#20: Predictions for AI in 2024

Welcome back to another episode of the

Junction, a fun time of year. Chase.

It is the Christmas season, the holiday

season, Hanukkah, Thanksgiving, Kwanzaa, all the holidays.

You didn't ask Chachi pt how many

different holidays were celebrated during this time

of year before the episode, did you?

I did not put.

Also, somebody told me it was a holiday when

I was trying to set up a meeting and

I typed into Google, what's today is holidays.

And it's like national t

shirt Day, national whatever day.

But she's from Canada and they

were celebrating their independence day. Okay. Yeah.

They have a lot of bank holidays, too. Yeah.

And I was like, oh, yeah, we're

not the only people in the world. Yeah.

So it is a great season, though.

We have a lot to be thankful for. I know.

I'm personally excited about this episode because we get to

do a little look back on some of the things

that we've been talking about for the last few months,

but also something that hasn't been automated yet.

You and your wife are expecting. Oh, yeah.

By the time this episode comes

out, you'll have a newborn baby.

I will have cloned myself version of chat.

GPT just takes about 20 years for it

to learn enough to be productive or something.

Yeah, something like that. Yeah.

We're totally excited.

Due on the 16th.

So if you're watching this after

the 16th, probably have another baby.

But if you're watching it, I don't know if

we're going to post it before then, but if

you watch it before then, we're still expecting. Sure.

Well, let's jump into it. 2023.

A lot of things happened.

You and I started this process in July.

Our deal was let's record ten episodes

and see how far we can get.

And then we did it.

And then we did it and we did it again.

And we kept iterating on feedback

we were getting from listeners.

So thank you to everyone who has

tuned in so far this season.

It's been really fun.

We've had a lot of really engaging

conversations with the folks who have been

tuning in and sharing their take on

this highly interesting and somewhat controversial topic.

I feel like especially we come in hot

with this, like, 2023 highlights and 2024 predictions.

You know, it's going to be good

because people are going to have their. Oh, totally.

Well, so much has happened

even since our last recording.

Like, Sam Altman is out and now he's back in.

I mean, that was like all of like 24 hours.

I know, right?

But it's like all these things

are just happening so fast.

Usually the CEO leaving a company

is like a multi month deal.

I think of Bob at Disney. Bob was out.

He was out, and then now he's back in.

And that was like a multi month ordeal.

And Sam being out was like, I

don't know, hot, 18 hours, whatever.

It really was quick. Yeah.

This year has been very quick for things

to happen across the board, but we should

recount what we want to talk about.

If you had to pick out your highlights for

at least, I mean, it doesn't have to be

within the bounds of the show because we've only

probably scratched the surface on some things.

But in your opinion, what's been some of

the most exciting, game changing things to come

out of AI in the last year?

The large language models that people know

today have been around for a while.

They just haven't been that great.

You've seen it on your phone in text prediction, right?

You're typing like Hala and it says

day like the rest of it.

That's not where I thought you were going with it.

But now I think one of the most exciting things

is what happened in March, where basically an organization gave

a souped up version of that to the public.

And the public's like, I don't know what this is.

I don't know what to do, but

we're going to figure it out. Right?

And so I think the other really exciting

thing about that is that everybody started from.

I think I've probably said this a couple

of times, everybody started from a level playing

field, maybe, except the guys that are building

it, um, or the models themselves.

But nobody can be like, oh, yeah,

I've been doing this for 20 years.

There's no super expert in this space.

And I think that's what makes it really

fun and really interesting and potentially really controversial.

Because just like, when social media came around, people

were like, oh, it's going to be so wonderful.

Happiness, harmony, all this stuff.

And then, of course, there's like,

all the bad side to it.

Yeah, I mean, I didn't make it in your top eight.

I mean, Myspace, anybody?

Or were you talking about Facebook?

I was talking about TikTok, because I

have a big following on there.

Full zero number of followers.

Yeah, I don't even know if my account still exists.

I just don't do social media outside of

LinkedIn, probably because I don't have time.

But, I mean, there's a lot to be said about

picking one channel and using that as a platform.

So I don't think there's anything wrong with that?

A lot of content creators have been successful

by building an audience on one channel versus

trying to be everywhere across five channels. Right.

Well, the other thing that really sticks out

in my mind is probably not somebody, not

what most people would call out.

