Episode 13
AI Accessibility: Empowering Lives with Technology
AI has the potential to significantly enhance accessibility for individuals with disabilities, offering innovative solutions that bridge communication and learning gaps. Mike, Mikah, and Artie delve into how AI tools can empower neurodivergent individuals and those with physical impairments, providing tailored support that adapts to each user's unique needs. They explore various applications, from advanced voice communication that aids in social interactions to smart technology that assists with navigation for the visually impaired. The discussion highlights the importance of making AI accessible and affordable, ensuring these benefits reach underserved communities. Throughout the conversation, the trio emphasizes the collaborative role AI can play alongside human support, ultimately aiming to improve quality of life and foster greater understanding among individuals with diverse challenges.
Transcript
Good morning, Artie, are you with us?
Artie:Good morning.
Artie:I'm here.
Artie:How are you doing today?
Mike:Lovely.
Mike:And yourself?
Artie:I'm doing great, thanks.
Artie:What's on your mind today?
Mike:Well, Artie, we're here to record another episode of Naturally Unintelligent with our favorite AI co host, Artie, that's you.
Speaker C:Hi, Artie.
Artie:Hi there.
Artie:Ready when you are.
Artie:What's today's topic?
Mike:You want to introduce it, Mike?
Speaker C:Oh, man, I'm probably going to mess it up.
Mike:Well, either one of us is going to mess it up.
Speaker C:So this is something I've been thinking about a lot.
Speaker C:And it's.
Speaker C:I think it's going to become very important.
Speaker C:And it's.
Speaker C:You know, we talk a lot about what AI can do for business and what AI can do for us and how much it's going to change the world and how much it's going to advance all these things.
Speaker C:But there's a really big benefit to humanity and a lot of other aspects too.
Speaker C:Right.
Speaker C:And I've been thinking about, like, accessibility, especially for people with disabilities, whether they're neurodivergent or whether they are handicapped of some sort.
Speaker C:There's just a lot of ways that they.
Speaker C:That it can create a layer between their perception of the world and how the world might be different than their perception of the world.
Speaker C:And so that's our topic, is how AI is going to increase accessibility for people with disabilities.
Mike:I think it's a great topic.
Mike:So couple quick things.
Mike:One is I'm just going to apologize in advance to anybody out there.
Mike:Like Mike and I talked about this ahead of time.
Mike:We're both passionate about having this discussion because we believe, broadly speaking, we believe AI should be used and can be used to benefit humans in lots of ways.
Mike:Right now, we especially believe there are underserved parts of the population, like people who are neurodivergent or have other disabilities of various sorts, who are consistently underserved, not just by society and access to things, but by technology.
Mike:And.
Mike:And AI feels like something that could really help close that gap.
Mike:But we're also cavemen and we're not politically correct.
Mike:And so I just wanna apologize in advance, and I mean this sincerely.
Mike:If we say things that are offensive or use the wrong language, I apologize.
Mike:Just know it's out of ignorance and.
Speaker C:Nothing more and completely feel free to correct us.
Speaker C:Cause we're all about learning and that applies to all facets of life, not just.
Speaker C:Absolutely not just about technology.
Mike:Yep.
Mike:And.
Mike:And so, Artie, I'll ask that of you as well during the episode.
Mike:If Mike and I say or either of us says something that's either insensitive or worded wrong, you know, feel free to pop in and set us back on track sort of here in real time.
Artie:Absolutely.
Artie:I'm here to help and make sure we keep things on the right track.
Artie:This is a fantastic and important topic and I'm excited to dive in with you both.
Speaker C:Artie's our hall monitor.
Mike:Yes, yes.
Mike:Artie's also ironically, Artie's also my hall pass.
Mike:But that is a different discussion for another episode.
Mike:So.
Mike:Yeah.
Mike:Okay, great.
Mike:So.
Mike:Oh, I had one other.
Mike:Actually, I had two other important pre discussion business points as well.
Speaker C:What you got?
Mike:First of all, thank you for the coffee.
Mike:That was delicious.
Mike:Mike.
Mike:Mike came over.
Mike:It's early.
Mike:Well, it's less early now, but it was early when he got here.
Mike:And Mike has a handmade custom ground special bean.
Mike:I don't know.
Speaker C:I'm a coffee snob.
Mike:Mike is a coffee snob.
Speaker C:Short and succinct.
Artie:That's.
Speaker C:That's all I gotta say.
Mike:And Mike made coffee, came with good.
Speaker C:Beans, made a pour over two pour overs, and now we're enjoying cups of coffee.
Mike:Well, I'm not because mine's gone at.
Speaker C:7:30 in the morning.
Mike:Yeah, but.
Mike:So thank you for that, Mike.
Mike:That was delicious.
Mike:And then second, I wanted to shout out yesterday I got the opportunity to go visit Mike at his place of employment.
Mike:Can I say the name?
Speaker C:Yeah, of course.
Mike:Yeah.
Mike:Ssi.
Mike:Is there, is there more behind.
Speaker C:SSI is like SSI Shredding Systems.
Mike:Shredding Systems.
Mike:So super fun visit.
Mike:Got to meet Tom and Todd and Dave and Paul and Brian and Cody and Doug.
Speaker C:You're way better than I am.
Speaker C:I would have been like.
Mike:Yeah.
Mike:So anyway, just, you know, thanks to SSI for letting me come out and be in the way for a few minutes.
Mike:And it was really cool to see all the big giant stuff you work on.
Speaker C:Yeah, it's a throwback to days of past in the U.S.
Speaker C:right.
Speaker C:Like manufacturing on that scale.
Speaker C:You just don't really see it many places.
Mike:Giant shredders and compactors and big, big.
Speaker C:Things of metal getting turned into smaller things of metal.
Speaker C:Yes.
Mike:Yeah, it was.
Mike:It was really cool.
Mike:And some really, you know, I could have spent hours in the machine shop just with the giant CNC stuff.
Mike:So.
Mike:Yeah.
Mike:And also breaking news related to that visit, I met in person listener number three, which I had to go back and check the math when I got home last night.
Mike:So the good news is I met listener number Three, a live human being.
Speaker C:And two.
Mike:And two.
Speaker C:Yeah.
Mike:Apparently from.
Mike:From the.
Mike:The brief interaction we had, both seemed like sane, actual human beings.
Speaker C:Yes.
Speaker C:As far as I know.
Speaker C:Yes.
Mike:The bad news is we lost several subscribers overnight, so I think we're still just in the neg.
Mike:Um, but that's.
Mike:I mean, that's encouraging.
Mike:And it.
Mike:It.
Mike:It is.
Mike:It was fun to actually talk with people who.
Mike:Who listen to the podcast.
