May 29, 2025 • 24 mins
In this episode of Growth Talks, I sit down with Abran Maldonado—AI thought leader, OpenAI Ambassador, and co-founder of Create Labs, a Black and Brown-founded startup transforming the fields of generative AI and digital human design.
Abran is also the creator of C.L.Ai.R.A., the first GPT-powered digital woman of color, demonstrating his commitment to inclusivity and representation in AI.
Together, we explore the critical issue of bias in artificial intelligence—how it manifests, why it matters, and what must be done to address it before it becomes deeply embedded in our digital infrastructure. Abran shares his insights on the ethical responsibilities of AI developers and the importance of inclusive design in creating equitable technologies.
We also delve into the future of AI, discussing emerging trends, potential societal impacts, and the evolving relationship between humans and intelligent machines.
Tune in to hear how we can build a more inclusive and ethical AI future—and why the time to act is now.
👍 Follow our guest here:
https://www.linkedin.com/in/abranmaldonado/
https://x.com/abran
📔 Read my book "Growth Talks"
https://amzn.to/3zfqRTu
🧲 Watch my free lead generation course:
https://gaito.link/skillshare
🙏 Subscribe to the channel:
https://gaito.link/subscribe
📚 Download the Reading List:
https://gaito.link/gtbooks
Abran is also the creator of C.L.Ai.R.A., the first GPT-powered digital woman of color, demonstrating his commitment to inclusivity and representation in AI.
Together, we explore the critical issue of bias in artificial intelligence—how it manifests, why it matters, and what must be done to address it before it becomes deeply embedded in our digital infrastructure. Abran shares his insights on the ethical responsibilities of AI developers and the importance of inclusive design in creating equitable technologies.
We also delve into the future of AI, discussing emerging trends, potential societal impacts, and the evolving relationship between humans and intelligent machines.
Tune in to hear how we can build a more inclusive and ethical AI future—and why the time to act is now.
👍 Follow our guest here:
https://www.linkedin.com/in/abranmaldonado/
https://x.com/abran
📔 Read my book "Growth Talks"
https://amzn.to/3zfqRTu
🧲 Watch my free lead generation course:
https://gaito.link/skillshare
🙏 Subscribe to the channel:
https://gaito.link/subscribe
📚 Download the Reading List:
https://gaito.link/gtbooks
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Chow, everyone, and welcome back to the show.
Speaker 2 (00:02):
I'm Ralph, your host, and my guest today is Abron Maldonado.
Speaker 3 (00:06):
Good to see you. I'm glad to be back at
AI week. I love it here.
Speaker 4 (00:10):
How he's going, How he's good, It's going great. I
always take some time to enjoy the culture first. So
I went to Florence and then came here to my
line and wrote to be a part of this, and
it was great the first time and the last time
I came here, AI was in a different place. And
you know, I call it like dog years with AI
because one year in real time and AI is like.
Speaker 3 (00:31):
Five years, is it? So a lot has happened between
now and then.
Speaker 1 (00:34):
I agree, I love it. I want to start with
a with a simple one.
Speaker 2 (00:38):
I just wanted to know more about your journey with
with with with the I.
Speaker 1 (00:41):
So where did it all started and where are you now?
Speaker 4 (00:45):
So in the beginning I came there's a there's a
word in in English, a phrase I don't know how
it translates here, multi hyphen it and multi hyphen it
means people that have multiple interests and where.
Speaker 3 (00:58):
I have done a lot of things, had different career experience.
Speaker 1 (01:01):
People need ye.
Speaker 3 (01:02):
Yeah, so I was. I worked as an artist.
Speaker 4 (01:06):
I worked in the music industry, entertainment, then moved to
education and then to tech.
Speaker 3 (01:12):
Right.
Speaker 4 (01:12):
So I took all that and brought that into technology.
And in twenty twenty when open Ai reached out to
me and gave me access to the invite only research
beta of GPT, I wasn't like everybody else. I was
very different because I was a creative and artists. I
was a teacher, so I had a different perspective on
(01:34):
how to use GPT. And when they saw that, they
were like, you know, can we make you an ambassador
to give us feedback research and you know, sign a
bunch of NDA's firsts right top, Yeah, and we'll give
you access to all the models before they released to
the public and give us your feedback.
Speaker 3 (01:52):
Right. And then in like twenty twenty one, they.
Speaker 4 (01:55):
Were like, oh, we're going to try this thing called
chat right, so this is two years before chat GBT,
and I was like, it's okay, it's cool. It's just
a chat, you know, what's what's the big deal? And
then now we see where it is. But now I
take all that and I look at real world problems
in the world and say, okay, when I apply AI
(02:18):
to it, how does it help the problem or make
it worse?
Speaker 3 (02:21):
And a lot of companies different industries come to me
and say, hey, how can you help us with AI?
Speaker 4 (02:26):
So it's way more than CHAGBT Now now I work
with Google and Nvidia and IBM and private companies, enterprise
and even nonprofits and charities to figure out how to
help them with AI.
Speaker 2 (02:36):
That that's like a great hook for me because I
wanted to ask what you know, what it's spotted you
to confunt, create a lobs, you a company, what's the corenation?
Speaker 3 (02:46):
So at first it was not enough of us.
Speaker 4 (02:49):
And when I say us, people from my community, people
of color, people who come from low income backgrounds, not
enough of us in those categories get to do innovative things, right.
We don't get to be in Silicon Valley like everybody else.
We don't get to play with robots, play with AI
(03:10):
and build things.
Speaker 3 (03:11):
We're just as creative as anybody else.
Speaker 4 (03:13):
And we see that with hip hop, right, a lot
of creativity comes from hip hop. You don't see a
lot of people in hip hop building robots or building AI, right,
So we wanted to change that we were like, we
see the creativity that.
Speaker 3 (03:23):
Comes from our community.
Speaker 4 (03:25):
People like it for entertainment, but what could we do
with it when it comes to innovation? And so I
recruit a lot of creative people to build AI with
me at Create Labs.
Speaker 3 (03:37):
And the goal was that's what we call the Create Labs.
Speaker 4 (03:39):
We just want people to be creative with these technologies,
either creating technologies or being creative with technologies.
Speaker 3 (03:46):
And now we leave the industry.
Speaker 4 (03:49):
When it comes to like removing bias from AI, from
adding diversity into AI and representation, we're one of the
leaders in those spaces.
