May 21, 2024 • 61 mins
Embark on an enlightening expedition with Clayton Smith of Minimap AI as he charts the uncharted territories of the media landscape. By transforming the way we visualize news distribution, Minimap AI's content cartography promises to arm you with a deeper understanding of the stories shaping our world, far surpassing the left-right dichotomy.
This episode maps out a revolutionary system that spatially arranges news topics, revealing trends and the breadth of coverage at a glance. The discussion with Clayton navigates through examples such as Tesla's activities in China, illustrating the profound impact a single story can have across multiple sectors. The conversation also sheds light on the platform's ability to connect seemingly unrelated stories, offering a nuanced perspective often lost in the chaos of traditional media consumption.
Lastly, we trace Minimap's evolution from a product search mapping tool to a beacon in the dense fog of news and social media content. Clayton and I dissect the challenges companies face in aligning their brand with their content output and discuss how Minimap's innovative algorithms can detect discrepancies that may slip through the cracks. By the end of our journey, you'll be equipped with insights into how Minimap AI not only redefines content consumption but also serves as a vital compass in the quest for authentic brand and media interaction.
Contact Clayton to learn more at https://minimap.ai
Click here and become an Insider and get a special shout-out on a future episode!
Please leave a review on Apple Podcasts.
Order your copy of "Cracking the Rich Code" today! Use code 'PODCAST' and get 20% off at checkout.
Join The Rich Code Club and take your business and life to the next level! Click here.
Are you a podcast host looking for a great guest or a guest looking for a great podcast? Join PodMatch! Click here.
Host a live stream, record an episode, deliver a webinar, and stream it all to multiple social media platforms! Try StreamYard today for free! Click here.
Record and edit your podcast episodes with the easiest-to-use drag-and-drop tools available! Try Alitu today! Click here.
Join Innovation Women today! Click here.
As an affiliate, I may earn a commission at no extra cost to you.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Man.
I really meant to get thisepisode out a while back, but
ran into some trouble withpost-production.
Deep breath.
It's okay, though.
We are back, and that's what'simportant.
Y'all know, a big passion ofmine is media literacy and, more
specifically, news literacy.
I truly believe that we andwhen I say we, I mean the
(00:20):
citizens of the United States inorder for us to thrive and to
move in the right direction as acountry, we need to have a
society of people that are aswell-informed as possible.
Notice, I did not just sayinformed, I said well-informed.
We need to have the capacity tozoom out from our narrow
(00:42):
understanding of the news thatwe consume, which typically
consists of news that confirmsour own bias and tells us what
we want to hear, makes us feelgood about the team we're on.
We need to be able to see thebroader news landscape to get a
better understanding of theentire media ecosystem.
My guest today has created aninnovative way to do just that,
(01:03):
but from a visual perspective.
Clayton Smith is the founder ofMinimap, which is a cartography
platform that uses artificialintelligence to visually map out
the content landscape.
It creates maps of largecontent spaces, specifically the
news, because when it comes toconsuming content.
Most of us consumers are in thedark, especially where it
(01:26):
concerns how much content isbeing produced by whom and for
what end.
To be able to see the contentand the producers of content in
a visual that helps us connectthe dots in a more comprehensive
way is incredibly valuable,because it helps us better
understand what's going on inthe world around us, or at least
what we're being.
What's going on in the worldaround us, or at least what
we're being told is going on inthe world around us.
(01:48):
The mini-map gives us a way tozoom out and see the bigger
picture Context about ourcontent matters, so that we can
think critically about theinformation we are consuming.
To hear more about thisfascinating topic of content
cartography, keep listening.
Speaker 2 (02:22):
Welcome to the Communication 24-7 podcast where
we communicate about how Jenit's so good to see you again.
Yeah, likewise.
Speaker 1 (02:41):
I'm excited about this conversation because we
have not been able to have aconversation in some time.
Speaker 2 (02:57):
When was the last time we actually spoke, released
our alpha of Minimap and westarted just talking about what
a content cartography platformlooks like and just how useful
it would be to visually see thecontent landscape.
Speaker 1 (03:13):
Yeah, that was really exciting to be able to see that
and just get the demonstrationfrom you and just have that
conversation about media andcontent and how we receive it
and how it's produced and all ofthose things.
So that's why I'm really happythat you agreed to be on the
show, because number one, forselfish reasons I want to get an
(03:34):
update on what's going on withyou and what's going on with the
program, but also just to beable to have a conversation with
someone.
I think we are aligned in ourbeliefs and values when it comes
to the media ecosystem and yeah, so I just want to kind of pick
up where we left off.
So for the audience, would youmind just introducing yourself
(03:56):
so they know, because I know whoyou are, but let the listeners
kind of catch up with who youare and what you've been doing.
Speaker 2 (04:04):
Yeah, absolutely so.
My name is Clayton Smith andI'm the founder of Minimap AI.
Minimap AI is a contentcartography platform, and so
what that means simply is wemake maps of content, and so the
value proposition that we'reputting out there is we're
taking the algorithm andbasically unfolding its brain,
(04:26):
projecting it out onto an actuallandscape and showing people
the content landscape as thealgorithm understands it, and
what that does for us being ableto spatially organize content
is allows us to, at a glance,zoom out and see how content is
manifesting itself, where it'sconcentrated, where it isn't,
(04:48):
and so it allows us to viewdeveloping stories or
collections of articles in waysthat we can't really see or
understand through basic metricsor flat lists of results.
And so that's what we're tryingto do at Minimap AI is just
show people the contentlandscape and give everybody a
medium, a common viewpoint, tounderstand this vast information
(05:11):
space that we're dealing within 2024 and beyond.
Speaker 1 (05:15):
Yeah, that's why I find it so fascinating, because,
you know, in my previous lifeas a media analyst, reading
through thousands and thousandsof news articles and podcasts
and just paying attention towhat's going on in the media
ecosystem, it can beoverwhelming the amount of
information.
But when you're really payingattention to it, close attention
(05:39):
to it, day by day, you begin tosee that there are within the
media ecosystem itself like yousaid, we tend to look at it flat
, you know, like what, whattends to be left, leaning what
tends to be right, leaning whattends to be covered the most
often.
And when you're able to have avisual representation of that,
(06:01):
it just causes you to be able toview some of these things in a
slightly different way.
And, yeah, I just, I reallylove that it is a visual
representation so that it getsus thinking about the news that
we are consuming.
Speaker 2 (06:18):
And that's one of the problems that we're trying to
solve is that, like, forinstance, me as a consumer, I
see articles that say, oh,everyone's talking about this,
or there might be a focus in ourfeeds of particular topics, and
the question I have is iseveryone actually talking about
these things?
(06:38):
To what degree of representationdoes this story have versus
other things, and why is thatthe focus of the media's
attention right now versus anyof the other topics that are out
there?
And so for me, as a consumer,being able to actually see how
these things stack up to eachother, the relative sizes of
them, that's what's importantand that's what I need when I
(06:59):
consume content, because therehave been times where I've seen
articles that say like, yeah,everyone's talking about this,
this is the big thing right now.
And I look and I do searchesand there's nothing else out
there.
These articles are singletons,but I had to do research to do
that.
I had to spend my time tryingto dig up the rest of the story,
and it's it's work.
(07:19):
It's a string and corkboardexercise that piece together,
like how these things actuallyfit into the greater content
space, and it's challenging andthat's what we're here to solve.
Speaker 1 (07:30):
Yeah, you bring up a very important point in that the
things that we are told or thethings that we read online are
not necessarily an accuraterepresentation of what's really
going on, especially if you'regetting most of your news
through social mediarepresentation of what's really
going on, especially if you'regetting most of your news
through social media.
You know, I know so many I can'tremember exactly what
percentage of people are gettingtheir news off of TikTok or you
(07:52):
know, any of the other socialmedia platforms, but it's a
pretty high percentage of peoplethat are getting their news
from these platforms.
And so when you're constantlybeing fed that this particular
topic is not being covered orthis particular topic is being
covered all the time, that maynot necessarily be the case.
It really just depends on whois saying this and how often is
(08:17):
that particular thing beingforced into your feet over and
over again.
So you could be under themisperception that one topic is
being covered a lot when it'snot, and then vice versa, like
you were just saying, and Idon't think a lot of people give
that enough considerationbefore sharing things and before
(08:38):
having a comment about certaintopics that happen in the news.
We get this emotional reactionright about certain topics that
happen in the news we get thisemotional reaction right.
Speaker 2 (08:46):
Yeah, 100%, and even in that just kind of rewinding
to something you said earlier onthere, like the author, the
publisher, who's actually saying, this matters just as much as
what they're saying, and that isthe context that we need to
communicate to people when theyconsume this information that we
want to make easily and readilyaccessible.
But the problem still is, isthat, okay, we found this
(09:10):
article from somebody.
Let's understand who thatsomebody is.
And now it's another exerciseof okay, here's a big list of
their content.
Let's go through and try topiece together and infer the
type of actor or person orpublisher that this entity is,
and it's just still a lot ofwork.
And so, for the end user, itcomes down to either time or
(09:31):
trust.
Are you going to spend the timeresearching these people, these
publishers, or are you justgoing to take it at face value?