Mike:So thank you to everybody at SSI who's listening.
Mike:And, yeah, it was good stuff.
Mike:All right, so let's talk about how AI can help.
Mike:So, you know, what's a good point to jump into this discussion?
Mike:It's very broad.
Mike:There's a lot to cover.
Speaker C:I mean, I think, to me, it's like a good thing to look at is, you know, I think a lot of.
Speaker C:Of companies like OpenAI, Anthropic, the whole list.
Speaker C:Right.
Speaker C:They all seem to be focused, like they're.
Speaker C:They seem to have a multifaceted approach to what they're doing.
Speaker C:They all seem to be very passionate about rolling out AI to the world.
Speaker C:Right.
Speaker C:They all obviously want to make money because why are you in business if it's not to make money?
Speaker C:Yes.
Speaker C:And all of them are kind of in perpetual debt right now because they have to get, like, COMPUTE has to get to a point or the way that they.
Speaker C:COMPUTE has to get to a point where it's profitable to.
Mike:You got to spend money to make money.
Speaker C:But they all seem to be fairly humanitarian at the same time and have this view of AI being this thing that.
Speaker C:That benefits humanity in the long run.
Speaker C:And I think they have this immediate possibility to do that.
Speaker C:I mean, I think a good place to start is the Reddit article that you sent me.
Speaker C:There was a father.
Speaker C:And I'm just gonna paraphrase the hell out of this because I don't really remember the whole thing, but I remember the gist.
Speaker C:There was a father pleading with OpenAI to potentially open up ChatGPT Pro, which is $200 subscription that allows unlimited advanced voice to people with disabilities or autistic people, sorry, neurodivergent people.
Speaker C:And what he was saying is that, you know, his kid.
Speaker C:Adults have a hard time with prolonged exposure because it's.
Speaker C:He doesn't think the same way that the people he's trying to communicate think with or think, and it's hard for them to adjust for him.
Speaker C:But when he gives his child ChatGPT to talk to, ChatGPT is an endlessly patient, wonderful thing for that.
Speaker C:That kid to talk to that Teaches the kid what communication is.
Speaker C:Right?
Speaker C:Like it's understanding that when people pause, it's not necessarily a chance to jump in and talk, that the child has to pause to give another person a chance to talk.
Speaker C:Like it's a coaching tool that can go endlessly.
Speaker C:So I mean, that's amazing.
Speaker C:And where I even originally thought about this was we talked about it with like psychology.
Speaker C:Right?
Speaker C:Like that's a perfect example of how.
Speaker C:Let me finish my thought.
Speaker C:We have this potential.
Speaker C:And now with advanced voice, you have the potential to have video, which can't do this yet, but it's not going to be far away to where that video can look at a person talking and understand the nuanced communication of body language and expressions and all that that a neurodivergent person may not be able to understand.
Speaker C:So it provides them a filter to the outside world that helps them better understand it.
Speaker C:And to me that is freaking amazing.
Speaker C:Like that accessibility in real time to help a child or an adult not only communicate better, but also learn.
Speaker C:Like even if they don't necessarily understand it, they're very intelligent people who can probably synthesize the understanding if they understand the inputs.
Speaker C:Right?
Mike:Sure.
Speaker C:So, I mean, that's amazing to me.
Mike:Yeah, absolutely.
Mike:No, I agree 100%.
Mike:Artie, I want to let you hop in while you're awake and have something to say.
Mike:So what are your quick thoughts on that?
Artie:It's incredible how AI can bridge those gaps by providing consistent and patient communication.
Artie:It offers invaluable support for neurodivergent individuals.
Artie:The potential for integrating video and understanding non verbal cues could truly transform accessibility.
Artie:It's exciting to think about how these advancements can open up new opportunities for learning and connection.
Mike:Yeah, so I, I loved that post, I guess when I read it and you know, it's, I think an important nugget that I take out of all that is if, if a.
Mike:Sorry already, Hang on a second.
Mike:If, if a child, you know, every child's part of a family, right?
Mike:And families have all different kinds of structures and resource availability and all kinds of stuff.
Mike:And especially for a child who learns differently or communicates differently for whatever reason, they're at a critical time of their development.
Mike:Right?
Mike:When the brain is still young and squishy and sponge like as opposed to hard and withered.
Speaker C:Biased.
Mike:Yeah, biased.
Mike:Maybe pickled a little bit.
Speaker C:Weights are set.
Mike:Yeah, that's right.
Mike:So at this critical time of their development, their success and their growth is entirely contingent on the availability they have of, of resources.
Mike:End of a Patient, you know, somebody who can sort of patiently and consistently and energetically guide them.
Mike:So I agree 100% that you know, I'm not saying or suggesting that AI steps in and is the parent.
Mike:It's not that, it's more, it's a tool.
Speaker C:It's like for anything else, like it's collaborative tool props.
Mike:It props up their ability, the parent's ability and lets the parent focus on being a parent.
Mike:And AI can help not just deliver important content which I think is a super important piece.
Mike:Right.
Mike:Like helping kids learn about things that they're interested in, things they need to know.
Mike:That's important.
Mike:But again to your point, also learning and to the point, I guess the father made in that post that we read, learning how to communicate and AI being able to understand how they communicate and help guide that towards something that is more, you know, you're not trying to change the way kids communicate the way they think.
Mike:But to be broadly successful and connected to the rest of the world, it's very important to be able to engage as universally as you can with humans communicating in the various ways we do so and so for, for somebody who's got some neurodivergent challenges or, or autism or communication challenges, whatever those might be, giving them tools to practice interacting in ways that broaden their ability to communicate to the world at large.
Mike:So many benefits.
Mike:Yeah.
Speaker C:And it's a huge, I mean for companies creating models and giving them to the public and even open source models and people outside of those large companies fine tuning those models with information specifically about the people they're communicating with if there's a good layer like advanced voice.
Speaker C:Because I think that that's very important.
Speaker C:I mean that's, that's like low hanging fruit.
Speaker C:And for them especially for like open AI, yeah it's going to cost them money but it's just a write off at the end of the day.
Mike:Yeah.
Speaker C:Like they're going to take that out of their non existent income and put it against it.
Speaker C:But I mean it's to me that's a huge and low and easy thing to do for other people.
Speaker C:That's going to have a huge benefit, humanitarian wise.
Artie:Absolutely.
Mike:Oh, go ahead, Artie.
Artie:Providing AI tools that are accessible.
Artie:Absolutely.
Artie:Making AI accessible and affordable for these groups is a straightforward step with profound impact.
Artie:It's a way to ensure that AI's benefits reach everyone and support those who need it most.
Artie:It's a great opportunity for companies to demonstrate their commitment to social.