Speaker 2 (03:58):
And I love that mentioned bias Ini. I think it's
one of the most important topic at the moments. This morning,
I had a talk about this. I wanted't you have
your point of view? So it's a huge problem, yes, Still,
do you think there is a solution and how many
(04:19):
eas do we need to fix it?
Speaker 1 (04:21):
If we can fix.
Speaker 4 (04:22):
Its mouth it's a bigger problem than AI because if
they're just taking a sample of data from society, the
bias in society is going to show up. And take
Wikipedia for example, many language models are trained on Wikipedia.
Speaker 3 (04:39):
Eighty seven percent.
Speaker 4 (04:41):
Of Wikipedia pages are made by white men. So whatever
biases in those Wikipedia pages are going to show up
in the LMS. So if we're more careful and thoughtful
with where we get our data from, then we can
control a little bit of what the data. The bias
is in the data. And still the bias is subconscious.
(05:02):
We don't even know our biases, so we have to
test for it, and we do evaluations. We do measurements
to see, okay, what level of bias. Let's take one
hundred answers from a chatbot that we make, how many
were biased? We find out seven Okay, that's seven percent.
That's still too high for us.
Speaker 3 (05:20):
Let's get it down to like under four percent, under
three percent.
Speaker 4 (05:24):
And we look at how is Chadgubt performing in bias
or claude and we were like, all right, if they're
at like seven to ten percent, let's get under them
and make sure that ours is better.
Speaker 3 (05:34):
And also involving experts.
Speaker 4 (05:37):
Right, So if I build, if I help Italy build
a chatpot based on your culture, and I do it
on my own, there's going to be biased because I'm
making assumptions.
Speaker 3 (05:48):
I'm guessing I don't know your culture like you do.
Speaker 4 (05:50):
But if I involve professors from your universities, teachers from
your schools.
Speaker 3 (05:56):
Then they would help me make it more accurate, you know, and.
Speaker 2 (06:00):
I love that you have a lot of experiments with
companies and non profit and a lot of different organization.
Speaker 1 (06:08):
What do you see when you talk with them about
this huge problem? I do they.
Speaker 2 (06:13):
Realize what bias in aire? Are they trying to do something?
What's the situation out there? I think I suppose that's
that's the question.
Speaker 4 (06:23):
They will you assume that there's always going to be biased,
The hallucinations are always going to be there, and it's
not something that you can control because you're using If
you're using CHAGBT, if you're using Gemini or Copilot, they
built it already and you're just using it. When you
can customize it and add your own data and fine
(06:44):
tune it, then you have more control over the responses.
And you're in the steering wheel, you know, you're in
the driver's seat. So I tell those companies, I'm like,
we're just using the technology, but we can build our
own with you. We can customize and control the levels
of bias, the levels of information, the type of information
that comes out, the tone, the voice, all of that
(07:07):
can be controlled with your guidance if you let us
build it for you from scratch, you know. But if
you're just using chatbt they built it already, you're just
the end user, you know. So I'm trying to get
them to not just be end users, but to be developers.
Speaker 2 (07:20):
And on the other side of the conversation, there are
in the huge companies usually you know American contech companies
that are building those models. What's the situation there on
that side? What do you think are they doing?
Speaker 3 (07:36):
En off?
Speaker 1 (07:37):
Is there something else that they could do about it?
Speaker 3 (07:40):
Yeah? So what I like about open ai is that
they do take.
Speaker 4 (07:44):
Active measures to address things that they don't like in
the responses, right, Well, Sam Altman just said on Twitter
that they were pulling back some updates because last one, Yeah,
it was people pleasing, right syco fancy where it's like,
you know, it's too agreeable, it's just telling you what
(08:04):
you want to hear, right, And they're like, oh, maybe
that's skewing too much that way, so let's let's pull
it back. And in real time they're like, oh, right,
let's make that change and pull that back. And I
think when people what they call building public when you
build in public, the community gives you feedback, and you
respond to that feedback, you know, and you say, okay,
(08:26):
let's make some changes.
Speaker 3 (08:27):
We hear you.
Speaker 4 (08:28):
We see what you don't like about our models, and
we're going to make some changes.
Speaker 3 (08:32):
Four point five.
Speaker 4 (08:33):
I think they said was the most emotionally intelligent and
sympathetic and empathetic. And you know, a lot of the
developers are like, oh, that doesn't help me.
Speaker 3 (08:44):
I use it for cold. I'm like, well, not everything
is cold.
Speaker 4 (08:47):
Sometimes you need it to be much yeah, you know, yeah, yeah, yeah.
Speaker 2 (08:51):
And what's the next step I mean, is that something
else that we should do right now? I mean we
as a society, people using ALI, people building AI, you know,
developments out there.
Speaker 3 (09:04):
Yeah.
Speaker 4 (09:05):
So I've always represented what's known as the no code community.
Speaker 3 (09:10):
And people who are experts.
Speaker 4 (09:13):
If you're a history teacher in a school, you shouldn't
have to give up your career to learn Java or
Python just to participate in AI.
Speaker 3 (09:25):
No stay as an expert in history.
Speaker 4 (09:27):
It will give you the tools where you can do
AI as an expert in history, you know, so these
tools where you can just communicate with them with your expertise,
and a lot of That is what's resulting into what's
called vibe coding, right, vibe coding or text prompting, where
like I give you the information and you turn that
(09:49):
into code, right, the AI turns it into code or agents.
I can write down on a whiteboard or we'll post
it and say I want my agent to do these steps.
Right you maybe that you have twenty post its on
a whiteboard, these are the steps that I want my
agent to complete. You could take a picture of that whiteboard,
may give it to chatgybt chatgubt will turn it into
(10:12):
an agentic wordflow and then you upload that to your agent.
All you did was with a pencil and a post
it to take your creativity, your ideas, and then the
AI did the rest.
Speaker 1 (10:23):
And I love that example because the creativity is the key.
Speaker 4 (10:27):
Yes, so that's your currency right now, exactly exactly, exactly exactly.
Speaker 1 (10:32):
So you just mentioned vibe coding.
Speaker 2 (10:33):
At now we know that it's like you know the
new buds wor out there.
Speaker 1 (10:38):
Is it just a bads word?
Speaker 3 (10:39):
I mean?
Speaker 2 (10:39):
Is it?
Speaker 1 (10:39):
Is it just pipe? Is it something that it's here
to stay?
Speaker 2 (10:42):
What's your take on the whole vibe coding and bymarketing
that it's happening right now.