Yeah, this checks out with therest of the stuff I heard in my
feed, so I'm just going to sendit full blast.
And so what we want to be ableto do is say like, not only like
okay, here are all the stories,here are all the articles that
(09:55):
are talking about this one story.
But we can also say like, okay,here are all the articles from
this one publisher and just, ata high level, be able to see the
footprint of that publisherwithout even reading any of
their articles, just see that,oh, they talk about sports a lot
or they talk about politics alot, and just understand that.
That that brief moment, beforeyou even start reading their
content.
Speaker 1 (10:15):
Yeah, and when one thing you just said about you
know figuring out on your ownthat's a really difficult
process, figuring out on yourown what's showing up on your
feed.
How does this compare to otherinstances of this topic being
talked about on other platformsor in other news organizations
on one particular social mediaplatform?
(10:36):
And that algorithm has figuredout what you like, what you
click on, what you tend to spendthe most time looking at, and
it's going to consistently justtry to feed you similar content
and then you know it's thatconfirmation bias.
(10:59):
It just completely reinforcesit.
So if you are only relying onwhat is showing up in your feed
to to confirm you know whatyou're already thinking about
this topic, you really don'teven know what your own blind
spot is when it concerns thattopic.
So, with with your mini map, ifI were to go out there and look
(11:21):
up particular topics, what itwould give me would it run the
entire gamut of everything forthat particular day or that
particular week.
Take me into how that mighthelp me figure out my own blind
spots so that I'm actuallyperceiving these topics a little
(11:41):
more accurately.
Speaker 2 (11:43):
Yeah, absolutely.
First, I'll start off just bydescribing how we map content
and what that actually lookslike, and then for the actual
user, for you or anyone who'sinterested in using Minimap,
just what that would feel liketo actually use the platform to
check their blind spots.
So for Minimap, we take atop-down approach when it comes
(12:05):
to rendering and materializingthe content landscape, and what
that means is we start with thehigh-level topics.
We start with sports, politics,weather all of these we have
about 50 right now that wecategorize all news content into
, and those topics are thenorganized spatially on the map.
This might be a good time toactually show what that looks
(12:25):
like for any of the viewers whoare watching this podcast.
Speaker 1 (12:30):
Yes, great idea.
So I know the vast majority ofyou listeners like to listen to
this on the audio format, but ifyou do have an opportunity, go
to the YouTube channel andyou'll be able to see an example
of the map that we're going toshow you.
We'll do our best to try toverbally for those of you who
(12:50):
are listening, though, so we canhelp you visualize, as we're
going through it, what it mightlook like, and then, later on,
you can go to the channel andtake a look at it.
All right, so we have the mapup here on the screen and it
looks so fascinating.
It really does capture peaksand valleys of these umbrella
(13:13):
terms, these umbrella topicsthat you were talking about with
entertainment, real estate,technology.
Yeah, so take us through this.
Speaker 2 (13:21):
Yeah, and so this is as I started to say.
We are organizing the landscapefirst by the topics, the major
buckets and categories that anynews article is most likely
going to fall under, and so thefirst concept here that we're
dealing with is that we want tobe able to maintain continuity,
a continuum between any topics,such that if you were to move
(13:44):
between energy and space, spaceand sciences, food and
entertainment, the articles thatyou see as you move from one
place on the map to anothermakes sense, and that there's a
gradual step between each ofthese places, and so that just
makes the content moreinterpretable as you explore the
map.
(14:04):
So that just makes the contentmore interpretable as you
explore the map.
Secondly, by doing this, we'reable to immediately see trends
within the new space.
For instance, right now we cansee that weather and sports are
the two most reported on things,and that makes sense.
We're going to have dailyweather reports, hourly weather
reports, weather reports for allpockets of the nation and
(14:25):
beyond, and so weather is goingto always be a dominant category
.
But then, second to that issports, and so this just shows
us that we're able to understandthe scale of these two topic
spaces.
And for me, just browsing thenews, just being active in the
(14:46):
news space, this makes sense andit matches what my
understanding was of these twobig buckets of content.
And then for the rest of them,we can see just how big
technology news and internetnews is, relative to finance or
business and all of these otherspaces where news is reported on
, and so this is just at a highlevel, how the content landscape
(15:08):
looks and feels, and so we cansee these peaks and valleys and
see where there areconcentrations, like the other
week there or the other monthnow there were some.
The California primaries wererolling in, votes were coming
out, and so we saw someinteresting news in the content
landscape.
We actually made a little videoabout it where, for politics,
we saw these two sharp spikes.
(15:30):
It wasn't even like a mound,like we saw with sports.
There were these two sharpspikes that were reporting out
the results from each of thedistricts, and those actually
came from two publishers thathad a auto-generated report for
each of the precincts, and sothat's the type of pattern that
we're able to immediatelyidentify.
(15:51):
We don't need metrics, we don'tneed chat summaries, we just see
that there are these big spikesthat are sticking out of the
landscape that match somemovement, a bleeding edge almost
, of what's happening in thenews space, and so that's the
content landscape.
And so then, for a user, whatthey're able to do is we're able
(16:15):
to show news stories, as I saidearlier, like the footprint of
news stories.
So, like here, this story isTesla receives their tentative
full self-driving approval inChina, and so what we're able to
do is show all of the articlesthat are talking about this one
story and all the topics thatthey fall into, and so this is
(16:38):
transportation, this is energy,this is technology and AI and
business and government, andthese are all of the categories
is technology and AI andbusiness and government, and
these are all of the categories,all of the angles that
contribute to this story.
And so, like, at a high level,we're able to understand the
scope, the range of this, thisbigger story that's, that's
evolving and taking place rightnow.
Speaker 1 (16:59):
So so let me jump in, you know, for anybody who's
listening, the interesting thingabout this is so you have, you
know, an article about Tesla and, from a news consumer's
perspective, when you thinkabout Tesla, regardless of how
you feel about the organization,you know the business,
(17:20):
regardless of how you feel aboutElon Musk, any of our opinions
the interesting thing here, likeyou just said, that article on
Tesla, when we pull it out andwe look at all of these other
topics that are connected tothat, there are all of these
different angles that thesearticles are covering, still
(17:44):
talking about Tesla, but itcould be from the technology
perspective, it could be fromthe finance perspective or, you
know, maybe governmentlegislation, you know,
perspective.
So that's that's reallyinteresting someone who is
interested, you know, in thisparticular topic, to be able to
(18:05):
just become better informedabout Tesla from all of these
different angles.
I think it could really helpbroaden your perspective of the
things that are said about thisone particular topic.
Speaker 2 (18:19):
I love it.
Speaker 1 (18:22):
You know, it takes a thousand people to build a
building and just one malcontentwith a stick of dynamite to
ruin it all.
Knowing how to reduce conflictand diffuse tension is key to a
happy life and, even though ittends to be difficult for most
of us, there are some simpletricks and tools that we can
learn that will really improveour life today.
(18:43):
Learn that will really improveour life today.
Hesha Abrams, a world-renownedmediator, negotiator and author
with over 30 years as an expertin resolving conflict,
implements innovative approachesand thought-provoking solutions
that obtain favorable outcomesfor even the most complex
matters.
Hesha's popular new bookHolding the Calm shares her
(19:03):
secrets of how to read asituation to resolve tension,
eliminate conflict and restoreharmony.
Get a copy of the book today,available on Amazon.
Click on the link in the shownotes.
Speaker 2 (19:18):
Absolutely, and just like how Ground News focuses on
breaking stories, developingstories into right-leaning,
left-leaning, and they're morefocused on the polarization of
news, we're focused on thehigh-level content or topics and
organization of not just newsbut content in general.
(19:38):
A little background aboutMinimap is news wasn't actually
our first application.
We started with Amazon'scatalog as the first data source
that we were playing with, butunfortunately it was expensive
to access their data.
News was readily accessible.
We felt that was impact.
We're able to.
By just being able to organizecontent this way, it does help
(20:03):
our users understand andnavigate these, these larger
content spaces.
Speaker 1 (20:08):
Yeah, it's so awesome to see how the map just fills
in, you know as you look out.
Speaker 2 (20:17):
It was a lot of fun learning how to write the front
end.
I got good at shaders, if youknow what those are, but I'm
very thankful that ChatGPT ishere.
But another interesting aspectof this and this is what I was
talking about earlier as ademonstration is we were saying
(20:39):
that not only do the articlesabout content matter, but also
the publisher of those articles,and so one of the things that
we can do is, for instance, herethis is the landmark, the logo
for us Us Weekly.
They're a news publisher, theytalk about entertainment, people
, et cetera, and so what you'reseeing here is this is Us Weekly
(21:02):
, placed as the algorithmunderstands them, and so that
means that they're in people,they're over in this area.
We had nothing to do with whereit lands.
That's all driven by how thealgorithm understands them in
the context of the contentlandscape.
But what, then, our users cando is like okay, I want to know
more about this publisher and wecan highlight all of their
(21:26):
articles on the map for whatevertime range that we're looking
at.
So, like here, we're looking ata 30-day window that we're
looking at, and so, like here,we're looking at a 30 day window
and we can see that they dopost or they do create content
that is people focused, that is,fashion and music and
entertainment focused media, andthey tend to stay out of the
legal and political scenes.