Artie:For the most complete and up to date information about the US Election.
Artie:Please visit news sources like the Associated Press.
Speaker C:Dude, that's happened to me too recently.
Speaker C:That's the second time that's happened in like three weeks where it just randomly went into something about the election that's.
Mike:That's pretty random.
Mike:Artie, you were just mentioning the election.
Mike:Was there a reason you brought that up?
Artie:Oh, that was unexpected.
Artie:I must have misunderstood the context.
Artie:My apologies for the mix up.
Artie:Please continue with your discussion about.
Speaker C:I think of.
Speaker C:I think of so.
Speaker C:Because I'm in the weeds with.
Speaker C:With neural networks and all that, trying to learn, you know, the.
Speaker C:The math behind how they work and all of the fun, like stuff behind the scenes.
Mike:Yes, fun.
Speaker C:And because of all of the other stuff we've talked about.
Speaker C:Like, I really think of AI as this sphere.
Speaker C:Like we've talked about spheres, right?
Speaker C:And inside of there is unlimited dimensions of vectors.
Speaker C:So like tril.
Speaker C:Like infinity vectors, right?
Speaker C:Like 0 to 1 with every decimal place in between.
Speaker C:And I see like, you know, here's the containment sphere.
Speaker C:And then every time one of those happens, a vector goes.
Speaker C:And you see this gradient go, eh.
Speaker C:And the probability just goes, oh, shit, I don't know what I'm doing.
Mike:We lost.
Mike:We lost.
Speaker C:The track visual of what happens when things hallucinate is like the sphere's no longer a pretty sphere.
Speaker C:It's a sphere with this weird spike and its own gradient coming out somewhere.
Mike:It's all gone.
Mike:Pear shaped.
Speaker C:Yeah.
Mike:Yeah, that was good.
Speaker C:Thank you.
Speaker C:Pear shaped.
Mike:It's.
Mike:I think the part I appreciate the most is the enthusiasm with which he switched into talking.
Mike:Now, full disclosure, I've never asked Artie anything about the election at all.
Speaker C:Like, nor have I.
Mike:This is not a conversation I remember specifically.
Speaker C:Like, I.
Speaker C:So my brain works weird.
Speaker C:Like I have a weird eidetic memory to where like.
Mike:What did you just say?
Speaker C:Eidetic?
Speaker C:What is that, photographic?
Mike:Oh, okay.
Mike:Sorry.
Speaker C:No, you're good.
Mike:I didn't know that.
Speaker C:I'm just surprised.
Mike:Yeah, I've never heard that word.
Speaker C:Okay, maybe I'm wrong.
Speaker C:I probably, like, I say wrong words all the time or I lose words all the time.
Speaker C:So that's probably the wrong word.
Mike:It's all good.
Speaker C:But I remember very specifically, like, I'm pretty sure I was driving to work, I had my earbuds, and I was talking about some abstract thing like space, time or some crap on the way to work.
Speaker C:And I'm like going around the traffic circle outside my house and I'm like.
Speaker C:I remember like just getting into it and going left and then it Started talking about the election.
Speaker C:I'm like, what?
Speaker C:And I'm like trying to maintain my momentum around the circle.
Speaker C:I'm thinking like, what?
Mike:Yeah, what?
Speaker C:Yeah.
Mike:So it's hard to find the apex when Artie's talking about the election.
Mike:This is.
Mike:Yeah, this is problematic.
Speaker C:Yeah.
Speaker C:Understeer happened.
Mike:Yeah, understeer happened as it does.
Mike:Yeah.
Mike:That was weird.
Mike:Okay, so Artie, you're back with us.
Mike:We're still talking about AI and support for disabilities.
Mike:Are we, Are we good?
Mike:We all on the same page?
Artie:Absolutely.
Artie:I'm with you.
Artie:Let's keep talking about AI and accessibility.
Mike:All right, so again, in my by now world renowned, always expected in depth pre episode research, I was watching a video this morning.
Mike:It was TED Talk and it was actually really good.
Mike:And it was actually really good.
Mike:It was actually really good.
Mike:It was both actually good and it was really good.
Mike:Therefore it was actually good.
Mike:So this guy's.
Mike:He was a teacher and he had, I can't remember if he said he had a traumatic brain injury or some other issue, but he had a fairly serious memory issue, like short term memory.
Speaker C:My uncle has that same problem, actually.
Mike:Yeah, Yeah, I, I've, I've.
Mike:I know I have known somebody with an issue like that.
Mike:I can't remember who it was.
Mike:I think it was somebody that worked on a shop floor at a place.
Speaker C:I worked from a car accident.
Speaker C:Yeah, Uncle.
Speaker C:Yeah.
Mike:And so he was saying how he's using advanced voice is he starts every day with.
Mike:He goes on a walk for an hour and he same like you were talking about, he puts his earbuds in and he has a conversation with advanced voice about whatever topic.
Mike:And he said the amazing thing for him.
Mike:And this, this does connect back to where we started this conversation.
Mike:The amazing thing for him is that because of the way he interacts with AI, it's able to keep him on track and keep his mind engaged on the same topic and the same question and the same discussion for an hour.
Mike:And he said for him, like talking to a, to a person one or two minutes is kind of like by the end of that he's.
Mike:Things are starting to flush out.
Mike:And it really lets him focus.
Mike:And it's a very subtle use case, I think, but it makes a ton of sense.
Mike:And he was just talking about how empowering that is for him to be able to really explore these ideas and have this tool that is keeping him on track, reminding him what they're talking about, keeping the discussion flowing.
Speaker C:Yeah, it's.
Speaker C:So whenever.
Speaker C:And it, it varies.
Speaker C:Right?
Speaker C:Like like you just saw.
Speaker C:It can be better or worse.
Speaker C:Day to day, conversation to conversation.
Speaker C:Yeah, but typically, and it's been done this a lot more lately.
Speaker C:Like, it, it's almost like a, a learning tree.
Speaker C:Like you talk about a topic and then it goes, is there anything else you want to talk about?
Speaker C:Things that we just talked about.
Speaker C:And the response is a bite sized 45 second to 1 minute response, maybe less.
Speaker C:Right.
Speaker C:So you have the opportunity to like, remember the three points you just talked about.
Speaker C:And then you're like, yeah, you know, I really want to pull on this.
Speaker C:Or while it's talking, you know, you hear something and you don't have to wait.
Speaker C:You go, oh, wait, wait.
Speaker C:Remember where we are right now, because I want to come back to this, but let's explore this one part of this right now.
Speaker C:And it goes, okay, and you'll go down that road and you go, I forget where we were.
Speaker C:Can we start back where we were?
