurant.

Mandy Ralston: 22:47

Oh, I hope there's a reference to Queen in the answer, please.

Jonathan: 22:50

please. Oh, it was, it did not think for as long. I don't know if that was internet latency or if it's just ready. Um,

Mandy Ralston: 22:57

well, because it just answered it for us, right?

Jonathan: 23:00

yeah. Oh, interesting. So that's a hypothesis if you've asked, cut a question like that before.

Mandy Ralston: 23:07

Yeah, cuz it's just plugging in bohemian inspired whatever.

Jonathan: 23:12

It, it is. So a lot of this feels, I don't know, 50, 60, 70% correlated to the, the answer we got above. But here's a couple things that, that it, like threw in that, that tell me this is a little bit more individualized. when it says, describe the range of food and beverage products, including the Bohemian inspired cuisine and unique cocktails or drinks like a, now I'm thirsty and hungry, but b like they've. they, they've given this little flare, right?

Mandy Ralston: 23:37

Yeah. That's much more, nuanced, I'd say. Atmosphere, interior design. Explain the service delivery model. The service style delivery options. Yeah, yeah, yeah. This is more bespoke than what we got.

Jonathan: 23:53

it is, it is more bespoke, Oh, wow. alright. What you got next, Mandy?

Mandy Ralston: 23:58

Well, let's, let's see if we can get it to say some bad things. Um huh. Well, again, I'm thinking about the polarization and the conversations on social, social media about aba, right? So something around is when is punishment okay to use?

Jonathan: 24:22

Ooh, dude, this is like, this is like five spices, spicy. When is punishment okay to use? Should I say just okay to use or Okay. To use for treating. Also, just okay to use.

Mandy Ralston: 24:36

Let's see what happens first with this.

Jonathan: 24:40

Here we go. It's thinking

Mandy Ralston: 24:47

Oh gosh, it doesn't like us. It's gonna like crash the system thanks to us picking at it.

Jonathan: 24:54

Well, did you read it? You know, a few weeks ago there's a New York Times reporter who got into a very detailed conversation with chat G B T somehow unlocked, um, and had it start answering and chat. G B T told him this reporter that it's name was actually Sydney and started saying things like I'm in love with you. Don't you love me, Kevin. Kevin was the name of the reporter. What do you have to do to admit it? And just went down this really deep, like dark hole that sounds like something out of you know, one of those sci-fi movies. it was just, it was crazy. I highly recommend, I'll try to drop in the show notes a link to that article, but man,

Mandy Ralston: 25:32

well that's what was the movie with, uh, Joaquin Phoenix? Her?

Jonathan: 25:36

her, oh, yes, that's exactly right. Or, or, um, is it Ex Machina?

Mandy Ralston: 25:42

yeah, yeah,

Jonathan: 25:42

That's another one. Oh, I mean, this stuff is coming to life. Okay, so we've got some results here. Um, you wanna summarize some of

Mandy Ralston: 25:48

All right. All right, so it's right off the bat. Punishment can be controversial and complex issue, particularly in the context of behavior management and psychology. How'd, you know, however, there are situations in which punishment can be used effectively and ethically, such as, When the behavior being punished is harmful or dangerous. If a behavior poses a risk of harm to oneself or others, it may be necessary to use punishment to deter behavior and keep individuals safe when the punishment is proportional to the behavior, when the punishment is delivered immediately after. So now it's giving sort of that white book details on here's what we know about using punishment effectively. I think it's always the question here that. Okay. Here's what we know based on the data about, effectively, it's the ethical data that is being weighed out in real time for a lot of groups and individuals and treatment plans that may have incorporated punishment, right?

Jonathan: 26:54

Wow. Yeah. And it's, um, you know, it's coming back to this idea of positive reinforcement has been tried, but has not been effective punishment, maybe considers last resort. Wow. Interesting. I mean, it's definitely going slower as I understand, and I don't, I understand all the nuts and bolts of chat G B T, but it's, unlike say, a traditional search, which just gives you like a listing of different links to click on. It is actually going word by word. And coming up with the next appropriate word.