Speaker 4 (10:47):
It's it's here to stay because it's just the next
iteration of prompting. Where you have text to text, you
have texted image, text to video where you write a
description and it generated it's a video output.
Speaker 3 (11:01):
So now it's text to code.
Speaker 4 (11:03):
You write a string of text in playing language, any language,
and it writes a string of code to execute your
idea or vision. And the better their real time voice gets,
it's going to be no different than a podcast. You're
going to have a conversation where like, all right, let's
me and you sit down and come up with an
idea for an app together and have a conversation while
(11:24):
the listening agent is.
Speaker 3 (11:26):
Writing the code in real time while we talk. Right.
Speaker 4 (11:30):
So, think of from an accessibility standpoint, people who who
don't have all their capabilities with their hands to type.
You know, I still don't know how to code, and
look how far I've gotten in this industry over the
last five years without coding a single line.
Speaker 1 (11:46):
That's why I use the word magic.
Speaker 2 (11:49):
I mean to me, in twenty twenty four, ideas still
feels like magic.
Speaker 1 (11:53):
It's unvalible what we can do right now with this technology.
Speaker 2 (11:56):
But I suppose that the right you know out there
developers or so name to the show, and maybe some
of them aren't worried about their future. Yes, you know
the whole thing. Everyone is going to be a colder.
You know, code is gonna be shitty. Yeah, we're going
to do our jobs.
Speaker 1 (12:12):
What to think about that?
Speaker 2 (12:13):
I mean, is there a future where we can have
both you know, the vibe coders and and you know
old school developers out there.
Speaker 4 (12:20):
Well, let's let's compare it to the automotive industry. I
drive an electric vehicle. I drive a Rivian, Right, I
love the car. I don't know what's going on under
the hood. Actually I open the hood, there's nothing there.
It's another trunk. Right, there's a battery underneath the car.
I don't know how that works. I drive it. I
know how to drive it very well. But you will
(12:42):
always need mechanics. They know how to fix that car.
Speaker 3 (12:45):
Right.
Speaker 4 (12:46):
So I can be a really good driver as a
Vibe coder, right, I could be a really good prompt engineer,
but you will always need the mechanic that knows how
to fix the code in case something goes wrong. They
knows what's happening under the hood. You know, and they
have their own technologies. Now, if you go to the
Rivian factory, they're also using AI as the mechanics, but
(13:08):
a different kind diagnostics testing safety protocols, right simulations for safety.
So they have their own versions of vibe coding on
their side for their protocols that they need to test for.
That's different than mine as a driver, you know. So
when I speak to the software developers, you know, the
other ambassadors that are expert software engineers, unit testing, debugging
(13:33):
is and even with vibe coding, like I like Replet,
I like Lovable hay Boss is another good one that
I use bolts that will get most people in this room,
they will get maybe seventy percent of the way they're
completing their vibe coding app. That last thirty percent you
need to find a developer and a coder that can
(13:54):
help you finish it, get it to production, publish it
to the Apple Store, connect it to a database, you know,
connected to sso for logging, like those are things like
I just made the pretty colors, someone else has to
do the rest.
Speaker 2 (14:06):
And you know, somehow I suppose we had maybe the
same no same people, you know, but we had people
also worried when I don't know, we we went from
assembly to I don't know, a bytone, and then you
had JavaScript and there was always someone saying, oh, you know,
this is shit, this is not real coding. We wear
the real coder back then. You know, I thought it's
maybe someone is always worrying about, you know, something new
(14:28):
right on the market.
Speaker 4 (14:29):
So but I mean there were also candle makers in Italy, right,
and what happened to them when we started with electricity? Yeah,
you know, And I come from the music industry.
Speaker 3 (14:38):
I remember when Sorito came.
Speaker 4 (14:41):
So I would go with DJs and they would be
carrying crates of records and then they see a guy
walk in with a laptop and they're like, oh my
back hurts. He's just carrying a small MacBook Air Like,
but if I carry the laptop, am I still a DJ?
Speaker 3 (14:58):
Right?
Speaker 4 (14:58):
And then eventually now everyone brings a laptop to the club,
you know. So it's just an evolution of tools, you know,
to do the same thing.
Speaker 3 (15:06):
You're still a DJ.
Speaker 4 (15:07):
Jazz Jeff, I don't know if they know Jazz Jeff
out here, who was Will Smith's DJ, very famous now
on the DJ court. He said when I'm doing a
big music festival with ten thousand people.
Speaker 3 (15:16):
The people all the.
Speaker 4 (15:16):
Way in the back have no idea what I'm using.
They're just enjoying the experience. You know, when I make
a beautiful video or film and AI, you don't know
what I used.
Speaker 3 (15:27):
So long as you're enjoying the experience, you.
Speaker 1 (15:28):
Know, I totally agree with you.
Speaker 2 (15:33):
It was funny when I was googling you and preparing
for this interview and find information and you know, cool
topics to talk about with you. On several websites, I
found this sentence that you are known as one of
the first prompt engineers out there, and and.
Speaker 1 (15:51):
It was huge, even maybe a couple of years ago.
Speaker 2 (15:55):
Yeah, using the words, you know, prompt engineer or prompt engineer,
is it.
Speaker 1 (16:01):
Something that is still there.
Speaker 2 (16:03):
I mean, in my experience, you're, My feeling using AI
tools every day is that they are getting better and
better and better.
Speaker 1 (16:11):
So I don't need.
Speaker 2 (16:12):
To be you know, like super NERD when I'm building,
I'm prompt.
Speaker 1 (16:17):
They really understand me. They understand what's my need.
Speaker 2 (16:21):
So I can go, you know with a like really
easy and simple prompt and I still get the result.
Speaker 1 (16:27):
So it's from the engineering just changing, evolving, or it's
just somehow these appearing.
Speaker 4 (16:33):
No, So there's different levels of prompting and working with
the best prompt engineers in the world and seeing how
they prompt and how I've learned from them in the
early days. Now, when I say early, this is twenty twenty,
so think how many years before CHAGBT that work was,
right and when most people went on the open AI
website and they learned how to prompt from the open
(16:55):
EY website. My mentors wrote those instructions, right. Andrew Mayne
and Russ Folks Smith and all the other ambassadors wrote
the initial instructions for everyone to learn how to prompt
engineer right in those early days. I like calling your
prompt design but prompt engineering same thing. And so if
you're using CHADGBT, you could write a single sentence to
get what you want right. But when we made the
(17:18):
Small Business AI Coach for MasterCard, that prompt is about
thirty pages long.