They don't tend to report onfinance and business, and so
(21:50):
they are.
We can see at a glance, withouteven looking at a single
article of theirs, that they areconsistent with their branding,
their messaging and the type ofcontent that they say that they
report on.
Speaker 1 (22:02):
Yeah, this I really, I really like that you're
bringing up this example.
Just this morning I saw Galluphad released, you know, talking
about topics that immigration,of course, naturally, you know
immigration remains a topconcern of voters.
But then follow up question assomeone who likes to follow the
(22:24):
news and you know I'm veryinterested in the news landscape
and how things are covered Oneof the conversations that I have
with other news consumers ishow important topics are covered
in different or by differentnews organizations and whether
(22:45):
or not is it covered more oftenby certain news organizations
than others.
There is an argument to be madewhen you talk about bias in the
news, bias by omission.
You know some newsorganizations that might be
right-leaning or left-leaningmight have a tendency to report
on particular topics more oftenthan others.
(23:06):
You know, and you had mentionedground news earlier you know
that's one of the things thatthey like to show, you know, in
their information, what'sconsidered to be a blind spot be
a blind spot on the rightversus a blind spot on the left.
And so if you were to bring upthe topic of immigration, I
would be just because this is atthe top of my mind seeing that
(23:29):
article from Gallup this morning.
I would be interested in seeingfrom this map you know,
visually how that would span outacross the.
You know the different areashere and, like you said, it
doesn't show left leaning orright leaning, but from a
(23:51):
topical perspective, you know,it is interesting to see where
all of the different dots thatyou have here you know falls
under government politics.
Regional news.
Speaker 2 (24:05):
Yeah, absolutely, and so here we just search for
immigration.
This is a search that's donejust by matching, it's not a
keyword search.
This is anything that isimmigration or immigration
similar, and so we can see allthe areas where immigration pops
up and seems to be a prettyubiquitous or universal concept
(24:30):
across all major categories,across all major categories, and
so we can see it, of course, inpolitics and government and
transportation and world news,but it appears to be coming up
in unexpected places as well,like real estate and technology
and entertainment and media.
Even sports has a couple of hitsfor immigration, and so this is
(24:50):
now we're able to see not justthe major or dominant topics
that we'd, of course, expectthis conversation to come up in,
but the unexpected places aswell, such as sports, and so
what I'd like to do here is do arefined keyword search to
actually bring up the articlesthat are specifically talking
(25:10):
about immigration, and then wecan see that this drops down to
77 articles in the last 30 daysthat we've been able to scrape,
and, of course, we're stillseeing a wide coverage, and
we're even seeing hits out innatural disaster, environment
and climate change, and so thatmight be immigration.
(25:31):
These could be stories, orlikely are, that are focused on
displacement because of climatechange issues or natural
disaster, and so that's.
Unfortunately, I lost my search, but that's what we're trying
to show is that there are, ofcourse, the dominant topics and
categories that we expect thisnews to fall into, but there are
(25:55):
blind spots that are going tobe difficult for people to reach
, to see and understand if theywere to only look at these
search results in a flat list oras summarized by a chat GPT
like algorithm.
Speaker 1 (26:11):
Right, yeah, that's so fascinating to be able to
look at this through thisparticular lens.
You know, thinking about USimmigration, and then you know
just some of the news articlesthat are news organizations that
I tend to pay attention to andhow they cover US immigration.
You know, it's interesting tobe able to drill down into this
(26:32):
map, you know, yeah, with thepolitics, and figure out who's
reporting on this from thatparticular aim or crime.
Or you know, like you justshowed climate change, which
news organizations are coveringthis particular topic from that
angle.
You know.
So, for anyone who is reallyinterested in doing a deep dive
(26:55):
and learning more about thisaspect of the news, and not only
what's being covered but who'scovering it and the angle you
know at which they'reapproaching these particular
topics, it's fascinating to beable to do this.
Speaker 2 (27:10):
Yeah, it's been rewarding.
This has been a dream or visionof mine for a long time, and
after my last job, I was in aposition where I needed to have
a crypto firm.
I came in being able to builddata pipelines.
(27:37):
I could do the technical work.
That wasn't a problem.
But crypto was, and still is,volatile.
It changes fast and if you'rein the crypto space, especially
if you're a decision maker, youhave to be able to react and
understand the news around you.
And, as almost a foreigner tothe space, I was not subscribed
(28:01):
to the right people.
I didn't know who to follow,and all crypto news broke on
Twitter first.
It did.
That's where crypto was.
It was on Twitter, it was on Xand without being able to use
X's search, because it's bad,it's really bad.
Earlier today I found somebodyon X.
I was just trying to figure outwho they are.
(28:23):
I looked at their account andit is difficult to even look at
the list of tweets that a singleperson put out there, because
it's conflated with ads.
And here are some other peopleand here are all their retweets.
It's like no, just show me whatthey've actually written.
(28:44):
And so just being unable to getthe information I needed to
understand the crypto landscape.
I felt like I was a liabilityand this had been something I
had been thinking about longbefore I even started that
company, and after that, I feltit was time, and so I took the
leap and started this andstarted building out Minimap and
just to take it a step furtherfor prospective users and anyone
(29:07):
that's interested, one of thethings that I wanted this
platform to be is an actualmini-map for users, and what
that means for us is that I wantto continue browsing Reddit or
Twitter or any of my news feeds.
As you said earlier in thepodcast, a lot of people get
their news information fromsocial media, and I don't think
(29:31):
that's going to change.
I don't think that readers aregoing to start moving over or
new content.
Consumers are going to startmoving over to the traditional
news websites New York Reddit,it's YouTube Shorts, where
people hear these news eventsand get these snippets of
information, and so what webuilt out is actually a browser
(30:07):
extension that tags along withthe user, and so if we do go to
AP News, there's a ton ofcontent here, and so right now I
have the browser extensionplugged in and it's saying to me
that there are 117 articles ontheir front page.
That is overwhelming.
(30:29):
That's a lot of information.
Yeah, I'm not going to.
I can't handle that.
That's too much for me.
And so what Minimap does forthe user is we have this browser
extension, drive by extension,and we can actually show what it
looks, what a user's newsfeedlooks like, and so on one window
(30:52):
I have AP News.
Doing my browsing as usual, andon the map I have all of those
articles highlighted.
I can see, oh sorry, it lookslike there's a bug.
These things are supposed tohide away, but on the map I can
see each of those articleshighlighted and I can see where
there are concentrations, if any, and where there's a sparsity,
(31:14):
if there's no coverage.
And it looks like on AP News'website.
Wow, here we go.
There is a gap in coverage forenergy and transportation.
I know what these mounds arebecause I spend every day with
the map, but these areas it's ahuge blank spot, exactly.
And so they're not talkingabout Boeing in the news, even
(31:35):
though I know Boeing's Starlineris starting to ramp up and they
have all the whistleblowercoverage.
That's right.
They don't have anyconversations out in the energy
sector.
I know that there's a lot ofnew battery tech happening right
now.
People are trying to find likelonger term, you know, zero
waste batteries, and so this istelling me just at a glimpse.
(31:59):
I'm not even reading a singlearticle.
What?
Oh, even here's another one.
Natural disaster is missing onthe map too, which might be a
good thing, you know, that'sactually probably a great thing
if there's no natural disastersto report.
Speaker 1 (32:11):
Right.
Maybe there's just notsomething going on right now,
but I'm so happy that you pulledthis up and you're using this
as an example.
You know, looking at AssociatedPress News, the website, and
then using your extension, andit is giving you an indication
of all out of all of thosearticles that are on the front
page right now, you know there'simportant news that's happening
(32:34):
that that we're aware of.
Just from reading other newsorganizations, you know
information that they're they'republishing.
But then here you can see thatthere there is a big, big area
looks like a desert coveringthat information, and we're not
saying it's good or that it'sbad, but it is notable that it's
(32:56):
missing off of the front page,and so does this extension.
It only scours that particularpage that you're on at that time
, right.
Speaker 2 (33:06):
Yeah, so we have a small list of websites that the
extension is allowed andapproved to work on, and so
that's us telling the Chromestore and the browsers like only
these sites.
And then when we visit, when theChrome extension sees or the
browser extension sees, oh okay,we're on AP news or we're on
(33:28):
Reddit.
If you try it on Reddit, it'sonly oldreddit, not with their
new layout, but AP News, cnnReddit then it will activate, it
will look at all of the contentthat's on the page in front of
the user and then it will sendthat content over to the map to
be understood by the algorithmand then plotted and mapped out
(33:50):
for the user and mapped out forthe user.
And then in that we're able tostill highlight and navigate the
content landscape as we wouldif this was just the bare or
native content landscape or ifwe were just on Associated Press
.
Because again, here's thesmaller, refined vertical feed
(34:12):
of news and I can from there goright back to the page on
Associated Press.
We're not looking to hostcontent ourselves.
We merely just want to be a,not a dictionary, but almost
like yellow pages, we want to beable to say like here, here's
where all the content is thatyou're looking for.
(34:32):
Here's where it lives, and letus get you there and organize
that content spatially for you.