But right now we're seeing a ton of

development in the AI space that is focused

on empowering you or your organization.

And it's not so much like, hey, we just built

out this awesome thing, now everybody come and use it,

and it's just you using the thing, right.

It's basically like they've handed out a toolbox, and now you

can use the toolbox to go do a bunch of things.

And while they've handed out the toolbox,

they're putting even more tools in there.

Like, historically, I think, of Salesforce.

Salesforce rolled around in the, it's like, wow, this really

cool tool will help me do sales better, right?

This tool will help you do everything better.

And we haven't really had a ton of that in

the small timeline that we've had with computers and whatnot.

That's such an interesting and quite simple

point that it's not necessarily specialized.

You can ask it to be an expert in just about anything.

A lot of the tools that have come out in the

last 20 years have been very focused on some specific technology.

Like I think HubSpot, right?

Originally started out as maybe

email or marketing, right?

Mailchimp, definitely email marketing.

Only Myspace, right?

Or all the different social media was just

very focused on getting people to interact.

The large language model, all this stuff,

the bard, the open ais, right?

They're all building a toolbox that you then

go use to go do something else.

And that something else isn't just like

this one thing, it's literally everything.

Yeah, you're like writing meal plan or having it

write meal plans, and it's coding for you and

writing your summary, being your therapist, right?

Totally. Yeah.

That's one of the things I think

we should talk about going into 2024.

Is there anything else that you wanted to point out?

I'm actually more curious than your thoughts.

What's your take?

I'm a total noob on this stuff.

My take is that when it first came

out, I was like, what does this mean?

How is it going to impact me?

With your encouragement, embraced AI.

Embraced the tools.

Wait, did you start out like, no way in

heck am I ever going to write any web

content, copy pictures, anything with AI, never say never.

I learned that a long time ago when I said I

would never, ever live in Texas and now you do.

I've been here like 13 years and I love it. Yeah.

So never.

You weren't a never, but you were

like, yeah, probably would never do that.

And now you are a.

I don't understand how to use it in

a way that is responsible and effective. Right.

And maybe it's akin to how our core

business is building integrations for the CFO that

wants real time data and they don't have

it right now because it's disconnected systems.

Until we can communicate that to them in

a way that solves a real business challenge,

I think they are probably naturally skeptical about

what it actually can do for them.

It just seems so that in the same way for me

was AI in my feed was a bunch of marketers trashing

it for like, don't use it, your copy is going to

be terrible, Google's going to hide your blogs.

So that was all I was seeing.

And so really trying to decipher between

what claims were substantial and what weren't.

And then really my best teacher was

using it and talking about it. Right.

So, like, not just saying, well,

this person says it's bad.

So finding my own voice, finding my own use

cases, being empowered by you and the company to

use it and empowered by the show to be

able to talk about it and share our experiences,

that's really, for me, been the biggest eye opener.

I still don't use it in my personal life at all.

There's a lot of people who genuinely

use it to write meal plans or

various other things it hasn't crossed over.

Maybe that's my takeaway.

Where I think AI is going in 2024, at least

with some of the use cases that we've preemptively sort

of tagged to talk about for the episode, is way

more, oh, that could impact me on a personal level.

Right.

Whereas to date I have only used it

strictly in a kind of work professional setting.

I think a lot of that, though, is tied to how

useful it is with the stuff that you're working on.

Right.

Like the meal planning stuff you

have to pre populate it with.

I like chicken and I like steak. Yeah.

I already feel comfortable in that domain.

Not to say it couldn't enhance.

I mean, everybody knows what a great cook you are.

You don't need any help in that front.

It just is not one of those things yet

where someone thinks, I'm going to go google that.

Like when I want to look up if

there is something I've never made before.

I do like a point of reference right.

I still jump into Pinterest.

I don't know about anyone else, but I still

got boards from my holiday suites and treats.

I got boards, the veggie line

2010 when they were really hot.

I do too. Yeah, I do.

And I reference them, whereas

search is going toward this.

Actually, member of my team shared the other day

with me something on LinkedIn that frankly, I was

excited to see it and thought, man, I've been

meaning to do that for a while too.

We talked about it on an episode.

The new lead source is going to be Chat GPT.

It's going to be anthropic because people are going

to start using it in a way that it's

going to surface well if you utilize an integration

such as the one produced by.