Speaker 2 (17:25):
Wow, and you mean single prompt the single prompts that's crazy,
And maybe fifteen.
Speaker 3 (17:31):
Pages are what not to do are the guardrails? I
love it.
Speaker 4 (17:36):
And the tone of voice of the MasterCard chatbot to
sound like everything else that MasterCard writes on their website,
their customer service. That was a meeting with twenty people
at MasterCard to write that one paragraph from the prompts.
Speaker 3 (17:50):
So that's how serious you have to.
Speaker 4 (17:52):
Take prompting frumpt engineering when it's creating a corporate product
for enterprise to release.
Speaker 3 (17:58):
You know, it's not just a single sentence.
Speaker 4 (18:01):
Yeah, a lot of descriptions, a lot of work goes
into every single set. And then if something goes wrong,
we go through those thirty pages and we change one
word and it changes the output.
Speaker 2 (18:11):
And do you think that that's still true even for
like small businesses out there, freelancers. I mean most of
the people listening to show and also here at the conference,
they don't corporate.
Speaker 1 (18:24):
Maybe it's five people team.
Speaker 2 (18:27):
So do you think that's still important even in those situations,
your small teams freelancers.
Speaker 4 (18:32):
Yeah, So I would say the best thing for them
to do is their thirty pages would be give your
life story, your complete business plan, that's your PRIs right,
write your business plan to chatch to bet give it
the core memory that it needs to give you advice
on future decisions that you make with the business right,
(18:55):
and then now chatch TOBT stores memory. Right, so do
this for years for you years you just brainstorm with
tragibt brainstorm and then now it has that core memory.
So then when a new deal comes up, right, and say, oh,
I have this big partnership, you know, opportunity with you
know AI week, what do I do? And it's like, well,
I can guide you based on how you do your business.
(19:16):
I can help you write your proposal because it has
learned that over time. And if you're an author, if
you're a creative writer, teaching it your style is your
prompt engineering. You know, practice It takes a lot of
work because you want to see anything that comes out
to look like your writing style, and that's going to
take a lot of work to teach it how.
Speaker 3 (19:37):
You write, you know.
Speaker 4 (19:38):
So there's a lot that they can accomplish. But that's
a good starting point. But then there's so many other
tools that you have to use. And I'm going to
go over there in my workshop of like you start
in TRAGBT, but then you take that out and put
it into mid journey, You take that out and put
that into VIO and Sora.
Speaker 3 (19:52):
You take that out and put it into Claude. There's
so many other.
Speaker 4 (19:55):
Tools that you have to use, but you're starting point
could be TRAGITBT with your core prompt.
Speaker 1 (20:00):
I love it, guys, this is golden. I love it.
Speaker 2 (20:05):
If you look at the future of generit Tyvai, what's
the thing, the trend, the technology that excited you the most.
Speaker 1 (20:14):
At the moment.
Speaker 3 (20:16):
I'm excited by film.
Speaker 4 (20:17):
I'm excited by entertainment because I don't like the idea
of Hollywood. I like that you can be what's known
as a world builder, right the next George Lucas or
James Cameron or Stan Lee. You don't have to like
be a billion dollar producer in Hollywood.
Speaker 3 (20:37):
Just to make beautiful things.
Speaker 4 (20:39):
You know, A child, a fifteen year old child in Milan,
with these tools can make the next Marvel, you know,
And now you have the tools to do that. And
if you want a show, if you want an Italian superhero,
you don't have to wait for Marvel to meet in
a Disney boom.
Speaker 1 (21:00):
Just build it.
Speaker 4 (21:01):
Just build it, Just build it with these tools that
I'm watching right now on that screen, and go build it.
Go build it and get your own following.
Speaker 2 (21:10):
Do you think this is something that we can do today.
I mean, how far away are away from from this?
Speaker 4 (21:16):
Right now?
Speaker 2 (21:17):
You know, building a movie in his own room.
Speaker 4 (21:21):
A kid building a movie in his own room, in
his own cartoon. There's a good example. I have a
ten year old son. He watches a Digital Circus right
and it was a YouTube show.
Speaker 3 (21:33):
Right now, there's.
Speaker 4 (21:35):
Four episodes ever made and it has gone so viral
on YouTube that they already have toys and Netflix picked
up dished their four episodes, not four seasons, four episodes,
And it could have been a fifteen year old that
made Digital Circus with AI who knows.
Speaker 3 (21:52):
Like it didn't look like a very sophisticated enemy.
Speaker 1 (21:55):
So it's real, it's opening, it's opening now.
Speaker 3 (21:57):
It's happening now.
Speaker 4 (21:58):
And if you if you put the follow only behind it,
the viewership, and you you know, the fan base, the
loyal fan base of someone who's making AI content can
get a Netflix deal tomorrow with their AI content, and
you start a new franchise.
Speaker 3 (22:11):
You don't have to move out to.
Speaker 4 (22:12):
Hollywood anymore and do the whole networking and connections and
sacrificing things about yourself that you don't like because you
want to get your foot into the door.
Speaker 3 (22:21):
That's done. That's completely done. Everyone.
Speaker 4 (22:23):
The access to knowledge from universities is out there through
CHATCHUBT and access to creative tools that were gate kept.
You know, the gatekeeping in Hollywood is now democratized too,
and now everyone has access.
Speaker 1 (22:38):
The last one.
Speaker 2 (22:39):
Is there something that concerns you looking at there of
this technology?
Speaker 3 (22:43):
Yeah? You in connection.
Speaker 4 (22:44):
I've met more people because of my tours with AI
speaking than I ever have before. And I love that
and I love that I get to experience your country.
But I don't like that people turn to it when
they're lonely, or they go to it for relationship.
Speaker 3 (23:00):
Advice or therapy. I built the therapy bot for a hospital.
Speaker 4 (23:05):
And I did it with the medical experts. I think
with a CHATGBT, that's like a one size fits all
general tool.
Speaker 3 (23:15):
It's not built for that you know you want.
Speaker 4 (23:19):
Because with the therapy about that we made, if something
concerning comes up, like self harm, it alerts the hospital,
It alerts your doctor, it alerts your counselor to call
you and say, hey, I just saw this in the
chatbout is everything okay? If you do that with the
chat YOUBT and say I want to hurt myself there's
no one's gonna call you on the other side and say, hey,
I saw this come up.