Speaker 1 (34:37):
Yeah, yeah, and I really like the idea of who's
talking about this today.
Speaker 2 (34:43):
Exactly yeah.
Speaker 1 (34:44):
Yeah, that's so fascinating.
So you really didn't have untilyou worked for that crypto
company and you were findingthat there was information that
you needed to be able to accessand you were having a difficult
time being able to access that.
That gave you the idea tocreate this, this mini map.
So have you been doing this onyour own?
(35:06):
Do you have, do you have,people who are helping you?
Like how this seems like it'san incredible amount of work.
To be honest with you, I don'tknow anything.
I don't know jack about coding,okay, but I can imagine this
looks like hella work.
Speaker 2 (35:21):
So I've had help.
I left the company, the cryptocompany DeFi Pulse, back in
December of 2022.
And I've been working onMinimap ever since.
Literally the next day after Ileft, I started working on
(35:42):
Minimap, and at first it wasjust me for several months and
then an old coworker of mineactually a former intern joined
me in, I think, july of 2023.
And so I'd been working on thisfor six months, just going
(36:03):
through and getting the.
Actually, if you give me aquick second, I have a fun
picture book that just shows theevolution of Minimap.
Speaker 1 (36:11):
That's just I love it .
I love it that you have apicture book.
Speaker 2 (36:16):
You know it's a, a it's a visual thing.
We're here to visuallycommunicate things, and so I
feel like it is fun that this isa visual representation of of
the news landscape.
Speaker 1 (36:27):
I think that's why I get so excited about it.
I'm a bit of a nerd that way,but fellow nerd nerds, if you
want to check it out, I highlyrecommend it.
Oh, here you go.
Okay, I see you, yeah.
Speaker 2 (36:40):
So this is the progress, our timeline In
December of 2022,.
I founded Minimap, Iincorporated, created the LLC,
all of that and more, and theoriginal idea was I wanted to be
able to map out search resultsfrom any search bar anywhere
where there's a search bar onthe internet and our algorithm
(37:00):
can understand that content.
That's that's the space that wewant to be, especially for,
like amazon's catalog, which isoverrun with drop sellers, so
there are so many duplicateproducts out there, and so for
us, what that would look like onthe landscape is like that
example earlier, where those twopublishers were making the
(37:21):
identical reports for everysingle precinct in California
and we saw those spikes.
That's what Amazon's catalogwould look like on the map,
where all of those drop sellerswould have hundreds of identical
products.
We don't care how many itemsare in the search results.
What we care about is how manyunique items there are, and so,
deduplicating those spatiallyand just stacking all the
(37:43):
similar things up, we wanted tobe able to show the full space
of search results for users.
And so here, like I'm looking atthis one sneaker as an example
and I just want to be able tohighlight where this is, and so
this was the initial idea andthis image I made by hand.
This isn't, and so the firstit's not to scale.
(38:04):
Is that what you're saying?
Yeah, it's not to scale, madeby hand, no algorithms involved.
This is the concept.
And so the first place to startis let's just start with two
things.
And so the first idea was okay,let's see if we can't make a
spectrum between boots andsneakers and have the algorithm
organize things as a gradientbetween those two, and so, on
(38:25):
the right, here we have sneakers.
On the left, here we have boots.
And then the idea was okay,well, can we have somebody, be
it Skechers or Keen Boots,define those polls, those
endpoints, and just that couldbe the business model.
And so, like, okay, brand youcan represent, like this poll
and other brand, that poll, andthen we can organize any and
(38:47):
every piece of content relativeto those two things.
And so that was where westarted.
And then this here is thatfirst implementation, and so
it's a diamond, for reasons thatdoesn't really matter.
What does matter is that as yougo from left to right, it goes
from most boot-like boot to mostsneaker-like sneaker, and at
(39:09):
the very top is the mostboot-like, most sneaker-like
thing.
Speaker 1 (39:13):
That's a nice combination of the two.
Speaker 2 (39:17):
The maximal combination and at the bottom we
had the minimal combination.
Like these aren't boots, thesearen't sneakers, these are cars
and airplanes.
Speaker 1 (39:24):
Yeah, and then it's like what are those doing?
Speaker 2 (39:27):
Yeah, and over here we have a car sneaker, but the
algorithm saw that, hey, this issubstantially sneaker, so we're
going to that belongs overthere and in some capacity it's
right.
And so this is showing us thatwe can organize things onto a
spectrum, a gradient between twounderstood topics, and also
(39:50):
distill a semblance of relevancywhere we, if you're sticking to
this top most border.
These are going to be the mostrelevant boots or sneakers, or
boot sneakers, whatever you wantto make the spectrum out of.
And so this was the firstsuccess that we had in February,
two months after I startedexperimenting with the algorithm
(40:11):
.
Wow, that's amazing.
And so then in March, we turnedthis into a full 3D thing, and
so now it's like, okay, we knowhow the math works for the 2D
scenario, let's make this 3D sowe can define these topics and
then organize several spectrumsat once.
And so what we're looking athere is these diamonds.
(40:35):
Upside down, diamonds representthe topics.
These are the boots, these arethe sneakers, these are anything
that we thought was relevantfor defining the content
landscape at a high level, andthen everything else, these
peaks and valleys and littlespikes.
That's the content, and this iswith Amazon's product catalog,
(40:59):
and then, of course, you can seethe logos, and so these are the
seller logos, and they areplaced based off of the
algorithm's understanding ofeither the seller themselves or
the content that they're selling.
And so behind logo, there's aBarbie thing over there.
There's in the far top leftcorner, there's mighty skins.
(41:19):
They do like skins for yourdrones or your like unibords,
all that kind of stuff.
But this was this was our firstsuccess with 3d mapping.
However, the math needed to beimproved, and so we started
experimenting.
So I started experimenting withdifferent approaches, some
weird spirals and donuts whentrying to organize these high
(41:42):
dimensional point clouds ofinformation figured out.
I figured out how to actuallymake this real, how to
meaningfully organize thecontent around these topics, and
so I can see that there's alandmark behind that valley over
there.
But that's, this is.
(42:03):
This is the first success, andthis is now the algorithm that
we're running with today.
And so then, afterwards, I hadto move on from Amazon.
As I said earlier, they'reexpensive, they're difficult,
they'd only give back 1,000search results and it would take
10 minutes.
They limit us any of theirvendors, through their regular
(42:26):
APIs, to two requests a second,and if they only give 1,000
requests or results in theirsearches, it would take 10
minutes to act for us toactually get a user's real time
search, and so that wasn't goingto be tenable for us.
And so what you're looking athere is the social media
(42:46):
platform, blue sky, and so nowwe've we, I turned to blue sky
to actually create content maps,and so what you're looking at
here on the right or, sorry, onthe left is the major topics by
the counts of documents and thensome relevancy scores.
(43:08):
This was my proof that I couldsearch for crypto content and
not just highlight all of thecontent that's on crypto island,
but all of the content that'soff of crypto island, because
there is content that isfinancial or not financial.
Let's space, let's say spacefocused, and it's like we
(43:28):
figured out how to useblockchain to do a thing on the
ISS, space, space, space, and sothat's a space dominant topic,
but that's crypto over in space,and so you otherwise probably
would have never been able tosee that exactly discovered it,
and so, like now we can.
I can see all of the places thatcrypto content is popping up
(43:51):
throughout the content landscapethat's outside of a crypto
dominant context, and so thiswas a big success.
And then this is the time whereone of our other software
engineers came on the platform.
We started experimenting withhighlighting by date,
highlighting by sentimentanalysis, refining the actual UI
, and so now it's getting alittle bit more sophisticated.
(44:12):
Refining the actual UI, and sonow it's getting a little bit
more sophisticated.
We're bringing on landmarksfrom from companies on Twitter.
So what I did here is I wentthrough and collected 20
companies and took their toptweets and gave them to the
algorithm, and now all of theselogos that you're seeing are
(44:33):
organized by how the algorithmunderstands their posts.
And one of the reallyinteresting findings here is
NVIDIA.
I thought we had a bug.
I thought that NVIDIA is atech-focused company.
They're AI-focused, they arehuge in the tech space and so
unfortunately, I didn't capturethe topic that this little
(44:56):
island represents.
But this is corporate andcorporate news,
corporate-focused stuff,business, also adjacent to
influencers and social media,and I was like why is NVIDIA
over in the corporate space?
Why is NVIDIA?
Why aren't they over in tech orAI?
And it turned out it's becausetheir top 10 tweets actually to
(45:20):
this day on LinkedIn, on Twitter, across all of their socials
nine out of 10 of their postsare about their CEO or their
executive teams or theirleadership and what they're
doing to further NVIDIA as acompany, and so they are talking
about themselves at a corporateor executive level always, and
so, as far as the algorithmunderstands it, they're not a
(45:42):
tech company, they're aenterprise company.
They're a social mediainfluencer.
They're a enterprise company,they're a social media
influencer.
They talk about themselvesversus the tech that they make,
and so this is an example of amisalignment between what the
company actually does and howthey portray themselves out on
the social landscape.
Speaker 1 (46:04):
If you've read my book cracking the rich code, you
know it is chock full offantastic advice from top
thought leaders and supersuccessful entrepreneurs from
around the world.