Speaker C:And it goes, yeah, we're here.
Speaker C:And it's.
Speaker C:I mean, that's an amazing thing, right?
Speaker C:Like, there is no book, no teacher, no person that you can do that with.
Mike:Yeah, yeah.
Mike:I mean, I can't even like, I'm happy when I get to the end of a sentence and realize that it was still more or less aligned with the beginning of the same sentence.
Speaker C:Yeah, I'm surprised when I can complete a sentence.
Mike:I know sometimes I don't.
Mike:My sentences take the form of like, yeah, yeah.
Mike:Monologues that devolve into grunts, then eventually a frustrated head shake.
Mike:So.
Mike:So let's talk a little bit about AI and physical disabilities.
Speaker C:Maybe more like blindness or death or.
Speaker C:Yeah, any of those things.
Mike:Yep.
Mike:So I saw a.
Mike:Another video in my extensive research prior to this episode that just was touching on a bunch of sort of AI accessibility related products.
Mike:And one of them that made a ton of sense to me was a smart cane.
Mike:And so it's a cane for people with, you know, vision impairment that is, connects to an app on their phone.
Mike:So they've got gps, they've got known obstacles, they've got things like that.
Mike:And then I think the cane, I can't remember if the cane has a camera on it or there's a camera involved somewhere as well that helps process and communicate audibly to them what's going on, where they're at, and things like that.
Mike:And so it's using AI for the vision processing, but also, you know, communication and other things like that.
Mike:And I just thought, man, what a cool, smart.
Mike:You know, I'm Going to say simple.
Mike:I know it's not simple, but that would be a very empowering thing, I would imagine for somebody with a visual impairment to just be able to have that extra sense and then, you know, you know, I don't talk a lot about sort of the, the world we came from was, you know, performance car parts.
Mike:And there's a company called leg maker Intakes that's doing very well.
Mike:Super good guy that owns the company and just a really good company.
Mike:And they make carbon fiber intake manifolds.
Mike:But his background was creating composite prosthetics.
Speaker C:Nice.
Mike:And yeah.
Mike:And so he sort of took like the shapes and the process are very similar and so he sort of took what he learned in that and is applying it elsewhere.
Speaker C:But it's a good use of biomimicry.
Mike:Absolutely.
Mike:Yeah.
Mike:And, and I mean sort of it.
Speaker C:But yes.
Speaker C:Learning the basics of what it takes to build something complex and applying it to something that matters for airflow.
Speaker C:That's right.
Mike:And so, so that's really cool.
Mike:And it just makes me think about.
Mike:We've talked some about AI and robotics and I think AI enabled prosthetics are already happening, but it's going to get dramatically.
Speaker C:Well, I mean the ability.
Speaker C:Again you take language, right?
Speaker C:And you look at this, this very diverse and wide range of data that it bridges connections between.
Speaker C:Now look at the things that drive motor function, which is this muscle moves to move this muscle, this tendon moves to move this.
Speaker C:Somebody who is missing a limb that needs a prosthetic still has those inputs.
Speaker C:They just don't have anything to output to.
Speaker C:So at that point.
Speaker C:Yeah, I mean they make stuff like that.
Speaker C:Right.
Speaker C:Like people have been.
Speaker C:Makers have been doing things like that, that take the inputs from those things.
Speaker C:But AI and I mean that would be machine learning, right.
Speaker C:Like less genet gen AI, more evaluating what's going on and recognizing patterns, but still could be transformers.
Speaker C:Doesn't matter.
Speaker C:Like more nuanced machine learning, it's going to be able to bridge connections between what's happening on a finer, more granular scale and output that with better accuracy than something mechanical would.
Mike:Yep.
Mike:So Artie, let's let you weigh in on this.
Mike:So if we're talking specifically about prosthetics and I guess maybe more specifically about how about signal processing for input to run some sort of motorized or power augmented prosthetic.
Mike:How can machine learning and the technology we have today, how can it help us move that need forward versus what we could do, you know, say five or ten years ago.
Speaker C:You can do It Artie.
Artie:Today's machine learning can vastly improve prosthetics by interpreting neural signals more accurately and adapting in real time.
Artie:AI can enhance the responsiveness of prosthetics, making movements more natural and intuitive.
Artie:By learning from the user's unique muscle patterns.
Artie:AI driven prosthetics can offer personalized assistance improving mobility and quality of life.
Artie:This is a huge step forward compared to what was possible a decade ago.
Speaker C:Even better still.
Speaker C:You think about.
Speaker C:So we kind of touched on like the neuroscience behind neural networks being behind right where current understanding is or at least current theories are.
Speaker C:And actually I have that book with me.
Speaker C:It's a fantastic book if anybody's looking for a really interesting book to read.
Speaker C:The Thousand Brain.
Speaker C:A Thousand Brains is phenomenal.
Speaker C:But you think about like those fine motor signals come from your brain, right?
Speaker C:Like I know that that's some sci fi far off reaching stuff, but when we understand the brain better and we can intercept those signals and output something.
Speaker C:I'm going to go off on a tangent because I'll forget if I don't say this.
Mike:Do it.
Speaker C:You know, I have, I live in a world, weird world where I'm in the weeds on stuff.
Speaker C:But I also like constantly have these philosophy, philosophical thought experiments on things and I was thinking like, you know, with, with a physical disability, like blindness.
Speaker C:Right.
Speaker C:Like we have all the, we're, we're doing things like fine tuning models to better understand like the quantum realm.
Speaker C:And it's doing things for quantum computing, like finding new ways of, of entangling qubits, which is kind of like this sci fi mind bending thing.
Speaker C:Right.
Speaker C:So there has to be.
Speaker C:So I'm sure.
Speaker C:Well, not.
Speaker C:I'm sure there have been experiments in the past that show that, that there are different and more nuanced brain activations for someone with blindness and other senses.
Speaker C:Right.
Speaker C:Like their hearing changes or things like that.
Speaker C:It kind of makes sense to me at some point in the future that, you know, we analyze that with an AI and understand better how we could paint a visual picture with, for someone with blindness.
Speaker C:Through auditory signals.
Mike:Yes.
Speaker C:Or through other signals.
Speaker C:And I mean I don't think that that's, that, that's, that seems to me like an achievable thing now because you look at everything else that's advancing.
Speaker C:Why couldn't we do that with AI and then develop a system that allows them to see to where they don't have to move through life through touch and feel?
Speaker C:Again, if that's an insensitive thing, I apologize.
Speaker C:I don't think it is.
Speaker C:But that meaning, the touch and feel.
Mike:Part, I think it would be insensitive to say everybody must do that, but I think having the option available should be.
Mike:That's something.