Speaker 3 (23:39):
Is everything okay? So you gotta be careful.
Speaker 4 (23:42):
And also these AI tools are just going to tell
you what you want to hear, and that's a scary thing.
Bias again, the bias the people he's in. You know, like,
I just want to be by myself. You don't want
the chatbout to be like, Okay, be by yourself, lock
yourself in that room. No, you want real connections with
real people. That should not go. If anything, it should
take away the boring work. We don't want to do
the spreadsheets, the reports, the proposals, so we could be
(24:03):
outside more and connect with real people and go visit,
you know, churches and see sunsets.
Speaker 1 (24:10):
I agree, I love it. It was a great pleasure.
Speaker 3 (24:14):
Thank you for watching. Probably be on the show all right,
thank you,
Chow, everyone, and welcome back to the show.
Speaker 2 (00:02):
I'm Ralph, your host, and my guest today is Abron Maldonado.
Speaker 3 (00:06):
Good to see you. I'm glad to be back at
AI week. I love it here.
Speaker 4 (00:10):
How he's going, How he's good, It's going great. I
always take some time to enjoy the culture first. So
I went to Florence and then came here to my
line and wrote to be a part of this, and
it was great the first time and the last time
I came here, AI was in a different place. And
you know, I call it like dog years with AI
because one year in real time and AI is like.
Speaker 3 (00:31):
Five years, is it? So a lot has happened between
now and then.
Speaker 1 (00:34):
I agree, I love it. I want to start with
a with a simple one.
Speaker 2 (00:38):
I just wanted to know more about your journey with
with with with the I.
Speaker 1 (00:41):
So where did it all started and where are you now?
Speaker 4 (00:45):
So in the beginning I came there's a there's a
word in in English, a phrase I don't know how
it translates here, multi hyphen it and multi hyphen it
means people that have multiple interests and where.
Speaker 3 (00:58):
I have done a lot of things, had different career experience.
Speaker 1 (01:01):
People need ye.
Speaker 3 (01:02):
Yeah, so I was. I worked as an artist.
Speaker 4 (01:06):
I worked in the music industry, entertainment, then moved to
education and then to tech.
Speaker 3 (01:12):
Right.
Speaker 4 (01:12):
So I took all that and brought that into technology.
And in twenty twenty when open Ai reached out to
me and gave me access to the invite only research
beta of GPT, I wasn't like everybody else. I was
very different because I was a creative and artists. I
was a teacher, so I had a different perspective on
(01:34):
how to use GPT. And when they saw that, they
were like, you know, can we make you an ambassador
to give us feedback research and you know, sign a
bunch of NDA's firsts right top, Yeah, and we'll give
you access to all the models before they released to
the public and give us your feedback.
Speaker 3 (01:52):
Right. And then in like twenty twenty one, they.
Speaker 4 (01:55):
Were like, oh, we're going to try this thing called
chat right, so this is two years before chat GBT,
and I was like, it's okay, it's cool. It's just
a chat, you know, what's what's the big deal? And
then now we see where it is. But now I
take all that and I look at real world problems
in the world and say, okay, when I apply AI
(02:18):
to it, how does it help the problem or make
it worse?
Speaker 3 (02:21):
And a lot of companies different industries come to me
and say, hey, how can you help us with AI?
Speaker 4 (02:26):
So it's way more than CHAGBT Now now I work
with Google and Nvidia and IBM and private companies, enterprise
and even nonprofits and charities to figure out how to
help them with AI.
Speaker 2 (02:36):
That that's like a great hook for me because I
wanted to ask what you know, what it's spotted you
to confunt, create a lobs, you a company, what's the corenation?
Speaker 3 (02:46):
So at first it was not enough of us.
Speaker 4 (02:49):
And when I say us, people from my community, people
of color, people who come from low income backgrounds, not
enough of us in those categories get to do innovative things, right.
We don't get to be in Silicon Valley like everybody else.
We don't get to play with robots, play with AI
(03:10):
and build things.
Speaker 3 (03:11):
We're just as creative as anybody else.
Speaker 4 (03:13):
And we see that with hip hop, right, a lot
of creativity comes from hip hop. You don't see a
lot of people in hip hop building robots or building AI, right,
So we wanted to change that we were like, we
see the creativity that.
Speaker 3 (03:23):
Comes from our community.
Speaker 4 (03:25):
People like it for entertainment, but what could we do
with it when it comes to innovation? And so I
recruit a lot of creative people to build AI with
me at Create Labs.
Speaker 3 (03:37):
And the goal was that's what we call the Create Labs.
Speaker 4 (03:39):
We just want people to be creative with these technologies,
either creating technologies or being creative with technologies.
Speaker 3 (03:46):
And now we leave the industry.
Speaker 4 (03:49):
When it comes to like removing bias from AI, from
adding diversity into AI and representation, we're one of the
leaders in those spaces.
Speaker 2 (03:58):
And I love that mentioned bias Ini. I think it's
one of the most important topic at the moments. This morning,
I had a talk about this. I wanted't you have
your point of view? So it's a huge problem, yes, Still,
do you think there is a solution and how many
(04:19):
eas do we need to fix it?
Speaker 1 (04:21):
If we can fix.
Speaker 4 (04:22):
Its mouth it's a bigger problem than AI because if
they're just taking a sample of data from society, the
bias in society is going to show up. And take
Wikipedia for example, many language models are trained on Wikipedia.
Speaker 3 (04:39):
Eighty seven percent.
Speaker 4 (04:41):
Of Wikipedia pages are made by white men. So whatever
biases in those Wikipedia pages are going to show up
in the LMS. So if we're more careful and thoughtful
with where we get our data from, then we can
control a little bit of what the data. The bias
is in the data. And still the bias is subconscious.
(05:02):
We don't even know our biases, so we have to
test for it, and we do evaluations. We do measurements
to see, okay, what level of bias. Let's take one
hundred answers from a chatbot that we make, how many
were biased? We find out seven Okay, that's seven percent.
That's still too high for us.
Speaker 3 (05:20):
Let's get it down to like under four percent, under
three percent.
Speaker 4 (05:24):
And we look at how is Chadgubt performing in bias
or claude and we were like, all right, if they're
at like seven to ten percent, let's get under them
and make sure that ours is better.
Speaker 3 (05:34):
And also involving experts.
Speaker 4 (05:37):
Right, So if I build, if I help Italy build
a chatpot based on your culture, and I do it
on my own, there's going to be biased because I'm
making assumptions.