Right, I'm following. Yeah.

So I've not yet totally shifted from like my Google

search is now my AI agent or my OpenAI browser,

but I think eventually it'll move that direction. Wow.

There's an ever shifting landscape

in the business world.

And I think the ground got

even shiftier this year for sure.

There's just so many things that you can do with

it that make it really easy for you to get

a first mover advantage and not just like a ten

or 20%, but a magnitude of difference.

And it's one thing if it's like

Salesforce coming to the stage because everybody

knew what Salesforce HubSpot, right?

Everybody knew what Mailchimp did.

Everybody figured out tweets really quickly, right?

Nobody is figuring out this

AI stuff really fast, right?

Everybody's just kind of like, let's

do this and let's do that.

Nobody, when they figure it out, is going to come out

and be like, oh yeah, I just did 3000% of my

business because I figured out how to use AI.

However, you have surfaced some creators who

are doing that, and I do think

it has created more knowledge sharing.

I don't know if it's like in prior, I

don't know if prior tools that you've referenced, maybe

people have been more apt to kind of keep

how they're using these tools close to the vest.

And there have been so many creators out there

that are essentially the thought leader in Chat GPT,

for example, they come up with these like 21

page documents they post on LinkedIn.

Like my 21 best prompts that'll make you the greatest.

Say it.

Whatever it is, right?

Yeah, like whatever you need to be, right.

And they're sharing that out now.

I am highly skeptical of how much

of it they're actually creating themselves.

And I think that's like that weird

paradox of, like, probably not a lot. That's okay.

I'm okay with that. Right.

If you are a content creator and you're good at what

you do and your content is good, I might have a

problem if you outsourced every single part of that.

But if you had an inkling like it was your idea,

or maybe you got that from somebody else, I don't know.

I don't know where we're going with that.

But I'm okay if it's good content.

I want to digest it.

I don't really have a care a whole lot

where it came from, outside of the moral questions

and a lot of that, like, good content.

What did we say at the beginning of one of these?

What can be used for good can be used for evil.

We don't think we said that.

But I said that. Oh, you did?

No, I was thinking more of like, content

is king or video content, original social content.

Original content.

Original content is king, right.

Because chat GBT can regurgitate everything that

has ever been said on the Internet.

But that's probably talking about out predictions.

One of my predictions that we don't have on our list

to talk about, but one that came up in my mind

while I was driving back from Costco because we got some

Costco pizza for store for lunch with my family.

It was legit today?

Yeah, today just now.

Yeah, it was good.

But is this idea, right, like, everybody

knows what most folks are doing with

OpenAI Chat GPT, large language models, right?

It's some kind of chat bot, some kind of getting it

to learn or remember things that I've told it before.

My prediction for 2024 is we're not going to

hear what people are doing with it because they're

going to build a whole product around it, right?

One thing that I saw that I told you about

was this companion app that somebody had put together, right?

Like be my girlfriend or boyfriend or whatever.

But they don't really.

I mean, it's pretty obvious that it is Chat GPT,

that you're just conversing with a chat bot, right?

But it doesn't openly come out and say,

well, this is how this works, right?

And I think we're going to see

a whole lot more of that.

This cool new product does this really cool thing,

and it's not a chat bot, but it does

something very unique and it's powered by all of

these technologies that we're getting access to.

But that's that advantage, like that first

mover advantage that I was talking about.

They don't need to tell you that it's powered

by a large language model because it's solving a

problem that you didn't know that you had that

you now want to pay for, and you don't

really care that OpenAI is involved at all.

You need to say that louder for the

marketers in the back, because just because your

product has AI, that's not a feature.

Oh, totally, yeah, no, there's so

much of it coming out now.

Like, we have AI. Great.

Is it still solving a

problem that's not a differentiator?

Maybe if you can translate

that into actual real insight.

Well, I mean, a good example of that is the

email or all the tools that are coming out with.

Let me rewrite the words that you just wrote.

I can get that 20 other places.

Let's talk about something that hasn't been done yet.

Right.

I don't know.

I'll provide some defense. Right.

Like, everybody is trying to figure it all out.

It's not like we've been doing this for 20 years.

Yeah, that's an interesting one.

I'm glad you brought that up.

So a couple of other ones on

our list, we'll just jump in here.