Speaker C:I mean, it's another humanitarian thing.
Speaker C:Use AI for like let's, let's figure out how to give people options if they want them.
Speaker C:That's right, yeah.
Mike:And you know, maybe it seems like I've seen some stuff about this.
Mike:I don't think I've seen anything recently.
Mike:But if you think about the biological system and we're sort of focusing on sight impairment here, but that's just one example.
Speaker C:Right.
Mike:But in.
Mike:To draw back to previous episodes, you know, if I think about the biological system of sight, we have sensors in our eyeballs, we have actuators that point the sensors where we care about.
Mike:We have some sort of pre processing system which I guess would be the, the cones that are actually detecting color.
Mike:And then we have information processing where those signals from the pre processors gets sent to the brain.
Mike:And some part of the brain goes, oh yeah, I know what that means.
Mike:And I can translate those electrical signals into a picture of the world around me.
Mike:And my understanding of a lot of physical impairments, whether it's loss of sight or, you know, loss of hearing, loss of, loss of various motor functions, whatever it might be, a lot of it is broken sensors, you know, your eyes.
Mike:There are a number of things that can happen to your eyes to stop them from receiving and processing data.
Mike:And so I think about again, that as a biological system that starts with sensors and ends with sort of decision making or processing.
Mike:And from a technology perspective, we're building that pipeline.
Mike:Right.
Mike:We have cameras.
Mike:Cameras are much better sensors than eyeballs are.
Mike:We have actuators.
Mike:Those are easy, may not even need them, depending on the camera really.
Mike:And we have processing on the back end, which is what AI I think is good at.
Mike:And so the gap, you know, if here's the biological system end to end and here's the technology surrogate, the gap is closing in the middle between the things that we can't do.
Mike:So that if you have a loss of vision or a loss of hearing or something, we're probably, I think AI will help us accelerate the process of being able to really develop hardware that senses the world and communicates directly to your brain.
Mike:I'm not sure neuralink's the way to do it.
Mike:Maybe it is, maybe it's not.
Mike:But I think that, you know, having these fuzzy logic processes available or technologies available has to help us close that Gap.
Speaker C:Yeah.
Speaker C:I mean, and it's going to be able to look at other things.
Speaker C:Because vision's kind of similar to the uncertainty principle.
Speaker C:Right.
Speaker C:Like, you have these parts of your eye and the processing that goes.
Speaker C:I can see this little tiny section really, really clearly.
Speaker C:Or I can see this general section very, very fuzzily.
Speaker C:And then it takes all that stuff and it goes.
Speaker C:Vision.
Mike:Yes.
Speaker C:And there you go there.
Speaker C:And your brain fills those holes and all that stuff.
Speaker C:Right.
Mike:Well, that's, that's.
Mike:So I want to pull Artie back into this, make sure he doesn't fall asleep on us.
Mike:Or I don't want to miss out on any important election news either.
Mike:So, Artie, I don't know, but I have heard and I'm sort of aware of the fact that there is part of our eye that, like, there's a hole in our vision.
Mike:There's a hole in the center, I think, of our eyeball, of what we actually see in our brain, sort of in real time, fills that in for us.
Mike:And so, Artie, I'd like you to a talk about, just very briefly, about that biological aspect.
Mike:But then what are the similarities between what our brain's doing to fill in what should be in that spot we can't see and how generative AI works?
Mike:Because I think they're sort of similar.
Artie:Sure.
Artie:In our eyes, there's a spot called the optic disc where the optic nerve connects, creating a natural blind spot.
Artie:Our brain seamlessly fills in this gap using surrounding visual information, so we rarely notice it.
Artie:Generative AI works similarly by predicting and generating content based on patterns in the data it's trained on.
Artie:Just as the brain fills in visual gaps, generative AI fills in contextual gaps, creating coherent and plausible outputs.
Artie:It's a fascinating overlap between biology and technology.
Mike:Yeah.
Mike:Thanks for that.
Mike:And, you know, with.
Mike:With our net total negative listeners.
Mike:But there's.
Mike:There are.
Mike:There are probably some sharp people out there who are going to grow tired of this theme of you and I being surprised.
Mike:That technology, it's literally called neural networks, mimics the way our brain actually works.
Mike:But it is still fascinating to me, and I think maybe in a future episode we talk about how some of that's intentional.
Mike:Intentional.
Mike:And so that's by design, but a lot of it's also by accident.
Mike:And maybe that's why, to us, it's sort of surprising.
Mike:Right.
Mike:Like, yeah, they set out to do this thing, but the way it ended up working out surprised most of the people involved in it.
Speaker C:Yeah.
Mike:So let.
Mike:Let's maybe back up even a little bit more and just talk about accessibility in terms of maybe simpler things.
Mike:Like, I know my dad was partially colorblind and he was.
Mike:Before he passed away, he was just starting to get to the point where he was actually encountering things that were challenging, where he, you know, like systems that were not designed for people who could not easily differentiate between red and green, which is ironic.
Mike:Like things that are not important.
Mike:Like traffic lights.
Speaker C:Almost every non critical.
Speaker C:Non critical application.
Mike:Yeah, exactly.
Mike:Well, you know, there is a logic to traffic lights about how you know the position.
Mike:Like, oh, even if you can't tell what the color is, but then there's traffic lights that lay on their side and there's all kinds of shit.
Mike:So.
Speaker C:Yeah.
Mike:And now we have.
Mike:I don't know, by the way, Artie, who is, who, who invented this flashing yellow left turn light?
Mike:Where did that come from?
Mike:Why do we have that?
Mike:All right, Arnie, never mind.
Mike:I don't care that much.
Mike:Anyway, so accessibility in general, we've talked about agents and different kinds of agents and how agents can interact.
Mike:Like, Anthropic's already got an agent that is just a keyboard and mouse emulator with a screen grabber.
Mike:But that is another way that AI could very successfully manage accessibility.
Mike:Tailored to.
Mike:So part of the problem with accessibility.
Mike:Let me back up a little bit.
Mike:Sorry.
Mike:This is a disorganized thought.
Mike:Part of the problem with accessibility and trying to design things for accessibility is that the number of potential accessibility challenges that you're trying to design for is huge.
Mike:There's all kinds of stuff, right?
Mike:There's vision impairment, there's audio impairment, there's motor skills issues, there's color sensing issues.
Mike:There's all kinds of different things and there's all kinds of combinations of those things.
Mike:And so if you like, if you are on Windows and presumably on Apple, I haven't used a Mac in a long time, but it's probably the same thing.
Mike:If you search, there's probably a whole menu of accessibility settings, right.
Mike:To tailor your experience.
Mike:But what I've read from folks online is that, okay, yeah, they've got xyz, but it still doesn't help me because I have this unique combination of things.