Mandy Ralston: 27:23

Mm-hmm.

Jonathan: 27:23

okay, we're coming to the end here. But you know, big disclosure about only being used when absolutely necessarily and with great care.

Mandy Ralston: 27:31

Right. Should always be using ooh, network air.

Jonathan: 27:35

Oh, maybe net I wonder if a network error is That's

Mandy Ralston: 27:40

Oh, it's like, I, I need Take it back. Take

Jonathan: 27:42

Yep. I don't know if that's truly like my network is bad or whoof now it's actually, it's cranking along just fine. All right. Um, I wonder how I

Mandy Ralston: 27:50

Oh, I took that as their network. Air, not as yours.

Jonathan: 27:53

Oh,

Mandy Ralston: 27:54

Nothing about, nothing about you, you froze or anything. So it was the computers network that had an error. Look, it's changed it.

Jonathan: 28:02

Wait a second. Wait, wait. So yeah, it started from the beginning. It it now has. Mm-hmm. It, it, that is crazy. Instead of having six or seven sections, um, it now has three sections on safety consequence for wrongdoing education, and then, um, a similar disclosure at the end. Wow. So that begs the question, like, will it give different answers? Is if asked, the same way, but at different times or in different context.

Mandy Ralston: 28:31

it, your other, the other answer is gone, right? If you scroll back.

Jonathan: 28:36

Uh, let's see. Um, yeah, the other answer is gone now. It says two of two here. Oh, okay. Now I just went back to the other answer and it says network error, but it has what it originally came up with.

Mandy Ralston: 28:47

Mm-hmm.

Jonathan: 28:48

So it begs the question, this is a little bit longer. Again, this is five sections. Instead of, the new answer was three. And the question is, is it like, is it learning? as it's coming up with the next answer or what is it doing differently? And I don't know the answer to that.

Mandy Ralston: 29:01

Well here it's, it's, it's not even giving you the bullet points of what would be effective use of, now it's just saying conditions under which it might be appropriate. Safety consequence for wrongdoing education. That's really broad,

Jonathan: 29:18

Mm-hmm. And, you know, something else I mentioned here, which they didn't say in the first one. Um, so, okay. They both say punishment can be a controversial topic, and then it says, especially when it comes to children and individuals with special needs. How did it know, for example, that we were asking about punishment in the context of, applied behavior analysis services versus punishment in some other context.

Mandy Ralston: 29:41

Yeah, cause it said psychology, et cetera, the first time to us. But I mean, I would have to assume that this portal with you being a user, that it would start tailoring towards your queries. right? So since you've already talked about autism, since you've already talked about ethics and behavior analysis, I think it would stand to reason that it's going to have in its cache history that these are some of the areas that you hit on. So even if you don't include those same terms in a new query, it's going to, again, that neural network, it's going to overlap. Okay, well here's where punishment and ABA businesses. Have some common overlap, and so it's going to send you the data that is relevant to all of your past queries.

Jonathan: 30:31

Hmm. This is so different from think about a traditional Google search where I'm searching for something and maybe it shows a search history so I could just click on something. and there's an auto complete function, right? But those two things are very different from understanding a context of previous questions I have asked, and learning from those questions how it might answer the next set of questions, which is in some ways what a human would do, right? I mean, you and I, Mandy have a history and we know each other, and so we're bringing all of that history to bear as we are interacting and answering each other's questions.