How would you like to be amember of an exclusive community
that provides that same how-tocontent from those same leaders?
What if you were able to attendmember-only live events and
(46:27):
interact with them?
I'd like to invite you to jointhe Rich Code Club.
It will change the way youthink about yourself, your money
and your life.
It's the only social mediaplatform purely focused on
helping you grow.
Join the Rich Code Club todayfor free by clicking on the have
(46:50):
.
In the current model, how manycompanies would you be able to
look up specific companies tosee the disbursement of topics
that they talk about or tweetabout or post about in all of
their different platforms?
Speaker 2 (47:05):
Yeah.
So right now, as I showed, weare able to highlight different
companies out on the socialmedia scape and whether or not
there's brand misalignment.
That's still pretty difficult.
What we have to do is then readlike, let's go back to Us
Weekly here.
Oh, are we going to get it?
(47:25):
It should be loading.
And so what it really means isthat just understanding the you
know what they're actuallytalking about versus where the
algorithm is placing them.
And so right now, we're placingthem based off the content that
they produce, and so what wecould do to actually highlight
this is there is a descriptionthat's going to load shortly
(47:47):
about what Us Weekly is, andthen there's all the content
that we have, and we can look atthe difference between how the
algorithm understands theirdescription versus how the
algorithm understands thecontent that they produced, and
that's how we could identify amisalignment or incongruency
between the brand and theircontent.
Speaker 1 (48:10):
Right.
So if I'm an organization, Iclaim to focus my content on
sports, but then I look on themap here and I see, wow, there
is an awful lot of informationthey're producing on politics.
That could be an indication ofa misalignment there.
Speaker 2 (48:29):
Yeah, that actually could be a very fun layer that
we could add to the map to showwhere these types of
misalignments are happening.
Yeah, it would be a fun sideproject for us.
Speaker 1 (48:43):
I'm good for that.
I'm good for for side projects.
Hey, can you?
You know it'd be cool.
Speaker 2 (48:50):
Yeah, but it's it's, it's real, it's impactful and
there are a lot of brands outthere, Like, for instance,
Wendy's food company, but theway they and I'll scroll- back
here.
Speaker 1 (49:01):
Yeah, I mean they're.
They're so famous for their,their tweets.
Speaker 2 (49:04):
Exactly, and so you can see out in the distance.
Here there's five guys justtowards the top of the screen,
but Wendy's isn't there, butfive guys in Wendy's burger
chains fast food.
Speaker 1 (49:15):
So why?
Speaker 2 (49:15):
aren't they next to each other?
And that's because Wendy'sportrays themselves as more of a
meme-y social media account.
They don't really push theirburgers as much as they push
just the idea of Wendy's andself-promotion, and that kind of
stuff, and so it evenhighlights the difference in
media strategies that thesecompanies are taking.
And so for users, again, thisgets back to being able to
(49:38):
understand these large entitiesthat have vast amounts of
content at a glance.
And there's no way that you cando that with any number of
metrics or chat descriptions.
It just doesn't work.
I like to give the example oflike I could draw a squiggly
circle and I could try todescribe to you that circle in
(50:00):
as many metrics as I want, andthere is no way that you could
recreate that circle.
It just it doesn't work that way, and so how do we expect to
understand the dynamics of thecontent landscape when all we
have are metrics, chatdescriptions or vertical scrolls
?
And so that's why I think wehave like something's going to
(50:21):
give, it's going to.
Things are going to change atsome point, whether it's by our
hand or somebody else, but justto continue on.
You know, we kept evolving, wekept iterating, like here's just
showing how similar storiesgroup together and occupy like
very close, like proximity onthe map, and then in September
(50:43):
and October that's when wereleased our beta the landscape
looks very different.
Now we had to change because wehad to make it easier to like
see things.
When you move your applicationfrom the 2D space to a 3D space,
everything goes out the windowin terms of UI design and color
dynamics and all of that, and soit's a tough challenge.
(51:04):
I was not prepared for thelevel of color science needed to
think the light and yeah, andso we just kept iterating.
We filed patents for themapping algorithm that we have,
and that brings us to today withthe content landscape as you
see it.
Speaker 1 (51:22):
Yeah, that's so fascinating.
So where do you see this goingnext?
Speaker 2 (51:28):
We are trying to raise on Kickstarter right now
to get ourselves to market andin front of users.
What we want to be is a mediumbetween users who need to
understand the content landscape.
We're at a point where that'sjust a must now, because brands
(51:48):
have big tools for this kind ofstuff.
There's Muckrak, there's Cision, there's all of these social
listening tools that can breakdown the content landscape into
any number of metrics and nuancy, descriptions and chat.
But it's that's that's stilllike a problem that we're trying
to solve.
Like again, I don't think thatwe can understand the content
(52:11):
landscape through just metricsand descriptions, but the
problem is these socialcompanies.
They have the tools needed tounderstand the landscape as it's
relevant to them, but theirunderstanding of the content
landscape is not the same ascontent consumers.
Content consumers are in thedark, and so we first want to be
(52:35):
a tool for the content consumerjust to get their bearings on
what's going on in the worldaround them, to be able to take
their random piece of socialcontent that they discovered on
reddit or whatever platform and,okay, show me everything else
that's like this, yeah.
And then, second, we want to bea medium that allows brands to
(52:57):
represent themselves brands,communities, anyone, those
landmarks.
Now there's an opportunity forbrands to represent themselves
in.
Now there's an opportunity forbrands to represent themselves
in contextually, in the spacesthat are relevant to the content
that they produce, to thespaces that that consumers are
looking for them.
And so it's either you know yousee walmart and then you see
(53:18):
walmart on the map, you're gonnabe like okay.
Or starbucks food you see theirlogo on the map, be like okay, I
, I kind of know what this isbecause I know what starbucks is
and so I have a sense of whatI'm going to find next.
Or vice versa the user is infood, the user sees a logo that
they've never seen before, butthat logo is in food and it's
next to starbucks and it's likeokay, this is probably going to
(53:40):
be a coffee shop and and so,just like Google Maps allows
users to look at a city at ahigh level and see logos and
landmarks and the little spoonand fork icons or museum icons
and understand like, oh, okay,food districts, entertainment
district, all of these things.
We want to be able to giveusers the ability to
(54:03):
contextualize content, landscapein that way and give brands the
ability to express themselvesand show themselves to users in
the same view and just have aunified view.
Companies and massive brandsview themselves, and how users
understand the world around themis vast.
There's nothing out therethat's attempting to bridge that
(54:34):
gap or even cater to the basicneeds of users with these
problems that we talked abouttoday.
Speaker 1 (54:40):
Yeah, that's fantastic, just from a brand
awareness, brand managementperspective, any organization,
it doesn't matter the industrythat you're in, I mean, you are
a brand, so being able to have amechanism in place to try to
figure out, okay, your consumersdo they have a perception of
(55:02):
you that matches what you hopeit is?
You know one thing that I say alot you know as a communication
person is the message you sendmay not necessarily be the
message received, and I thinkthis is a perfect example of
that.
If you are a brand, whateveryour, you know whatever
organization, wherever you are,whatever it is that you're
selling, or you know the contentthat you're producing, it's
(55:25):
really important to be able tokeep that in mind.
Is the message we're sending,the message that's being
received?
And if there's a misalignmentthere, we need to be aware of
that so that we can figure outwhat changes we might need to
make.
You know in our messaging andhow we're getting it out, where
we're getting it out.
You know, yeah, that'sfantastic.
Speaker 2 (55:45):
Yeah, absolutely.
I firmly believe that there is.
There's a lot to be gained fromjust a shared understanding of
the content landscape and thenfrom there just building out.
Speaker 1 (55:58):
Yeah, absolutely so, if anyone is interested in
playing around with the map orthey want to get in contact with
you or they want to help insome way, because I know that
you have a Kickstarter campaigngoing on right now, so give us,
give us, all the informationthat we need for that.
Speaker 2 (56:17):
Absolutely.
We're looking for feedback.
We're trying to get users onthe platform to understand how
they engage and what their firstimpressions and experience are
with the map.
They're the first contact andso getting feedback.
Join our Discord, give us thatfeedback and, if you like what
you see and you get the visionand you understand where we're
(56:39):
trying to go with this, becausewe want to be multi-domain,
multi-modal in the sense thatit's not just news, but news,
social media, e-commerce,anything that the algorithm is
capable of understanding.
That's what we want to be ableto put onto a map.
My goal.
It's nowhere within reach rightnow, but I'd love to have a map
like this of music.
(56:59):
I'd like to take a snippet of asong that just I really jam to
like.
Sometimes they're just thesesmall 10 seconds.
I'm like that.
Find everything with that vibeand give it to me.
Speaker 1 (57:09):
That would be fun, that would be great yeah.
Speaker 2 (57:11):
Yeah, exactly, and that's where we want to go with,
again, anything that thealgorithm is capable of
understanding.
And so right now we have ourproof of concept, our minimal
viable product out there thatdemonstrates the concepts at
play here, and what we need isto get the funding to bring this
to market, to get stablerevenue and get to the point
(57:34):
where we can expand thealgorithm to our algorithms to
encompass social media,e-commerce and even images and
video things that are out ofdomain of just text, but
multi-domain, multimodal and,yeah, maybe even at a point
where we can do music and audio.