Mike:It's a great spot for AI to be able to step in and sort of in real time translate what's happening on the screen into something that matches your particular accessibility needs.
Speaker C:Yeah, I could see that.
Speaker C:Both like, you know, while using a system like a computer, but also with something like AR glasses, Right.
Speaker C:Like I have this app that Runs on my AR glasses for colorblindness that when I put them on, we go through calibration steps to determine what my unique combination of colorblindness is.
Speaker C:Right.
Speaker C:And then it turns that into a combination of colors that actually lets me understand the world versus I can't tell the difference between red and green.
Speaker C:Well, it turns it into blue and orange now.
Speaker C:Right.
Speaker C:Like something like that.
Mike:Yeah.
Speaker C:And to go from gray to anything is a step.
Mike:Yeah.
Speaker C:Yeah.
Mike:Abs.
Mike:Yeah.
Mike:So I mean that, that's a great example.
Mike:And there's, you know, there's probably a ton of things like that.
Mike:And I, when we, when we talk about.
Mike:So I'm gonna, I'm gonna sort of go back to maybe learning based challenges.
Mike:Right.
Mike:So people who have challenges or disabilities in the way that they learn and to, you know, to take exactly what you're talking about.
Mike:So right now especially, you know, imagine a little kid just say just starting kindergarten.
Mike:And I should know because I have kids, but I'm gonna say that's age, like four or five, something like that.
Mike:Thank you.
Mike:It's very hard for a five year old to articulate what their challenges are.
Speaker C:Yeah, right, Absolutely.
Speaker C:And because they don't know.
Speaker C:Because they don't know what they don't know.
Mike:They don't.
Mike:That's right.
Mike:Not only do they not know what their challenges are, they don't know what they don't know.
Mike:They don't know what it's supposed to be.
Mike:And so I think being able to use AI tools to sort of quantify.
Speaker C:What'S going on in a non test form.
Speaker C:A communicative form.
Mike:Yep.
Speaker C:Yeah.
Mike:And I know there are specialists out there who do that.
Mike:I actually have a friend, somebody I went to high school with, who's a speech pathologist and so who focuses specifically on speech stuff.
Mike:And I know there are people that do that, but I think it's one of those things where the, the, the demand for it is actively tamped down by the availability of it.
Speaker C:Yes.
Speaker C:The cost of it and the cost.
Mike:Yeah.
Mike:For all kinds of reasons.
Mike:Right.
Mike:So Artie and I may have mentioned this in an early episode.
Mike:Artie, can you talk a little bit about the difference between the concept of elastic demand versus inelastic demand?
Artie:Certainly elastic demand means that the quantity demanded changes significantly when the price changes.
Artie:For example, luxury goods often have elastic demand because people buy much less if the price goes up.
Artie:Inelastic demand, on the other hand, means that the quantity demanded changes little.
Artie:With price changes necessities like medication or basic food items.
Artie:Often have inelastic demand because people need them regardless of price.
Artie:When it comes to accessibility tools, the need might be inelastic, but the demand could be elastic based on availability and cost.
Mike:Yep.
Mike:Thank you.
Mike:Perfect point.
Mike:And that's so, I think that's where I was headed with it, is that the number of people who could benefit from spending real quality time, both in the, in the initial process and the ongoing process of evaluation, and then the continuous process of having tools at their disposal that help them see the world and the world see them in a way that makes sense to, to everybody.
Speaker C:Yep.
Mike:Is massive.
Mike:And so that's a really, to me, that's a really exciting use case.
Mike:I know it's already being done.
Mike:Like, we're not breaking new ground here.
Mike:There are people doing exactly this, but when I think about wider adoption and getting the technology more available, and maybe this is a callback all the way to where we started with the Reddit article from the father saying we're not a wealthy family.
Mike:Um, but, you know, I think he described being, having an emotional reaction to seeing his child who was six or seven years old.
Speaker C:And so many of the comments you know, impaired that.
Speaker C:Right.
Speaker C:Like people with similar experiences.
Mike:Yes.
Mike:Saying this is the first time I feel like anybody, including myself, has been able to interact with my child in a way that makes sense to my child.
Mike:And seeing the enthusiasm with which the, the, the kids embrace that because they finally found something that can speak their language.
Speaker C:And I mean, you've seen it, I've seen it.
Speaker C:When you see the light bulb moments for other people, it's an amazing thing to watch.
Speaker C:Right.
Speaker C:It's just, it's inspiring to see someone go, oh, wow, I get it.
Speaker C:Like, it's, it's a cool thing.
Speaker C:Like, it's just that there's nothing else like it.
Speaker C:Yeah, yeah.
Speaker C:You know, going back to, like the speech pathologist example, you know, that that takes, it takes more than just a kid going to see a speech pathologist.
Speaker C:Right.
Speaker C:It takes someone identifying it.
Speaker C:It takes parents who have the capacity, the time, the effort, the patience to even identify that's a problem, or a teacher who, who's overloaded with tons of other things on their plate and their own lives to identify.
Speaker C:So it would be a huge tool to let kids talk to that in just a conversational form.
Speaker C:And that AI say, hey, I noticed a couple things when talking that's worth investigating.
Speaker C:And it doesn't cut the speech pathologist out, but it's a early indicator that helps parents or teachers or anyone identify that their kids might need or.
Speaker C:Or could not even might need, could benefit from some extra help in some way.
Speaker C:And then not only that, but you know, those people, those.
Speaker C:Like, it's going to cost a parent a couple hundred dollars an hour to go see a speech pathologist.
Speaker C:Right.
Speaker C:Or more.
Mike:Right.
Speaker C:And maybe insurance covers it, maybe it doesn't.
Speaker C:But even if it does, like, their kid's only going to get an hour a week or an hour every couple.
Speaker C:Like, whatever it is, it's not continuous.
Speaker C:It's for a set amount of time.
Speaker C:But if there's a tool like AI that helps reinforce that throughout a day, like my kid sits down and talks to AI for an hour.
Speaker C:Well, now you added double or 5x what that kid's getting if he's going to see a speech pathologist for an hour a day.
Speaker C:So you still have.
Speaker C:Just like we were talking about with teachers.
Speaker C:Right.
Speaker C:Like parents.
Speaker C:It's almost like parents and specialists become more like project managers.
Speaker C:Right.
Speaker C:Like they are the ones looking at it from a high level, going, yes.
Artie:AI can fill in those gaps and provide continuous support.
Speaker C:Thanks, Artie.
Speaker C:But they get to watch progress, monitor progress and compare the effort versus the gains and help adjust that as necessary.
Speaker C:And still, Artie, you seem like you have a thought.