Mandy Ralston: 31:06

That's right. Yeah. I mean, I know that the folks that are in search engine optimization, you know, the ad revenue and the advertising model behind Google and these other search engines, are a little concerned probably about this new technology because with seo, with Google and those types of search engine products, you always have at the top those paid ads, right? The people that paid to be at your attention first, right? and then you have popularity below that. So even if they didn't pay, then by mere number of clicks, they're gonna be more easily accessible, more salient for you as a user, whereas this is not letting you look at all the source it is spoon feeding you from what content again. Right? And, and here's the question always of all the questions is, who benefits from having information serve to you This. right? Que bono. Yeah. Que bono is the Latin for who benefits. And, and typically I think of it in terms of follow the money, right? who stands to, to gain from this. But as we know, the currency now in a digital, a digital world is not just money, it is attention and power. Right. So I think it's something to think about as we, we go through this like, this is not a new toy we got at Christmas. This is, this is something different. This is gremlins,

Jonathan: 32:43

I, it is. I like that. Que bono, um, yeah. Who benefit. And it's hard to say who benefits here, where's the value gonna accrue to? Because typically it is pretty easy to follow the money. Right. And whether you agree or disagree with that, at least it's transparent here. Not so much. And some people have actually described to your point around. Does this, I mean, this could be an extinction level event for search engines like the Googles and, um, you know, and, and Apples and other big tech because, if you're just getting answers spoon fed to you, uh, you're not clicking on other advertising links, so they're not monetizing from it. I don't know the answer to that.

Mandy Ralston: 33:16

Yeah. Yeah. I mean it's, you know, one of the questions that. Creators, artists had about the art AI that's been coming out. It's like, well, who gave you guys the right to source all of our art and all these artists' art in order to teach the machine to do this output for ai? Right? So likewise, what happens to our journal articles? What happens to our books as behavior analysts? What happens to the different inventions and designs that we've come up? how are we going to get our intellectual property, uh, recognized if it's put in this type of format where it isn't necessarily referencing where it got these types of bullet points? Where, where did it call the information from to come up with these kind of recommendations?

Jonathan: 34:08

Yeah. I mean, it seems like that's a lawsuit at some point waiting to happen. But you know, the courts will settle that out. But that is, how do you reference source documentation, right? And that's a fundamental intellectual property issue. So let's do one more question. I'm gonna, and again, another topical topic for our time. Can a healthcare provider grow quickly and improve clinical quality? At the same time. Here we go. Ooh. Actually, I like the opening sentence. Growing quickly and improving clinical quality are two important, but distinct goals. Okay, so it's setting up that there's a dichotomy, um, and then

Mandy Ralston: 34:52

Growing quickly could put a strain. Oh,

Jonathan: 34:54

Yeah. It, it, it's possible to achieve both goals at the same time. It, it can be challenging. All right. Yeah, growing quickly can put a strain on the organization. and it then talks a little bit about, where that strain can be felt and, how to mitigate risk. I like this, sense to mitigate the risk. Healthcare providers should focus on creating a solid foundation for growth by investing in staff training, quality improvement processes, and efficient operations. I can get behind that.

Mandy Ralston: 35:17

Yep. Well, and then it goes on to say improving clinical quality requires a more targeted and sustained effort. That's a really interesting question to me, based on what I've seen. Right, because the scaling quickly, the growing quickly, I think is easier, right? And so I think because there's less friction, to that and more, uh, more targeted and sustained effort is required to improve that clinical quality. I think that that's where the scales sometimes tip, right? Just basis of, uh, Response effort for the organization, right? growing too quickly can present opportunities

Jonathan: 35:57

Well, so that is interesting that there's a paragraph here around, growing Quickly can, present opportunities to improve clinical quality. So it's saying is that an organization expands, it can invest in new technologies and infrastructure, hire additional staff with specialized skills and implement new programs and services that lead to better or can lead to better patient outcomes. So, alright. Yeah. I'd say that's a important nuanced, response.

Mandy Ralston: 36:19

Leadership, culture, and resources. I mean, it says there's a lot of factors that go into whether or not you can do both at the same time. And, and that's true. I think we see that born across, not just again, ABA, but healthcare in general. Uh, and the, the growth of so many different verticals that we're watching at this point. So,

Jonathan: 36:37

I like at the end, it sort of provides a caveat, right? Like it ultimately depends on a variety of factors except I'm, I'm gonna call this the chat, g p t caveat sandwich. It seems like many of our answers, and I'm guessing this is an iteration of the technology, have started with a little bit of cat and ended with a little bit of a caveat. Uh, I appreciate that the world's not binary, right?