But that's far in the future,and so right now we're just
(57:56):
looking to get funded to getthis crowdsource campaign off
the ground.
Speaker 1 (58:02):
Okay, yeah, so how long is that campaign going to
last?
Speaker 2 (58:05):
So we have 16 more days of our campaign.
All right, we need 23rd.
Speaker 1 (58:09):
So got to get cracking on that.
Speaker 2 (58:12):
Oh, I know, yep, I'm watching the countdown.
Speaker 1 (58:16):
Yeah Well, how can we get in touch with you, any
listeners who might want tocontact you directly?
You mentioned Discord.
Are you on any other socialmedia platforms?
Speaker 2 (58:25):
You mentioned Discord .
Are you on any other socialmedia platforms?
I'm on Discord and LinkedIn.
I don't go on Twitter too much,it's just I have not enjoyed
their experience, for obviousreasons, and so yeah, and if
anyone wants to contact medirectly, it's just Clayton at
Minimapai and I always read myemails, and so you know I'd love
(58:46):
to.
Speaker 1 (58:46):
I'd love to have a conversation.
Yeah, that's fantastic, andI'll make sure to have the link
to the website in the show notesas well.
Well, my friend, it has beenexciting seeing this journey
since the first time you and Ihad a conversation talking about
the media landscape and youintroduced what it was that you
were working on, and just to seesome of the changes even
between then and now, I'm superexcited to see where this is
(59:10):
going.
So we're definitely going tohave to have another
conversation in the future sothat we can think back to this
moment in time and do the wholeremember when kind of thing,
kind of you know conversation,because I think it's fantastic.
So many applications and justfrom a consumer perspective, a
(59:30):
news consumer perspective Ithink it is filling an important
need that exists out there, andthe fact that you're coming at
it from a visual perspective, Ithink it provides something that
all of the other metrics cannotprovide at the moment, and so I
(59:51):
think more information isalways better.
So I just appreciate whatyou're doing and I appreciate
that you're out there makingsure that we can get as informed
as possible, because it's socritical.
We need to have a society, weneed to have a community that is
as well informed as possible.
So yeah, yeah.
Speaker 2 (01:00:13):
Thank you for doing this.
No, thank you.
I appreciate that it'smotivating.
Speaker 1 (01:00:18):
Oh yeah, I'm a fan, so for the listeners more to
come, make sure that you jointhis Kickstarter campaign.
Let's make sure that the thiskeeps growing and the usability,
I think when you get out thereand you start playing around in
there and, like Clayton said,they're looking for people to go
to the website and and giveyour feedback.
(01:00:39):
So that's really important.
Be a part of it now.
How exciting is that you can?
Speaker 2 (01:00:43):
be a part of it and Join us day one.
It's exciting, it really is.
Thank you, jen, I appreciate it.
Speaker 1 (01:00:51):
Absolutely All right.
You have a great rest of yourday and, listeners, I'll have a
great rest of your day as well,and we'll see you next time.
Take care now.
Thanks for listening.
If you enjoyed this episode andyou'd like to help support the
(01:01:11):
podcast, please share it withothers, post about it on social
media or leave a rating and areview.
Man.
I really meant to get thisepisode out a while back, but
ran into some trouble withpost-production.
Deep breath.
It's okay, though.
We are back, and that's what'simportant.
Y'all know, a big passion ofmine is media literacy and, more
specifically, news literacy.
I truly believe that we andwhen I say we, I mean the
(00:20):
citizens of the United States inorder for us to thrive and to
move in the right direction as acountry, we need to have a
society of people that are aswell-informed as possible.
Notice, I did not just sayinformed, I said well-informed.
We need to have the capacity tozoom out from our narrow
(00:42):
understanding of the news thatwe consume, which typically
consists of news that confirmsour own bias and tells us what
we want to hear, makes us feelgood about the team we're on.
We need to be able to see thebroader news landscape to get a
better understanding of theentire media ecosystem.
My guest today has created aninnovative way to do just that,
(01:03):
but from a visual perspective.
Clayton Smith is the founder ofMinimap, which is a cartography
platform that uses artificialintelligence to visually map out
the content landscape.
It creates maps of largecontent spaces, specifically the
news, because when it comes toconsuming content.
Most of us consumers are in thedark, especially where it
(01:26):
concerns how much content isbeing produced by whom and for
what end.
To be able to see the contentand the producers of content in
a visual that helps us connectthe dots in a more comprehensive
way is incredibly valuable,because it helps us better
understand what's going on inthe world around us, or at least
what we're being.
What's going on in the worldaround us, or at least what
we're being told is going on inthe world around us.
(01:48):
The mini-map gives us a way tozoom out and see the bigger
picture Context about ourcontent matters, so that we can
think critically about theinformation we are consuming.
To hear more about thisfascinating topic of content
cartography, keep listening.
Speaker 2 (02:22):
Welcome to the Communication 24-7 podcast where
we communicate about how Jenit's so good to see you again.
Yeah, likewise.
Speaker 1 (02:41):
I'm excited about this conversation because we
have not been able to have aconversation in some time.
Speaker 2 (02:57):
When was the last time we actually spoke, released
our alpha of Minimap and westarted just talking about what
a content cartography platformlooks like and just how useful
it would be to visually see thecontent landscape.
Speaker 1 (03:13):
Yeah, that was really exciting to be able to see that
and just get the demonstrationfrom you and just have that
conversation about media andcontent and how we receive it
and how it's produced and all ofthose things.
So that's why I'm really happythat you agreed to be on the
show, because number one, forselfish reasons I want to get an
(03:34):
update on what's going on withyou and what's going on with the
program, but also just to beable to have a conversation with
someone.
I think we are aligned in ourbeliefs and values when it comes
to the media ecosystem and yeah, so I just want to kind of pick
up where we left off.
So for the audience, would youmind just introducing yourself
(03:56):
so they know, because I know whoyou are, but let the listeners
kind of catch up with who youare and what you've been doing.
Speaker 2 (04:04):
Yeah, absolutely so.
My name is Clayton Smith andI'm the founder of Minimap AI.
Minimap AI is a contentcartography platform, and so
what that means simply is wemake maps of content, and so the
value proposition that we'reputting out there is we're
taking the algorithm andbasically unfolding its brain,
(04:26):
projecting it out onto an actuallandscape and showing people
the content landscape as thealgorithm understands it, and
what that does for us being ableto spatially organize content
is allows us to, at a glance,zoom out and see how content is
manifesting itself, where it'sconcentrated, where it isn't,
(04:48):
and so it allows us to viewdeveloping stories or
collections of articles in waysthat we can't really see or
understand through basic metricsor flat lists of results.
And so that's what we're tryingto do at Minimap AI is just
show people the contentlandscape and give everybody a
medium, a common viewpoint, tounderstand this vast information
(05:11):
space that we're dealing within 2024 and beyond.
Speaker 1 (05:15):
Yeah, that's why I find it so fascinating, because,
you know, in my previous lifeas a media analyst, reading
through thousands and thousandsof news articles and podcasts
and just paying attention towhat's going on in the media
ecosystem, it can beoverwhelming the amount of
information.
But when you're really payingattention to it, close attention
(05:39):
to it, day by day, you begin tosee that there are within the
media ecosystem itself like yousaid, we tend to look at it flat
, you know, like what, whattends to be left, leaning what
tends to be right, leaning whattends to be covered the most
often.
And when you're able to have avisual representation of that,
(06:01):
it just causes you to be able toview some of these things in a
slightly different way.
And, yeah, I just, I reallylove that it is a visual
representation so that it getsus thinking about the news that
we are consuming.
Speaker 2 (06:18):
And that's one of the problems that we're trying to
solve is that, like, forinstance, me as a consumer, I
see articles that say, oh,everyone's talking about this,
or there might be a focus in ourfeeds of particular topics, and
the question I have is iseveryone actually talking about
these things?
(06:38):
To what degree of representationdoes this story have versus
other things, and why is thatthe focus of the media's
attention right now versus anyof the other topics that are out
there?
And so for me, as a consumer,being able to actually see how
these things stack up to eachother, the relative sizes of
them, that's what's importantand that's what I need when I
(06:59):
consume content, because therehave been times where I've seen
articles that say like, yeah,everyone's talking about this,
this is the big thing right now.
And I look and I do searchesand there's nothing else out
there.
These articles are singletons,but I had to do research to do
that.
I had to spend my time tryingto dig up the rest of the story,
and it's it's work.
(07:19):
It's a string and corkboardexercise that piece together,
like how these things actuallyfit into the greater content
space, and it's challenging andthat's what we're here to solve.
Speaker 1 (07:30):
Yeah, you bring up a very important point in that the
things that we are told or thethings that we read online are
not necessarily an accuraterepresentation of what's really
going on, especially if you'regetting most of your news
through social mediarepresentation of what's really
going on, especially if you'regetting most of your news
through social media.