Speaker 3 (05:48):
I'm guessing I don't know your culture like you do.
Speaker 4 (05:50):
But if I involve professors from your universities, teachers from
your schools.
Speaker 3 (05:56):
Then they would help me make it more accurate, you know, and.
Speaker 2 (06:00):
I love that you have a lot of experiments with
companies and non profit and a lot of different organization.
Speaker 1 (06:08):
What do you see when you talk with them about
this huge problem? I do they.
Speaker 2 (06:13):
Realize what bias in aire? Are they trying to do something?
What's the situation out there? I think I suppose that's
that's the question.
Speaker 4 (06:23):
They will you assume that there's always going to be biased,
The hallucinations are always going to be there, and it's
not something that you can control because you're using If
you're using CHAGBT, if you're using Gemini or Copilot, they
built it already and you're just using it. When you
can customize it and add your own data and fine
(06:44):
tune it, then you have more control over the responses.
And you're in the steering wheel, you know, you're in
the driver's seat. So I tell those companies, I'm like,
we're just using the technology, but we can build our
own with you. We can customize and control the levels
of bias, the levels of information, the type of information
that comes out, the tone, the voice, all of that
(07:07):
can be controlled with your guidance if you let us
build it for you from scratch, you know. But if
you're just using chatbt they built it already, you're just
the end user, you know. So I'm trying to get
them to not just be end users, but to be developers.
Speaker 2 (07:20):
And on the other side of the conversation, there are
in the huge companies usually you know American contech companies
that are building those models. What's the situation there on
that side? What do you think are they doing?
Speaker 3 (07:36):
En off?
Speaker 1 (07:37):
Is there something else that they could do about it?
Speaker 3 (07:40):
Yeah? So what I like about open ai is that
they do take.
Speaker 4 (07:44):
Active measures to address things that they don't like in
the responses, right, Well, Sam Altman just said on Twitter
that they were pulling back some updates because last one, Yeah,
it was people pleasing, right syco fancy where it's like,
you know, it's too agreeable, it's just telling you what
(08:04):
you want to hear, right, And they're like, oh, maybe
that's skewing too much that way, so let's let's pull
it back. And in real time they're like, oh, right,
let's make that change and pull that back. And I
think when people what they call building public when you
build in public, the community gives you feedback, and you
respond to that feedback, you know, and you say, okay,
(08:26):
let's make some changes.
Speaker 3 (08:27):
We hear you.
Speaker 4 (08:28):
We see what you don't like about our models, and
we're going to make some changes.
Speaker 3 (08:32):
Four point five.
Speaker 4 (08:33):
I think they said was the most emotionally intelligent and
sympathetic and empathetic. And you know, a lot of the
developers are like, oh, that doesn't help me.
Speaker 3 (08:44):
I use it for cold. I'm like, well, not everything
is cold.
Speaker 4 (08:47):
Sometimes you need it to be much yeah, you know, yeah, yeah, yeah.
Speaker 2 (08:51):
And what's the next step I mean, is that something
else that we should do right now? I mean we
as a society, people using ALI, people building AI, you know,
developments out there.
Speaker 3 (09:04):
Yeah.
Speaker 4 (09:05):
So I've always represented what's known as the no code community.
Speaker 3 (09:10):
And people who are experts.
Speaker 4 (09:13):
If you're a history teacher in a school, you shouldn't
have to give up your career to learn Java or
Python just to participate in AI.
Speaker 3 (09:25):
No stay as an expert in history.
Speaker 4 (09:27):
It will give you the tools where you can do
AI as an expert in history, you know, so these
tools where you can just communicate with them with your expertise,
and a lot of That is what's resulting into what's
called vibe coding, right, vibe coding or text prompting, where
like I give you the information and you turn that
(09:49):
into code, right, the AI turns it into code or agents.
I can write down on a whiteboard or we'll post
it and say I want my agent to do these steps.
Right you maybe that you have twenty post its on
a whiteboard, these are the steps that I want my
agent to complete. You could take a picture of that whiteboard,
may give it to chatgybt chatgubt will turn it into
(10:12):
an agentic wordflow and then you upload that to your agent.
All you did was with a pencil and a post
it to take your creativity, your ideas, and then the
AI did the rest.
Speaker 1 (10:23):
And I love that example because the creativity is the key.
Speaker 4 (10:27):
Yes, so that's your currency right now, exactly exactly, exactly exactly.
Speaker 1 (10:32):
So you just mentioned vibe coding.
Speaker 2 (10:33):
At now we know that it's like you know the
new buds wor out there.
Speaker 1 (10:38):
Is it just a bads word?
Speaker 3 (10:39):
I mean?
Speaker 2 (10:39):
Is it?
Speaker 1 (10:39):
Is it just pipe? Is it something that it's here
to stay?
Speaker 2 (10:42):
What's your take on the whole vibe coding and bymarketing
that it's happening right now.
Speaker 4 (10:47):
It's it's here to stay because it's just the next
iteration of prompting. Where you have text to text, you
have texted image, text to video where you write a
description and it generated it's a video output.
Speaker 3 (11:01):
So now it's text to code.
Speaker 4 (11:03):
You write a string of text in playing language, any language,
and it writes a string of code to execute your
idea or vision. And the better their real time voice gets,
it's going to be no different than a podcast. You're
going to have a conversation where like, all right, let's
me and you sit down and come up with an
idea for an app together and have a conversation while
(11:24):
the listening agent is.
Speaker 3 (11:26):
Writing the code in real time while we talk. Right.
Speaker 4 (11:30):
So, think of from an accessibility standpoint, people who who
don't have all their capabilities with their hands to type.
You know, I still don't know how to code, and
look how far I've gotten in this industry over the
last five years without coding a single line.
Speaker 1 (11:46):
That's why I use the word magic.
Speaker 2 (11:49):
I mean to me, in twenty twenty four, ideas still
feels like magic.
Speaker 1 (11:53):
It's unvalible what we can do right now with this technology.
Speaker 2 (11:56):
But I suppose that the right you know out there
developers or so name to the show, and maybe some
of them aren't worried about their future. Yes, you know
the whole thing. Everyone is going to be a colder.
You know, code is gonna be shitty. Yeah, we're going
to do our jobs.
Speaker 1 (12:12):
What to think about that?
Speaker 2 (12:13):
I mean, is there a future where we can have
both you know, the vibe coders and and you know
old school developers out there.