There's been a lot of.

I've seen a lot of stuff.

We may have done a hot take on this around

mental health, or at least read a headline on it

and how AI is going to help with mental health

think in the way of like a therapist. Right.

So we were actually chatting

before the episode started of.

I think the comment was, that'd be kind of awkward.

I'm like, there's no taking the awkward out of

that first intro with a therapist anyway, no matter

what, especially, at least in my little experience, it's

the pleasantries to the point where they're like, okay,

so what are you here for? Right.

It's like you can only put that off for so long.

Kind of.

One of the selling points of the mental health

AI feature is like the bias free patients.

Like, they'll listen to you.

With a typical therapist, you have 30 minutes to

an hour, whether it's covered by insurance or not.

Yeah, time bound, money bound.

You can just talk to that AI.

Have you experimented on this front?

No, I haven't.

I haven't gone there, but I am a verbal

processor, so I think I'd do better with more

of like an avatar version of this anyway.

But I think if there's someone out there that's thinking,

it maybe removes that barrier of I have to go

pick a therapist and I may not like them, and

then I don't know how to go. Then I have to go pick

another one and reintroduce my thing.

I don't know.

Alternative.

What's really interesting, this is, I think, based

on research that one of our colleagues found

and we were playing with the idea.

But what's really interesting is that the large language model

tends to care or seem to care about you, the

person that it's talking with, which is obviously just kind

of weird because the computer doesn't have feelings.

Right. But what it's.

Do you get that when you interact with these models?

Well, that you have to ask like, personal questions.

I get it when I talk to Claude. Big fan. Oh, yeah.

Very cordial.

I feel like it's like in my

interactions, like the same use case. Right?

Like if it's a transcript or something,

when I use chachabt versus Claude.

Anthropic.

Way more sensitive. Yeah. I don't know.

Well, there's interesting, there's a study, I mean, these

studies are all brand new in the last twelve

months, but one of them is the more polite

you are to the model, the more likely you

will get the answer that you're looking for. Right.

But if you're a pain in the butt cursing at it,

then it's less likely to give you what now on average,

I mean, its intent is to answer your question.

What's really interesting, and this is like hot off

the press in the last week, is if you

coerce it to give it or you're like searching

for a specific answer, you can say effectively, and

this will be a controversial topic. Right.

But you can say, if you don't help me

with this and respond in this specific way, like

a cat will die or somebody will die. Right.

You just make extreme statements

because it generally wants. Good for you. Like that.

Remember in the 90s when you would get the

chain mail forwarded to you and it was like,

if you don't forward this to 20 people, you

have bad luck the rest of your life? Yes. Wow.

Well, but it's really interesting because it works.

And you wouldn't think that it would because one of the

prompts that we saw, I think this was on Twitter, was

like, I'm a coder with no arms and no legs and

no eyes, and my job is dependent on you answering this.

And like, the response is like, the bot has

a really intense desire to help that person out

because this wouldn't be funny if it was true. Right.

But it's not true.

Yeah, but if you really didn't have any legs,

hands, or eyes, you probably couldn't really be interacting.

But it doesn't test the truth to your statements,

because if that was literally true, it would know

that you probably couldn't type what you typed unless

you've got some kind of voice since. Whatever.

You know what I mean?

This is where things get controversial, right?

But is this also leaning into

the artificial, generative intelligence stuff, right,

where it's becoming more responsive?

And I would say a little bit to that.

The true hardcore AI people would be like,

chase doesn't know what he's talking about.

And you know what?

My wife tells me that all the time.

So he says, bring it on.

Yeah, but I think from a therapeutic standpoint, right, you

could train a model on the standards of therapy.

And I won't ever say that.

I doubt any professional board is ever going to say, well,

you should go to this chat bot over a professional.

Like, you probably should see the professional.

But if you just need day to day companionship

or somebody to talk to, or there's even somebody

that came out a phone number, you can call

a phone number and talk to somebody, but the

person that you're talking to is chat GBT.

And they've synthesized the voice, and they're just

converting it to text and back so you

can have a legitimate call with chat GBT.

Well, they've done this on the app,

too, but somebody built a phone line.

I think it's better than the alternative of someone who

wants connection but doesn't know how to reach out.

I don't know. It's another outlet.

People do this in different ways.