You know, I know so many I can'tremember exactly what
percentage of people are gettingtheir news off of TikTok or you
(07:52):
know, any of the other socialmedia platforms, but it's a
pretty high percentage of peoplethat are getting their news
from these platforms.
And so when you're constantlybeing fed that this particular
topic is not being covered orthis particular topic is being
covered all the time, that maynot necessarily be the case.
It really just depends on whois saying this and how often is
(08:17):
that particular thing beingforced into your feet over and
over again.
So you could be under themisperception that one topic is
being covered a lot when it'snot, and then vice versa, like
you were just saying, and Idon't think a lot of people give
that enough considerationbefore sharing things and before
(08:38):
having a comment about certaintopics that happen in the news.
We get this emotional reactionright about certain topics that
happen in the news we get thisemotional reaction right.
Speaker 2 (08:46):
Yeah, 100%, and even in that just kind of rewinding
to something you said earlier onthere, like the author, the
publisher, who's actually saying, this matters just as much as
what they're saying, and that isthe context that we need to
communicate to people when theyconsume this information that we
want to make easily and readilyaccessible.
But the problem still is, isthat, okay, we found this
(09:10):
article from somebody.
Let's understand who thatsomebody is.
And now it's another exerciseof okay, here's a big list of
their content.
Let's go through and try topiece together and infer the
type of actor or person orpublisher that this entity is,
and it's just still a lot ofwork.
And so, for the end user, itcomes down to either time or
(09:31):
trust.
Are you going to spend the timeresearching these people, these
publishers, or are you justgoing to take it at face value?
Yeah, this checks out with therest of the stuff I heard in my
feed, so I'm just going to sendit full blast.
And so what we want to be ableto do is say like, not only like
okay, here are all the stories,here are all the articles that
(09:55):
are talking about this one story.
But we can also say like, okay,here are all the articles from
this one publisher and just, ata high level, be able to see the
footprint of that publisherwithout even reading any of
their articles, just see that,oh, they talk about sports a lot
or they talk about politics alot, and just understand that.
That that brief moment, beforeyou even start reading their
content.
Speaker 1 (10:15):
Yeah, and when one thing you just said about you
know figuring out on your ownthat's a really difficult
process, figuring out on yourown what's showing up on your
feed.
How does this compare to otherinstances of this topic being
talked about on other platformsor in other news organizations
on one particular social mediaplatform?
(10:36):
And that algorithm has figuredout what you like, what you
click on, what you tend to spendthe most time looking at, and
it's going to consistently justtry to feed you similar content
and then you know it's thatconfirmation bias.
(10:59):
It just completely reinforcesit.
So if you are only relying onwhat is showing up in your feed
to to confirm you know whatyou're already thinking about
this topic, you really don'teven know what your own blind
spot is when it concerns thattopic.
So, with with your mini map, ifI were to go out there and look
(11:21):
up particular topics, what itwould give me would it run the
entire gamut of everything forthat particular day or that
particular week.
Take me into how that mighthelp me figure out my own blind
spots so that I'm actuallyperceiving these topics a little
(11:41):
more accurately.
Speaker 2 (11:43):
Yeah, absolutely.
First, I'll start off just bydescribing how we map content
and what that actually lookslike, and then for the actual
user, for you or anyone who'sinterested in using Minimap,
just what that would feel liketo actually use the platform to
check their blind spots.
So for Minimap, we take atop-down approach when it comes
(12:05):
to rendering and materializingthe content landscape, and what
that means is we start with thehigh-level topics.
We start with sports, politics,weather all of these we have
about 50 right now that wecategorize all news content into
, and those topics are thenorganized spatially on the map.
This might be a good time toactually show what that looks
(12:25):
like for any of the viewers whoare watching this podcast.
Speaker 1 (12:30):
Yes, great idea.
So I know the vast majority ofyou listeners like to listen to
this on the audio format, but ifyou do have an opportunity, go
to the YouTube channel andyou'll be able to see an example
of the map that we're going toshow you.
We'll do our best to try toverbally for those of you who
(12:50):
are listening, though, so we canhelp you visualize, as we're
going through it, what it mightlook like, and then, later on,
you can go to the channel andtake a look at it.
All right, so we have the mapup here on the screen and it
looks so fascinating.
It really does capture peaksand valleys of these umbrella
(13:13):
terms, these umbrella topicsthat you were talking about with
entertainment, real estate,technology.
Yeah, so take us through this.
Speaker 2 (13:21):
Yeah, and so this is as I started to say.
We are organizing the landscapefirst by the topics, the major
buckets and categories that anynews article is most likely
going to fall under, and so thefirst concept here that we're
dealing with is that we want tobe able to maintain continuity,
a continuum between any topics,such that if you were to move
(13:44):
between energy and space, spaceand sciences, food and
entertainment, the articles thatyou see as you move from one
place on the map to anothermakes sense, and that there's a
gradual step between each ofthese places, and so that just
makes the content moreinterpretable as you explore the
map.
(14:04):
So that just makes the contentmore interpretable as you
explore the map.
Secondly, by doing this, we'reable to immediately see trends
within the new space.
For instance, right now we cansee that weather and sports are
the two most reported on things,and that makes sense.
We're going to have dailyweather reports, hourly weather
reports, weather reports for allpockets of the nation and
(14:25):
beyond, and so weather is goingto always be a dominant category
.
But then, second to that issports, and so this just shows
us that we're able to understandthe scale of these two topic
spaces.
And for me, just browsing thenews, just being active in the
(14:46):
news space, this makes sense andit matches what my
understanding was of these twobig buckets of content.
And then for the rest of them,we can see just how big
technology news and internetnews is, relative to finance or
business and all of these otherspaces where news is reported on
, and so this is just at a highlevel, how the content landscape
(15:08):
looks and feels, and so we cansee these peaks and valleys and
see where there areconcentrations, like the other
week there or the other monthnow there were some.
The California primaries wererolling in, votes were coming
out, and so we saw someinteresting news in the content
landscape.
We actually made a little videoabout it where, for politics,
we saw these two sharp spikes.
(15:30):
It wasn't even like a mound,like we saw with sports.
There were these two sharpspikes that were reporting out
the results from each of thedistricts, and those actually
came from two publishers thathad a auto-generated report for
each of the precincts, and sothat's the type of pattern that
we're able to immediatelyidentify.
(15:51):
We don't need metrics, we don'tneed chat summaries, we just see
that there are these big spikesthat are sticking out of the
landscape that match somemovement, a bleeding edge almost
, of what's happening in thenews space, and so that's the
content landscape.
And so then, for a user, whatthey're able to do is we're able
(16:15):
to show news stories, as I saidearlier, like the footprint of
news stories.
So, like here, this story isTesla receives their tentative
full self-driving approval inChina, and so what we're able to
do is show all of the articlesthat are talking about this one
story and all the topics thatthey fall into, and so this is
(16:38):
transportation, this is energy,this is technology and AI and
business and government, andthese are all of the categories
is technology and AI andbusiness and government, and
these are all of the categories,all of the angles that
contribute to this story.
And so, like, at a high level,we're able to understand the
scope, the range of this, thisbigger story that's, that's
evolving and taking place rightnow.
Speaker 1 (16:59):
So so let me jump in, you know, for anybody who's
listening, the interesting thingabout this is so you have, you
know, an article about Tesla and, from a news consumer's
perspective, when you thinkabout Tesla, regardless of how
you feel about the organization,you know the business,
(17:20):
regardless of how you feel aboutElon Musk, any of our opinions
the interesting thing here, likeyou just said, that article on
Tesla, when we pull it out andwe look at all of these other
topics that are connected tothat, there are all of these
different angles that thesearticles are covering, still
(17:44):
talking about Tesla, but itcould be from the technology
perspective, it could be fromthe finance perspective or, you
know, maybe governmentlegislation, you know,
perspective.
So that's that's reallyinteresting someone who is
interested, you know, in thisparticular topic, to be able to
(18:05):
just become better informedabout Tesla from all of these
different angles.
I think it could really helpbroaden your perspective of the
things that are said about thisone particular topic.
Speaker 2 (18:19):
I love it.
Speaker 1 (18:22):
You know, it takes a thousand people to build a
building and just one malcontentwith a stick of dynamite to
ruin it all.
Knowing how to reduce conflictand diffuse tension is key to a
happy life and, even though ittends to be difficult for most
of us, there are some simpletricks and tools that we can
learn that will really improveour life today.
(18:43):
Learn that will really improveour life today.
Hesha Abrams, a world-renownedmediator, negotiator and author
with over 30 years as an expertin resolving conflict,
implements innovative approachesand thought-provoking solutions
that obtain favorable outcomesfor even the most complex
matters.
Hesha's popular new bookHolding the Calm shares her
(19:03):
secrets of how to read asituation to resolve tension,
eliminate conflict and restoreharmony.
Get a copy of the book today,available on Amazon.
Click on the link in the shownotes.
Speaker 2 (19:18):
Absolutely, and just like how Ground News focuses on
breaking stories, developingstories into right-leaning,
left-leaning, and they're morefocused on the polarization of
news, we're focused on thehigh-level content or topics and
organization of not just newsbut content in general.
(19:38):
A little background aboutMinimap is news wasn't actually
our first application.