Speaker 4 (12:20):
Well, let's let's compare it to the automotive industry. I
drive an electric vehicle. I drive a Rivian, Right, I
love the car. I don't know what's going on under
the hood. Actually I open the hood, there's nothing there.
It's another trunk. Right, there's a battery underneath the car.
I don't know how that works. I drive it. I
know how to drive it very well. But you will
(12:42):
always need mechanics. They know how to fix that car.
Speaker 3 (12:45):
Right.
Speaker 4 (12:46):
So I can be a really good driver as a
Vibe coder, right, I could be a really good prompt engineer,
but you will always need the mechanic that knows how
to fix the code in case something goes wrong. They
knows what's happening under the hood. You know, and they
have their own technologies. Now, if you go to the
Rivian factory, they're also using AI as the mechanics, but
(13:08):
a different kind diagnostics testing safety protocols, right simulations for safety.
So they have their own versions of vibe coding on
their side for their protocols that they need to test for.
That's different than mine as a driver, you know. So
when I speak to the software developers, you know, the
other ambassadors that are expert software engineers, unit testing, debugging
(13:33):
is and even with vibe coding, like I like Replet,
I like Lovable hay Boss is another good one that
I use bolts that will get most people in this room,
they will get maybe seventy percent of the way they're
completing their vibe coding app. That last thirty percent you
need to find a developer and a coder that can
(13:54):
help you finish it, get it to production, publish it
to the Apple Store, connect it to a database, you know,
connected to sso for logging, like those are things like
I just made the pretty colors, someone else has to
do the rest.
Speaker 2 (14:06):
And you know, somehow I suppose we had maybe the
same no same people, you know, but we had people
also worried when I don't know, we we went from
assembly to I don't know, a bytone, and then you
had JavaScript and there was always someone saying, oh, you know,
this is shit, this is not real coding. We wear
the real coder back then. You know, I thought it's
maybe someone is always worrying about, you know, something new
(14:28):
right on the market.
Speaker 4 (14:29):
So but I mean there were also candle makers in Italy, right,
and what happened to them when we started with electricity? Yeah,
you know, And I come from the music industry.
Speaker 3 (14:38):
I remember when Sorito came.
Speaker 4 (14:41):
So I would go with DJs and they would be
carrying crates of records and then they see a guy
walk in with a laptop and they're like, oh my
back hurts. He's just carrying a small MacBook Air Like,
but if I carry the laptop, am I still a DJ?
Speaker 3 (14:58):
Right?
Speaker 4 (14:58):
And then eventually now everyone brings a laptop to the club,
you know. So it's just an evolution of tools, you know,
to do the same thing.
Speaker 3 (15:06):
You're still a DJ.
Speaker 4 (15:07):
Jazz Jeff, I don't know if they know Jazz Jeff
out here, who was Will Smith's DJ, very famous now
on the DJ court. He said when I'm doing a
big music festival with ten thousand people.
Speaker 3 (15:16):
The people all the.
Speaker 4 (15:16):
Way in the back have no idea what I'm using.
They're just enjoying the experience. You know, when I make
a beautiful video or film and AI, you don't know
what I used.
Speaker 3 (15:27):
So long as you're enjoying the experience, you.
Speaker 1 (15:28):
Know, I totally agree with you.
Speaker 2 (15:33):
It was funny when I was googling you and preparing
for this interview and find information and you know, cool
topics to talk about with you. On several websites, I
found this sentence that you are known as one of
the first prompt engineers out there, and and.
Speaker 1 (15:51):
It was huge, even maybe a couple of years ago.
Speaker 2 (15:55):
Yeah, using the words, you know, prompt engineer or prompt engineer,
is it.
Speaker 1 (16:01):
Something that is still there.
Speaker 2 (16:03):
I mean, in my experience, you're, My feeling using AI
tools every day is that they are getting better and
better and better.
Speaker 1 (16:11):
So I don't need.
Speaker 2 (16:12):
To be you know, like super NERD when I'm building,
I'm prompt.
Speaker 1 (16:17):
They really understand me. They understand what's my need.
Speaker 2 (16:21):
So I can go, you know with a like really
easy and simple prompt and I still get the result.
Speaker 1 (16:27):
So it's from the engineering just changing, evolving, or it's
just somehow these appearing.
Speaker 4 (16:33):
No, So there's different levels of prompting and working with
the best prompt engineers in the world and seeing how
they prompt and how I've learned from them in the
early days. Now, when I say early, this is twenty twenty,
so think how many years before CHAGBT that work was,
right and when most people went on the open AI
website and they learned how to prompt from the open
(16:55):
EY website. My mentors wrote those instructions, right. Andrew Mayne
and Russ Folks Smith and all the other ambassadors wrote
the initial instructions for everyone to learn how to prompt
engineer right in those early days. I like calling your
prompt design but prompt engineering same thing. And so if
you're using CHADGBT, you could write a single sentence to
get what you want right. But when we made the
(17:18):
Small Business AI Coach for MasterCard, that prompt is about
thirty pages long.
Speaker 2 (17:25):
Wow, and you mean single prompt the single prompts that's crazy,
And maybe fifteen.
Speaker 3 (17:31):
Pages are what not to do are the guardrails? I
love it.
Speaker 4 (17:36):
And the tone of voice of the MasterCard chatbot to
sound like everything else that MasterCard writes on their website,
their customer service. That was a meeting with twenty people
at MasterCard to write that one paragraph from the prompts.
Speaker 3 (17:50):
So that's how serious you have to.
Speaker 4 (17:52):
Take prompting frumpt engineering when it's creating a corporate product
for enterprise to release.
Speaker 3 (17:58):
You know, it's not just a single sentence.
Speaker 4 (18:01):
Yeah, a lot of descriptions, a lot of work goes
into every single set. And then if something goes wrong,
we go through those thirty pages and we change one
word and it changes the output.
Speaker 2 (18:11):
And do you think that that's still true even for
like small businesses out there, freelancers. I mean most of
the people listening to show and also here at the conference,
they don't corporate.
Speaker 1 (18:24):
Maybe it's five people team.
Speaker 2 (18:27):
So do you think that's still important even in those situations,
your small teams freelancers.
Speaker 4 (18:32):
Yeah, So I would say the best thing for them
to do is their thirty pages would be give your
life story, your complete business plan, that's your PRIs right,
write your business plan to chatch to bet give it
the core memory that it needs to give you advice
on future decisions that you make with the business right,
(18:55):
and then now chatch TOBT stores memory. Right, so do
this for years for you years you just brainstorm with
tragibt brainstorm and then now it has that core memory.