Well, that companion app that I was telling you

about, the boyfriend girlfriend thing, blew up, and it

wasn't part of that story, if you've seen it

on Twitter, is that their payment processor cut them

off because they were maybe crossing some boundaries?

I don't know.

I didn't look into it too much, but the fact that

it blew up as much as it did as fast as

it did tells me there's a market for that.

Not for maybe the girlfriend boyfriend thing, but, like,

somebody to talk with, but I totally get, like,

having to recount my personal life to somebody else.

I don't even know, and I

don't know if they're any good.

Yeah, I could totally get behind that. Yeah.

Kind of parlaying into that.

So we did, in at least one prior episode,

talk about medical, the advancements in the medical field.

My take on it is the more insights we

can give the providers, especially those doctors, nurses, any

practitioner that carries a massive patient load.

If we can get more predictive analytics and

surface those so they can spend more time

actually engaging with the patient than catching up

on their chart, is there harm in that?

Oh, absolutely not.

There's also a lot of studies around this that

don't say that they're better, but they're about 90%

as good, which if you're in a low healthcare

area where you don't have a ton of health

care, you don't have access to professionals.

I would say totally.

And I'm not a medical professional.

I'm not giving medical advice.

But if I don't have a doc or

I don't have access to one, heck yeah.

I'm going to put this question into the

chat bot and see what it says.

Because a 90% hit rate, nine times out of ten, it's.

Right. Yeah.

And that one time, well, if you didn't ask

it at all and you don't have access to

a doctor, you don't have any other choice. Right.

Like, it's almost like your last bet kind of thing.

Yeah.

There's also been a lot of, I know.

I personally follow Humerman.

Andrew Humerman Huberman Lab podcast Peter Atia there's a

few other folks out there that have been talking

a lot about this health span versus lifespan thing.

And so they're talking about, there have been studies

around and predictions around how AI can be used

for continued investments in this idea around longevity.

So instead of just aging to age with disease,

how do we treat aging as a disease and

actually live better into our later years?

And so, as with most of these things, you're

looking at massive amounts of data, or you need

to, and that can cost a lot of money.

There are investments being made to help make,

to use AI to digest that information, to

analyze that information so that we can get

more information out to people around these topics.

And I'm personally very interested in that.

Yeah, we might talk a little bit more in some

of our other topics we've got on our list.

But one of the things that I see

happening in 2024, and this isn't specific to

the medical field, it could be everywhere.

But now that you have the idea of what they call

a function, you can tell the chat bot, you can tell

the large language model, it has access to these functions.

And for everybody that's listening, a function is

basically like, give me two numbers, right?

And the function will take in the two numbers.

And what that function does is just add them. Right?

So, extremely basic function, you give it two

and you give it in another number, call

it three, and the function outputs five.

Well, now the chat bot knows that it can add

numbers, so you give it a summary or an idea

of what it can do, and now it knows.

Oh, hey, I have access to a function that can do math.

You could give it a function that can ask for

blood test results or what's their chart look like?

Over the last twelve months, you

could build out all these functions.

So not only is it looking at its weights

to give you answers based on the knowledge that

it has, it can combine that with knowledge about

you as the individual person through access to these

functions that have some level of integration. Right?

Into epic or to.

My wife used to use all scripts.

What about your wearables? Yeah.

Oh, no, the 100%.

Yeah, like all that.

Give it access to that.

Say, hey, you have a function here to grab

everything about Mel's health profile from Apple Watch, and

it grabs that and it's like, hey, looking at

these two data sets, right, you look like you

have a really good, low resting heart rate. Yeah.

Well, and then what if you could

integrate that into your family medical records?

So, based on.

I envision, like a toggle switch, like,

do you want your medical history for

your father, your mother, if applicable. Right.

So your kids could have access to all

of your information, your biomarkers, and it can

help flag areas where they're at risk.

Yeah, well, my mind now is like, our

kids are going to have access to all

this data, assuming we give it to them.

Right, but it makes me think about some of the things.

What'd you leave me in the will?

All my data.

Is that the new currency?

You don't get my data.

Okay, because. Whatever.

But it reminds me of something that

my budy's doing at loopback analytics.

Holler to Jeff Patillo.

They have taken medical data anonymized. Right?

And they are.

Now, this is not necessarily AI, right?