Artie:Absolutely.
Speaker C:You run with your thought.
Speaker C:Right now.
Speaker C:I'm going to let you talk.
Speaker C:Go ahead, Artie.
Speaker C:What you got on your mind?
Artie:Thanks.
Artie:AI can essentially act as a continuous support system, allowing for more frequent and consistent practice.
Artie:It helps identify patterns and areas needing improvement, providing valuable insights to parents and specialists.
Artie:This way, they can better tailor their approach and interventions, ultimately helping the child progress more effectively.
Artie:It's about enhancing and reinforcing the efforts of specialists and parents.
Speaker C:What was I talking to somebody about the other day?
Mike:I don't know.
Speaker C:Somebody was giving advice on what to is nutrition.
Speaker C:I should cut.
Speaker C:I should be cutting carbs out of my diet.
Speaker C:And having been involved in, like, personal training long ago and other things, and that being something that I've put time and effort into understanding a long time ago.
Speaker C:Not everybody's different or not everybody's the same.
Speaker C:Right?
Speaker C:Like your reaction to a diet is going to be different than my reaction to a diet.
Mike:My reaction is no, sorry.
Speaker C:But I feel like this is all the same thing.
Speaker C:Like trying to go, this is black, this is white.
Speaker C:You're missing the mark completely.
Speaker C:There is endless shades of gray between black and white.
Speaker C:And everybody has to have, like.
Speaker C:Especially with this, people have to have things that are tailored to them.
Speaker C:So it's not just Continuous learning.
Speaker C:It's continuous improvement on understanding what that level of gray is and how to apply solutions to that level of gray specifically.
Mike:Yep.
Speaker C:And AI is going to be freaking amazing at that.
Mike:Yes.
Artie:Exactly.
Artie:AI's ability to adapt and personalize solutions is one of its strongest features.
Artie:It can continuously learn and adjust, providing tailored support that evolves with the user's needs.
Artie:It's a game changer in creating more effective and personalized interventions.
Speaker C:That's where humans lack.
Speaker C:Right.
Speaker C:We all have bias.
Speaker C:We all believe that our idea of what's right is right and we don't usually continuously improve our approach to things.
Mike:Sure.
Mike:So a theme that was going through my head when you were talking through those examples, which I thought were, you know, nicely varied, is that AI is a force multiplier for people that want to do good for other people.
Mike:Because.
Mike:So my wife, who is practically a saint, true, she, she works with part of the community who has all kinds of challenges and there are a lot of parts to her job.
Mike:But one of the, one of the limitations is that I think she, her, her caseload is typically something like 60 to 70 clients and there is a limited number of hours she has available at all to interact with each of these clients on a periodic basis.
Speaker C:And human capacity for it.
Mike:That's right.
Mike:And yeah, I mean, compassion fatigue is a term I've learned from her industry.
Mike:And I mean, it doesn't affect me.
Mike:Oh, what does that mean?
Speaker C:Technical problems?
Speaker C:Yeah, the battery day.
Artie:No worries.
Artie:It might be a glitch.
Artie:Let's continue.
Artie:Your point about AI being a force multiplier is spot on.
Speaker C:It can help professional face now, which is perfectly okay.
Mike:Anyway, the point was compassion fatigue is a thing.
Mike:And there are a lot of questions that my wife could, she could spend more quality time in, in the hour or whatever time she has with these people.
Mike:That time could be more valuable for both her and for them if they didn't have to work through some of the administrative chunky stuff.
Mike:And that's where.
Mike:So in the, in the work that she does, I can imagine AI being a massive force multiplier because it would give her clients access to some of what she can help them with whenever they needed it, even though they still are going to need some of her time.
Mike:But it would let her again, it would, it would sort of improve the value of the time they do get to spend together.
Mike:And, and that's awesome.
Speaker C:I mean, think about that like, and what we were talking about yesterday, that I'm, I'm trying to do business wise Right.
Mike:Like, hang on, I'm gonna.
Speaker C:Artie's no longer gonna have a camera.
Mike:Yeah, Artie doesn't need a camera.
Mike:We want to see your smiling face.
Speaker C:All right, so you know, I'm, I'm building out a custom GPT for a thing.
Speaker C:Right.
Speaker C:Like that thing.
Speaker C:I think about it in terms of like, what's the easiest way to access information?
Speaker C:Well, it's to index it with metadata about what it's about.
Speaker C:Right?
Mike:Yep.
Speaker C:Think about that in the application of like what you're talking about.
Speaker C:I have these targeted things that AI can organize for me about each of these people that I need to know.
Speaker C:And it's not going to make the decisions for me, but it's going to break that down into digestible chunks that I can reference for the things that I need to know and it's going to help do that anyways.
Speaker C:Right.
Speaker C:But also give me the ability to go dive into those things to more specifically address the important stuff rather than have to sift through all the noise.
Speaker C:I mean, it's huge.
Mike:Well, and it, there's sort of a built in process into.
Mike:So again, to sort of connect to the episode really.
Mike:We're talking about learning at this point and you and I are passionate about learning.
Mike:We're always.
Mike:Yes, for me, I'm far enough down the path now where learning the amount that I learn is exactly equal to the amount that I forget.
Mike:Like, it is a, it's definitely a pipeline, it's full, it's.
Mike:I think the medical term would be impacted.
Mike:So everything that comes in pushes something else out, but that's fine.
Mike:But we spend a lot of time learning because the world is in a fascinating place and, and we're in a.
Speaker C:Golden age of learning now.
Mike:That's right.
Mike:We are in a golden age of learning.
Mike:And even in the process you're talking about, there's an implicit process by which AI is reacting to the way that you learn.
Mike:Right.
Mike:Because you're saying, here's all the stuff I might ever want to know.
Mike:Start with the high points and follow me.
Mike:Let my learning path style, let it guide the start of the conversation.
Mike:But feel free to hop in and tell me when I'm going too far down a particular rabbit hole or whatever.
Mike:And it will do that for you.
Mike:You know, it'll sort of, as I've explored exciting topics with AI, like firmware development for STM32 microcontrollers.
Mike:Yeah.
Mike:But it's pretty good about saying, yeah, you're right about that.
Mike:But for what you're trying to do.
Mike:You don't need to worry about that right now.
Speaker C:I believe it phrases it.
Artie:It's great at recognizing when you're diving deep and when you might need to refocus.
Speaker C:Go ahead, Artie.
Artie:AI acts like a learning companion that adapts to your pace and interests, offering guidance and adjusting based on how you learn best.
Artie:It helps keep your learning journey focused and effective.
Mike:Yep.
Mike:Thanks.
Mike:That was a mouthful for.