So then when a new deal comes up, right, and say, oh,
I have this big partnership, you know, opportunity with you
know AI week, what do I do? And it's like, well,
I can guide you based on how you do your business.
(19:16):
I can help you write your proposal because it has
learned that over time. And if you're an author, if
you're a creative writer, teaching it your style is your
prompt engineering. You know, practice It takes a lot of
work because you want to see anything that comes out
to look like your writing style, and that's going to
take a lot of work to teach it how.
Speaker 3 (19:37):
You write, you know.
Speaker 4 (19:38):
So there's a lot that they can accomplish. But that's
a good starting point. But then there's so many other
tools that you have to use. And I'm going to
go over there in my workshop of like you start
in TRAGBT, but then you take that out and put
it into mid journey, You take that out and put
that into VIO and Sora.
Speaker 3 (19:52):
You take that out and put it into Claude. There's
so many other.
Speaker 4 (19:55):
Tools that you have to use, but you're starting point
could be TRAGITBT with your core prompt.
Speaker 1 (20:00):
I love it, guys, this is golden. I love it.
Speaker 2 (20:05):
If you look at the future of generit Tyvai, what's
the thing, the trend, the technology that excited you the most.
Speaker 1 (20:14):
At the moment.
Speaker 3 (20:16):
I'm excited by film.
Speaker 4 (20:17):
I'm excited by entertainment because I don't like the idea
of Hollywood. I like that you can be what's known
as a world builder, right the next George Lucas or
James Cameron or Stan Lee. You don't have to like
be a billion dollar producer in Hollywood.
Speaker 3 (20:37):
Just to make beautiful things.
Speaker 4 (20:39):
You know, A child, a fifteen year old child in Milan,
with these tools can make the next Marvel, you know,
And now you have the tools to do that. And
if you want a show, if you want an Italian superhero,
you don't have to wait for Marvel to meet in
a Disney boom.
Speaker 1 (21:00):
Just build it.
Speaker 4 (21:01):
Just build it, Just build it with these tools that
I'm watching right now on that screen, and go build it.
Go build it and get your own following.
Speaker 2 (21:10):
Do you think this is something that we can do today.
I mean, how far away are away from from this?
Speaker 4 (21:16):
Right now?
Speaker 2 (21:17):
You know, building a movie in his own room.
Speaker 4 (21:21):
A kid building a movie in his own room, in
his own cartoon. There's a good example. I have a
ten year old son. He watches a Digital Circus right
and it was a YouTube show.
Speaker 3 (21:33):
Right now, there's.
Speaker 4 (21:35):
Four episodes ever made and it has gone so viral
on YouTube that they already have toys and Netflix picked
up dished their four episodes, not four seasons, four episodes,
And it could have been a fifteen year old that
made Digital Circus with AI who knows.
Speaker 3 (21:52):
Like it didn't look like a very sophisticated enemy.
Speaker 1 (21:55):
So it's real, it's opening, it's opening now.
Speaker 3 (21:57):
It's happening now.
Speaker 4 (21:58):
And if you if you put the follow only behind it,
the viewership, and you you know, the fan base, the
loyal fan base of someone who's making AI content can
get a Netflix deal tomorrow with their AI content, and
you start a new franchise.
Speaker 3 (22:11):
You don't have to move out to.
Speaker 4 (22:12):
Hollywood anymore and do the whole networking and connections and
sacrificing things about yourself that you don't like because you
want to get your foot into the door.
Speaker 3 (22:21):
That's done. That's completely done. Everyone.
Speaker 4 (22:23):
The access to knowledge from universities is out there through
CHATCHUBT and access to creative tools that were gate kept.
You know, the gatekeeping in Hollywood is now democratized too,
and now everyone has access.
Speaker 1 (22:38):
The last one.
Speaker 2 (22:39):
Is there something that concerns you looking at there of
this technology?
Speaker 3 (22:43):
Yeah? You in connection.
Speaker 4 (22:44):
I've met more people because of my tours with AI
speaking than I ever have before. And I love that
and I love that I get to experience your country.
But I don't like that people turn to it when
they're lonely, or they go to it for relationship.
Speaker 3 (23:00):
Advice or therapy. I built the therapy bot for a hospital.
Speaker 4 (23:05):
And I did it with the medical experts. I think
with a CHATGBT, that's like a one size fits all
general tool.
Speaker 3 (23:15):
It's not built for that you know you want.
Speaker 4 (23:19):
Because with the therapy about that we made, if something
concerning comes up, like self harm, it alerts the hospital,
It alerts your doctor, it alerts your counselor to call
you and say, hey, I just saw this in the
chatbout is everything okay? If you do that with the
chat YOUBT and say I want to hurt myself there's
no one's gonna call you on the other side and say, hey,
I saw this come up.
Speaker 3 (23:39):
Is everything okay? So you gotta be careful.
Speaker 4 (23:42):
And also these AI tools are just going to tell
you what you want to hear, and that's a scary thing.
Bias again, the bias the people he's in. You know, like,
I just want to be by myself. You don't want
the chatbout to be like, Okay, be by yourself, lock
yourself in that room. No, you want real connections with
real people. That should not go. If anything, it should
take away the boring work. We don't want to do
the spreadsheets, the reports, the proposals, so we could be
(24:03):
outside more and connect with real people and go visit,
you know, churches and see sunsets.
Speaker 1 (24:10):
I agree, I love it. It was a great pleasure.
Speaker 3 (24:14):
Thank you for watching. Probably be on the show all right,
thank you,
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
New Heights with Jason & Travis Kelce
Football’s funniest family duo — Jason Kelce of the Philadelphia Eagles and Travis Kelce of the Kansas City Chiefs — team up to provide next-level access to life in the league as it unfolds. The two brothers and Super Bowl champions drop weekly insights about the weekly slate of games and share their INSIDE perspectives on trending NFL news and sports headlines. They also endlessly rag on each other as brothers do, chat the latest in pop culture and welcome some very popular and well-known friends to chat with them. Check out new episodes every Wednesday. Follow New Heights on the Wondery App, YouTube or wherever you get your podcasts. You can listen to new episodes early and ad-free, and get exclusive content on Wondery+. Join Wondery+ in the Wondery App, Apple Podcasts or Spotify. And join our new membership for a unique fan experience by going to the New Heights YouTube channel now!