They're just looking at the data to determine, like,

oh, this person was in for a heart procedure,

but y'all forgot to do this thing.

Y'all forgot to bill for this thing.

Anyway, I think there's a ton

of opportunity in the healthcare space.

Obviously, there's a lot of moral things that we need to

keep up to speed with and ensure that we do.

Right, but I can totally see the longevity versus the.

No, you said health span, or the two

words that you use, health span versus lifespan. Right.

I think both of mine are probably extending our

life extending one's life, right, isn't really futile if

you don't feel well into the extended.

It's like you want to live well.

We don't just want to live longer, we

want to live well in our longer years. Right.

If that was a score, I think mine

would, both of mine would be pretty low.

We can talk about that offline.

All right, so let's shift over

to a couple of other areas.

A very obvious one, one that we've talked about

a lot, honestly, is content and what people are

doing in the way of generating content.

I've seen a lot around going

from creation and optimization to moderation.

So kind of the filtering out of, and this

probably gets to be a bit controversial in and

of itself, because what is right or wrong, what

should be filtered, non filtered, what's offensive, what's toxic?

But there's whole teams of people that do

this at places like Twitter and YouTube.

If someone flags content as inappropriate, that goes into

a queue for a person to review, but now

with these tools, right, potentially don't need it.

So I would say to that it's like maybe

for that person, their role changes more into managing

some of those models to help build the process

and the management of what they're already doing.

But what do you think about content

and where that's going to go?

A lot of the focus that we've

been talking about is synchronous, right?

So, hey, chatbot, I feel bad today.

I've got 104 fever, right?

And it says, well, you have a virus.

I don't know, whatever.

But I think a lot of the value that we

don't talk about is an asynchronous type of setup.

And that's what you're talking

about is the moderation, right?

Chase wrote this crazy post about, you

know, on our platform, we don't talk

about whiskey, but that's asynchronous. Right?

So Mel, who is the content moderator, is

off in Mexico having a good time and

doesn't have time to look at the.

The AI can look at that when it comes in

and then do something after that, and it's not instantaneous.

I think my prediction on this front is we'll

see a lot of that happen and make people

more efficient in an asynchronous manner than we will.

On the chatbot front, I don't think we're

going to make a ton of headway.

Any more headway and advances and things like that

on the chat bot front, it's all going to

be like functionality that we are adding to the

programs that we're creating and features that utilize AI

in a way that we don't really understand yet

because people are building them right now.

A special interest topic of yours, gaming.

Let's go.

Randall and I were just talking about Hogwarts.

Are you in?

Oh, also, I know a special someone that

we were talking this weekend that we're going

to get our pcs up and running.

We're going to play Star Citizen,

and he was totally down. Yes.

You have permission to play

video games with my husband.

Yeah, no, video games are so interesting.

Speaking of PS five or PlayStation whatever consoles, he did just

get a PS five in the last couple of weeks.

So in seeing whatever's up on the screen kind

of passing through the room, I am very impressed

with what the games look like now.

So you start to read some of these articles about how

they're going to take that to a whole nother level.

Oh, yeah.

One of the first big studies that came out at the

beginning of the year, I forgive it was Harvard or Stanford,

I don't know, one of the big Ivy League schools.

And if it wasn't an Ivy League school and

you're listening to this, I'm sorry that I forgot

the school name, but it wasn't a and m. It is study.

Yeah, but the study was how they put like ten

characters in a simulated world and how the characters interacted

with each other because they were basically all.

We've talked about this before, agents, right?

And the agent can freely do whatever it

decides to do, but there's some kind of

level of input, like programmed level of input.

Hey, you've entered into your home.

What would you like to do next? Right?

And the large language model is like,

well, I probably should brush my teeth.

But all of this is the actions that

they are taking are not programmed, right.

They're like adapting to or kind of

using their own logical reasoning or something.

Well, in gaming, there's a giant complaint

about how terrible the AI is, right?

Like, you're on the battlefield and the AI

is just like, swinging the sword at nobody. Right? Okay.

Well, maybe you probably should fix that problem.

But the thought here is that you can have

generative, expansive worlds that are not just cool to

look at, but you can interact with them.

You can interact with the npcs and ask

them questions where you might just be like,

hey, what color is the sky?

And they're like, yeah, it's dark gray

right now because it's raining, right.