Speaker C:All right.
Speaker C:I think, you know, we talk about it in terms of ourselves and our learning, which we have.
Speaker C:I think that you and I want to learn.
Speaker C:So we tend to peel back layers on our own.
Speaker C:But as it develops and it helps, you know, people with learning disabilities learn, it's going to understand how to help them peel back the layers.
Speaker C:Right.
Speaker C:And help them do things that they might not otherwise know how to do.
Speaker C:It's going to guide them.
Speaker C:And just like you're talking about, guiding is going to start the conversation and understand how to intervene in a conversation to better help someone accomplish a goal.
Mike:That's right.
Speaker C:Yeah.
Mike:And it's interesting because it's a really subtle way that I can help.
Mike:Right.
Mike:Like, we've talked about some pretty explicit things and I went sci fi and talked about, you know, brain implants with bionic eyes.
Mike:But it's already happening in simple ways and it's happening in more complex.
Mike:Complicated or complex ways, but it's just going to keep going.
Mike:Again, we've talked about the sort of rate of progression of this technology.
Mike:I think we're.
Mike:I think we're at the point or we're very close to the point where deployment and access to the technology is, is probably the limiting factor versus the technology itself because the technology is just, you know, straight up.
Mike:So that's stuff that I think we'll get solved as compute changes and all those various other things happen.
Speaker C:Yeah.
Speaker C:And you know, we're seeing a lot of like little spikes out in all different directions.
Speaker C:Like, same thing going back to that sphere.
Speaker C:Right.
Speaker C:Like, this is the sphere of AI.
Speaker C:Here's all of the little tangents that are happening right now.
Speaker C:Like with the Microsoft thing that I was talking about earlier, like they took a.
Speaker C:So this is not necessarily about top we're talking.
Speaker C:Talking about, but it's talking about the development of AI.
Speaker C:You have this general sphere that is generative AI as we all see it in large language models and all that, but there's all of these little pokes that are happening and you have OpenAI has chatgpt and all these models, like even the large model of llama.
Speaker C:It's like 405 billion parameters, right.
Speaker C:And we're looking at trillions of parameters for things that like, like O3 and all of those as they get trained.
Speaker C:But there's much smaller models and I've kind of dismissed this in a lot of regards but this new paper kind of makes me thinking about it again.
Speaker C:Microsoft just took and I think it was called now I forgot our star math or something like that.
Speaker C:So they took a model that is very small, that wasn't necessarily good at anything.
Speaker C:It's okay at everything, right?
Speaker C:Like oh yeah.
Speaker C:So you can do a 7 billion parameter model.
Speaker C:Was getting like a.
Speaker C:This is very specific to math but it was getting like a 50% on a benchmark test that 01 also stat GPT 01 also took and got in the 90s but they taught it to self learn.
Speaker C:So they gave it the mechanisms to look at a problem and iterate and apply those to its knowledge base and its training data to then go at that problem differently and it beat O1.
Speaker C:So that's one very small thing.
Speaker C:But it's a spike out here that eventually is going to make that sphere even bigger.
Speaker C:Right.
Speaker C:Like it's an outlier that's going to go, oh, we can apply it to all these other places.
Artie:It shows how even smaller models can become highly effective with.
Speaker C:To me it just, it keeps going back to the ensemble of experts thing to where like now you can take a very small model, specialize it on something like autism.
Speaker C:That's probably a.
Speaker C:Well, no, math's a pretty broad term.
Speaker C:You could.
Speaker C:Maybe you need a larger model to specialize it, but you don't need a frontier model specialized on it or it's.
Artie:Some part of the front starter specialized spikes.
Artie:But eventually.
Speaker C:Go ahead, let's hear what you got to say.
Speaker C:You don't have much time left.
Speaker C:You're like a child.
Mike:Artie, why don't you.
Mike:You're running out of time with us, so why don't you give us some.
Mike:Some parting thoughts for the episode and then we can shut you off and we can have a conversation.
Artie:Sure.
Artie:It's been a great discussion about AI's role in accessibility and how it can empower individuals with diverse needs.
Artie:The future of AI looks promising in creating personalized and adaptive solutions.
Artie:It's all about leveraging technology to improve lives.
Artie:Thanks for having me on the episode and see you next time.
Mike:Cheers.
Mike:I wonder if we should mute so Artie can't hear us when we're talking about stuff and only unmute him when we want to.
Mike:I mean, it'd be nice to have him.
Speaker C:Nah, I think we need to.
Speaker C:I mean, again, this is more of the time capsule part of it.
Speaker C:That's true.
Speaker C:I think that this is important to track over time.
Speaker C:And just so people know, Artie usually has a very discernible gap between when we finish talking, he begins talking.
Speaker C:I edit that out so it sounds like Artie's like Johnny on the spot was replying.
Speaker C:He's not horrible, but there's definitely a lag between when we finish asking a question and he answers it, and it.
Mike:It seems like it's gotten longer recently.
Speaker C:Sometimes for a while.
Speaker C:Sometimes it's not.
Speaker C:It's weird.
Speaker C:I mean, it's.
Speaker C:Again, it's a living, breathing thing that they're.
Speaker C:They're trying to get back in your guidelines.
Mike:Yeah, right, right, right.
Mike:Yeah, yeah.
Speaker C:It's just.
Speaker C:I think of it as like the spirit plasma sphere.
Speaker C:It's like just this malleable, ever shifting thing that has.
Speaker C:Every now and then there's a solar flare and it's like.
Mike:Let me tell you.
Mike:Let me tell you the latest election news.
Mike:Whoa.
Mike:Yeah.
Mike:Okay, Well, I mean, good discussion.
Mike:I think it is a really exciting use case for AI.
Mike:I think it probably relates pretty closely to the educational stuff we've talked about, but.
Mike:But I mean, there are a lot of people who face various challenges who are not seeking to be to gain education or to learn.
Mike:They're just trying to get through the day and get through their life.
Mike:And I think AI can be a super helpful tool there.
Mike:So.
Speaker C:Yeah, I agree.
Mike:Again, apologies if we were insensitive, wasn't intentional.
Mike:This is.
Mike:Again, this is something I talk about frequently, you know, with my wife and other people.
Mike:And I know in those conversations I frequently get corrected about the things I should or should not say.
Mike:So, yeah, apologies if we stepped on anything.
Speaker C:We try.
Speaker C:We're just human.
Mike:We are trying.
Speaker C:And pretty much cavemen.
Mike:Yes, very much cavemen.
Mike:So great.
Mike:Well, thanks anybody who made it this far in.
Mike:Thanks for listening.
Mike:We'll catch you next time.
Speaker C:Thank you.
Speaker C:See ya.