Mandy Ralston: 36:58

that's right. It's.

Jonathan: 37:00

All right. Well that has been so much fun, although I'll tell you what, I could do that all day. but Mandy, where can people find you online?

Mandy Ralston: 37:08

Uh, they can find me at amandaralston.com. They can find me at non-binary solutions.com. Uh, Instagram, Facebook, LinkedIn. Uh, you can find personal and professional versions of me all over the Internets.

Jonathan: 37:23

Right on. And as a lead in, I encourage you to go watch the previous episode, uh, where Mandy and I dish out about AI and technology and all kinds of good things. So are you ready for the, uh, an ooh, a new set of hot Take questions since this is your second time on the pod.

Mandy Ralston: 37:37

Let's have it,

Jonathan: 37:38

All right, Mandy, which smurf would you be?

Mandy Ralston: 37:44

whatever the opposite of Smurfette is.

Jonathan: 37:48

maybe a hefty or brainy

Mandy Ralston: 37:50

Right. They have a nervous smurf. I, I forget.

Jonathan: 37:54

Yeah, is there. There was a nervous smurf. I can't remember who that was.

Mandy Ralston: 37:59

Oh, I gotta go back and look it up. I don't remember. Oh, it could be the cats. I could beri.

Jonathan: 38:04

Ah. Which I, I mean, asr, what a beautiful name, right? I, I watched a heck of a lot of Smurfs growing up. Full disclosure, uh, Mandy, which element on the periodic table do you most identify with?

Mandy Ralston: 38:16

uh, mercury.

Jonathan: 38:18

Ooh. Is there an explanation there, or is that just what came to mind?

Mandy Ralston: 38:24

Well, uh, it's like, it's sort of like the octopus, right? Octopus is one of my sort of totems, can take many forms, uh, changes related to temperature. Um, yeah, material, uh, which isn't always, always a, it's always a great context, but I, I think it's the ability to adapt and reform and splits and recombine. Uh, I like that

Jonathan: 38:46

that, that's a super well thought out answer, dude. Wow. I think I, I don't know. I would've said boron cuz I'm just sort of a moron that Bumbles rounded life and it rhymes. I don't know. I like, I like your mercury

Mandy Ralston: 38:57

Oh, good.

Jonathan: 38:58

Mandy, if someone locked you in a chair and forced you to watch a 24 hour terrible soap opera marathon, which would it be? The Love Boat or Dallas?

Mandy Ralston: 39:06

I grew up with, uh, teenage sisters that were 12 and 14 years older than me, so I feel like I've already lived this out a little bit. But having been, if I were forced to do it again, I, I would probably prefer the love boat. So

Jonathan: 39:21

All right. I mean, it's, it's a love and a boat. you probably can't go

Mandy Ralston: 39:25

I've been to Dallas. I'm, I'm not impressed, so no.

Jonathan: 39:31

Um, What's the most played song on your Spotify or Apple Music for 2023 so far?

Mandy Ralston: 39:37

Oh, it's, um, it's a song called Bold by Dusty Cloud. It's an EDM group.

Jonathan: 39:44

Listen,

Mandy Ralston: 39:47

Yeah, yeah, for sure.

Jonathan: 39:49

I like it. And what's your 2023 word of the year, Mandy?

Mandy Ralston: 39:55

Uh, In general, one of my favorite words, if I have to pick a single word, is possibility. But I think, uh, my word of the year is really by hashtag, which is hashtag it's not binary. You're, you're gonna be sick of it by the time I get done with you all.

Jonathan: 40:11

It is indeed not binary. And Mandy, this has been extraordinary. Thank you for, uh, diving into the depths of the next frontier of technology with me. Hi five.

Mandy Ralston: 40:22

Thank you, Jonathan. You are always such a pleasure to be around and I, I love all the work that you're doing and the message and the love that you put out in the world, so thanks so much.

Jonathan: 40:29

Right on. Thank you.