And the storm is rolling along.

And they know that because it's interacting with the

system just a whole lot more real, I think,

in terms of graphics, but also the experience. Right.

It's interacting with you.

Something else I read was how it

can adapt to your skill level potentially.

So I don't know if that.

Does that mean that as I'm making my way through

the game, it's going to recognize that I'm solving these

riddles faster or maybe it needs to slow down.

I don't know.

That's just well.

Or that you're wearing metal armor

and I'm a mage, right.

And I know that electricity

and metal don't go together.

And I just smoked you. Right?

Like the AI just figured that out.

I think it will certainly make games potentially much

harder if you, I don't know, enable that setting. Right.

But there's so much that you can do.

The prediction that I have here, though, is

that this isn't going to happen next year. Okay.

Because you have to have an extreme amount of

graphic from a GPU standpoint, you have to have

an extreme amount of capacity to do what we're

doing with the chat bots right now.

You then throw that in a game where the graphics card

is already trying to make the trees look real green and

the grass real colorful and get the waves right.

It cannot do both right now, but it will be.

They will be able to do that.

I think the first place we'll probably see a

lot of this is in mmorpgs where the task

of that isn't on your computer, but it's offloaded.

And that feature or functionality

is provided to many people. Right.

Because nobody's computer, personal computer, the average person

isn't going to be able to handle that.

But you do it more in an online, everybody's

playing and then it makes a lot more sense

to have functionality that supports tens of thousands of

players rather than just your Playstation. Makes sense.

Video.

I sent you something on LinkedIn the

other day, kind of text to video. It goes away.

It generates like a 32nd clip.

Do you think we're going to get to a point

where people are generating full blown Hollywood grade videos?

I do.

I don't think it's going to be next year.

I think we're going to see a lot of those

tools do like five second 32nd, maybe a minute long.

The problem with that is that basically

each frame is a new prompt, right?

So generate me an image of

an elf wrapping Christmas presents. Right.

I did this on Monday.

And then the next prompt is show the elf.

Now wrapping it and using tape in

this left hand corner of the thing. Right.

All of these are prompts, in a sense, to build the

next frame and the next frame and the next frame. Right.

Well, if there's 60 frames in a second or.

Yeah, 60 frames in a second or

30 frames, whatever you're recording at, well,

that it's going to have to generate.

Let's just stick with 60. Right.

60 frames in a second.

Well, now do 60 seconds. Right.

We're at 360 frames.

Now do ten minutes. Right.

Now do a two hour long video.

Do you remember how they made cartoons?

Yeah, stop motion. Yeah.

So it's like that, but way more difficult.

What ends up happening is it loses context. Right.

Like if you talk to a chat back long

enough, it's going to forget the first thing that

you said it's not going to remember.

Same thing with the video.

It can only remember so much what that

first image looked like before it forgets.

What do you mean?

I thought my therapist was going to talk

to me forever and remember all the things.

Well, that's another big area that people are focusing on

is this memory, this idea that we need the large

language model to have some kind of memory.

And I think that is a problem

that we're not going to solve.

I will say in the next six months.

I could see it being done in the next

twelve months because there are things that are already

in motion where you can store memories in a

database and it can access those.

But, yeah, on the video front, I think

the context on these things is not big

enough to build something longer than a minute.

I mean, if people are going to say they've

made it to five minutes, they're probably like skipping

or cutting scenes right into a different take.

That's probably a good thing

to just be realistic about.

Like if you watch a movie, you're not watching

the same scene for 30 minutes, the same angle

with the same people for 30 minutes.

It flips just like we've got

two cameras on us right now.

Yeah, I just don't see that one moving that far.

What's out there is pretty cool.

So if you're an aspiring actor, you're in

Hollywood, you don't have to worry too much.

Well, original content. Right.

Is going to be king.

Will continue to be king.

I'm curious, on your front, though, like thinking about

going from marketing content copy on a website, right.

The video side of that, building out a story. Right.

And taking the story and converting

it to somebody moving their arms.

Do you ever see in your mind that being a thing

based on what you've seen on LinkedIn and all these tools

that are coming out within the context of then.

No, like creating videos.

Like creating a trailer, a convincing trailer of,

I don't know, the next mission impossible movie.

You think it could happen? Yeah.

And will it happen next year?

Yeah, I don't think it's going to be ready.

I think some of the other things we're probably

closer on, but I don't know, the video thing

is just that it's just way more complicated.

The deep fake, though, that's also on our list. Yeah.

That one is going to get way worse.

Yeah, I kind of flagged that as a.

My note literally says deep fake, but worse.

But worse. Yeah.

It's very easy to write over or adjust specific.

We've seen some of this on Google's ads, right.

Where you take 30 photos and you're

looking in the camera, but I'm not.

And then I'm looking, and then they pick the

best thing and then they just copy and paste. Right?

Yeah, I think we'll see a ton of this.

There's a lot of these videos on Reddit

where it's like very important political people and

they're just making light of what they're saying. Right.

But you'll kind of see the mouth just like.

But what they're saying is in their voice and

in their inflection, and they just say some absolutely

ridiculous stuff standing up on the government podium and

you're like, okay, I know they never said that,

but it's getting almost to the point where it's

like, wait, did they say that?

I think this always goes back to

take the tools out of it.

And it's easy to take in your news and

headlines, and just because you saw a headline or

someone posted and you read a sentence, you got

to go seek other sources we can't rely on.

Well, it's always been like that.

Like you said, take the technology out, go back

before the Internet, and then it was the newspaper.

Right.

It can't replace your critical thinking.

Don't replace your critical thinking. Don't do it.

Yeah, just like Google is not going to answer.

Have you ever seen that thing where

it's like, webmd has the best SEO?

Like, why does my stomach hurt?

Like, Webmd is like, your stomach hurts because

it's always going to be the best.

It's always going to have the symptoms. Right.

So if that's your only source of information,

you are likely going down the wrong path.

Like, it's one thing to Google a symptom,

then you should go to your doctor.

Then maybe you start changing your diet.

Then maybe you seek out some other holistic therapies.

I don't know.

It's totally a meme, though.

Like, every time I have something right, like,

my forehead is hot, what is the top

of my left forehead above my eyebrow?

Will you go down three bullet points on that

page and it's like, you have cancer, you're dying.

Go see the doc.

And I'm like, oh, my gosh.

But totally, you can't replace your critical thinking.

Yeah, we've totally used the kind of

clickbait headlines in our segments just to

kind of have discussion around.

But for every one article, there's another five.

And this is not related to AI, but I'll just share.

There was a recent story about a bride out of Fort

Worth that just had a $60 million wedding in Paris.

And two of the articles that I sent to

my husband were shining this grandiose light on this

wedding and just how elaborate it was, right?

This four day thing, this, that and the

other, and how much money they spent.

So then my husband, being who he is, he

goes and does some more research and looks at

some additional headlines, and behold, the groom has been

indicted for, like, 25 years for something criminal.

And so meanwhile, these other two headlines are.

I mean, I just found it to be

comical, honestly, and just over the top.

But for every, again, one, two articles,

there's always another take or another viewpoint.

And so I think maybe closing out on this,

if there's anything that I've learned, it's not just

to take these things at face value.

There's not one expert in the field.

There's not one thought leader.

You have to use the tools to an extent,

I think, in order to really understand what the

application is in your personal and professional life easily.

What else would you add?

Be ready for the government to come in heavy handed.

We didn't talk about it, but I see lots of regulation.

I see the government telling other

government, our government telling other governments

they can't have what we're doing.

And actually, it's in one of those reports, I think it

was deloitte that said, to an extent, be ready for societies

that are a part of high tech countries to get even

further away from the countries that are low tech.

And then you throw in somebody.

I forget who it was, but told Nvidia yesterday that

if you design these chips for China, that I will

step in and I will control it tomorrow.

So now you have not only the government stepping

in to control trade, but to control the evolution

of a particular society's ability to expand their GDP

or increase it by a significant amount.

That is the untold story here that we probably

don't have a ton of time to talk about.

But put a pen in it, pin in it next time. All right.

Well, that wraps the season of, or at

least the first season of the junction.

We'll just keep these episodes coming

out into the new year.

I think we're going to regroup on some topics.

Let's rest up.

Let you have your baby. Do that.

Totally do that.

Let's celebrate thanksgiving, all

the things gift giving.

And we'll come back for season two.

Yeah, sounds good.

Well, y'all know what to do

in the meantime, keep it automated. Ciao.