We started with Amazon'scatalog as the first data source
that we were playing with, butunfortunately it was expensive
to access their data.
News was readily accessible.
We felt that was impact.
We're able to.
By just being able to organizecontent this way, it does help
(20:03):
our users understand andnavigate these, these larger
content spaces.
Speaker 1 (20:08):
Yeah, it's so awesome to see how the map just fills
in, you know as you look out.
Speaker 2 (20:17):
It was a lot of fun learning how to write the front
end.
I got good at shaders, if youknow what those are, but I'm
very thankful that ChatGPT ishere.
But another interesting aspectof this and this is what I was
talking about earlier as ademonstration is we were saying
(20:39):
that not only do the articlesabout content matter, but also
the publisher of those articles,and so one of the things that
we can do is, for instance, herethis is the landmark, the logo
for us Us Weekly.
They're a news publisher, theytalk about entertainment, people
, et cetera, and so what you'reseeing here is this is Us Weekly
(21:02):
, placed as the algorithmunderstands them, and so that
means that they're in people,they're over in this area.
We had nothing to do with whereit lands.
That's all driven by how thealgorithm understands them in
the context of the contentlandscape.
But what, then, our users cando is like okay, I want to know
more about this publisher and wecan highlight all of their
(21:26):
articles on the map for whatevertime range that we're looking
at.
So, like here, we're looking ata 30-day window that we're
looking at, and so, like here,we're looking at a 30 day window
and we can see that they dopost or they do create content
that is people focused, that is,fashion and music and
entertainment focused media, andthey tend to stay out of the
legal and political scenes.
They don't tend to report onfinance and business, and so
(21:50):
they are.
We can see at a glance, withouteven looking at a single
article of theirs, that they areconsistent with their branding,
their messaging and the type ofcontent that they say that they
report on.
Speaker 1 (22:02):
Yeah, this I really, I really like that you're
bringing up this example.
Just this morning I saw Galluphad released, you know, talking
about topics that immigration,of course, naturally, you know
immigration remains a topconcern of voters.
But then follow up question assomeone who likes to follow the
(22:24):
news and you know I'm veryinterested in the news landscape
and how things are covered Oneof the conversations that I have
with other news consumers ishow important topics are covered
in different or by differentnews organizations and whether
(22:45):
or not is it covered more oftenby certain news organizations
than others.
There is an argument to be madewhen you talk about bias in the
news, bias by omission.
You know some newsorganizations that might be
right-leaning or left-leaningmight have a tendency to report
on particular topics more oftenthan others.
(23:06):
You know, and you had mentionedground news earlier you know
that's one of the things thatthey like to show, you know, in
their information, what'sconsidered to be a blind spot be
a blind spot on the rightversus a blind spot on the left.
And so if you were to bring upthe topic of immigration, I
would be just because this is atthe top of my mind seeing that
(23:29):
article from Gallup this morning.
I would be interested in seeingfrom this map you know,
visually how that would span outacross the.
You know the different areashere and, like you said, it
doesn't show left leaning orright leaning, but from a
(23:51):
topical perspective, you know,it is interesting to see where
all of the different dots thatyou have here you know falls
under government politics.
Regional news.
Speaker 2 (24:05):
Yeah, absolutely, and so here we just search for
immigration.
This is a search that's donejust by matching, it's not a
keyword search.
This is anything that isimmigration or immigration
similar, and so we can see allthe areas where immigration pops
up and seems to be a prettyubiquitous or universal concept
(24:30):
across all major categories,across all major categories, and
so we can see it, of course, inpolitics and government and
transportation and world news,but it appears to be coming up
in unexpected places as well,like real estate and technology
and entertainment and media.
Even sports has a couple of hitsfor immigration, and so this is
(24:50):
now we're able to see not justthe major or dominant topics
that we'd, of course, expectthis conversation to come up in,
but the unexpected places aswell, such as sports, and so
what I'd like to do here is do arefined keyword search to
actually bring up the articlesthat are specifically talking
(25:10):
about immigration, and then wecan see that this drops down to
77 articles in the last 30 daysthat we've been able to scrape,
and, of course, we're stillseeing a wide coverage, and
we're even seeing hits out innatural disaster, environment
and climate change, and so thatmight be immigration.
(25:31):
These could be stories, orlikely are, that are focused on
displacement because of climatechange issues or natural
disaster, and so that's.
Unfortunately, I lost my search, but that's what we're trying
to show is that there are, ofcourse, the dominant topics and
categories that we expect thisnews to fall into, but there are
(25:55):
blind spots that are going tobe difficult for people to reach
, to see and understand if theywere to only look at these
search results in a flat list oras summarized by a chat GPT
like algorithm.
Speaker 1 (26:11):
Right, yeah, that's so fascinating to be able to
look at this through thisparticular lens.
You know, thinking about USimmigration, and then you know
just some of the news articlesthat are news organizations that
I tend to pay attention to andhow they cover US immigration.
You know, it's interesting tobe able to drill down into this
(26:32):
map, you know, yeah, with thepolitics, and figure out who's
reporting on this from thatparticular aim or crime.
Or you know, like you justshowed climate change, which
news organizations are coveringthis particular topic from that
angle.
You know.
So, for anyone who is reallyinterested in doing a deep dive
(26:55):
and learning more about thisaspect of the news, and not only
what's being covered but who'scovering it and the angle you
know at which they'reapproaching these particular
topics, it's fascinating to beable to do this.
Speaker 2 (27:10):
Yeah, it's been rewarding.
This has been a dream or visionof mine for a long time, and
after my last job, I was in aposition where I needed to have
a crypto firm.
I came in being able to builddata pipelines.
(27:37):
I could do the technical work.
That wasn't a problem.
But crypto was, and still is,volatile.
It changes fast and if you'rein the crypto space, especially
if you're a decision maker, youhave to be able to react and
understand the news around you.
And, as almost a foreigner tothe space, I was not subscribed
(28:01):
to the right people.
I didn't know who to follow,and all crypto news broke on
Twitter first.
It did.
That's where crypto was.
It was on Twitter, it was on Xand without being able to use
X's search, because it's bad,it's really bad.
Earlier today I found somebodyon X.
I was just trying to figure outwho they are.
(28:23):
I looked at their account andit is difficult to even look at
the list of tweets that a singleperson put out there, because
it's conflated with ads.
And here are some other peopleand here are all their retweets.
It's like no, just show me whatthey've actually written.
(28:44):
And so just being unable to getthe information I needed to
understand the crypto landscape.
I felt like I was a liabilityand this had been something I
had been thinking about longbefore I even started that
company, and after that, I feltit was time, and so I took the
leap and started this andstarted building out Minimap and
just to take it a step furtherfor prospective users and anyone
(29:07):
that's interested, one of thethings that I wanted this
platform to be is an actualmini-map for users, and what
that means for us is that I wantto continue browsing Reddit or
Twitter or any of my news feeds.
As you said earlier in thepodcast, a lot of people get
their news information fromsocial media, and I don't think
(29:31):
that's going to change.
I don't think that readers aregoing to start moving over or
new content.
Consumers are going to startmoving over to the traditional
news websites New York Reddit,it's YouTube Shorts, where
people hear these news eventsand get these snippets of
information, and so what webuilt out is actually a browser
(30:07):
extension that tags along withthe user, and so if we do go to
AP News, there's a ton ofcontent here, and so right now I
have the browser extensionplugged in and it's saying to me
that there are 117 articles ontheir front page.
That is overwhelming.
(30:29):
That's a lot of information.
Yeah, I'm not going to.
I can't handle that.
That's too much for me.
And so what Minimap does forthe user is we have this browser
extension, drive by extension,and we can actually show what it
looks, what a user's newsfeedlooks like, and so on one window
(30:52):
I have AP News.
Doing my browsing as usual, andon the map I have all of those
articles highlighted.
I can see, oh sorry, it lookslike there's a bug.
These things are supposed tohide away, but on the map I can
see each of those articleshighlighted and I can see where
there are concentrations, if any, and where there's a sparsity,
(31:14):
if there's no coverage.
And it looks like on AP News'website.
Wow, here we go.
There is a gap in coverage forenergy and transportation.
I know what these mounds arebecause I spend every day with
the map, but these areas it's ahuge blank spot, exactly.
And so they're not talkingabout Boeing in the news, even
(31:35):
though I know Boeing's Starlineris starting to ramp up and they
have all the whistleblowercoverage.
That's right.
They don't have anyconversations out in the energy
sector.
I know that there's a lot ofnew battery tech happening right
now.
People are trying to find likelonger term, you know, zero
waste batteries, and so this istelling me just at a glimpse.
(31:59):
I'm not even reading a singlearticle.
What?
Oh, even here's another one.
Natural disaster is missing onthe map too, which might be a
good thing, you know, that'sactually probably a great thing
if there's no natural disastersto report.
Speaker 1 (32:11):
Right.
Maybe there's just notsomething going on right now,
but I'm so happy that you pulledthis up and you're using this
as an example.
You know, looking at AssociatedPress News, the website, and
then using your extension, andit is giving you an indication
of all out of all of thosearticles that are on the front
page right now, you know there'simportant news that's happening
(32:34):
that that we're aware of.
Just from reading other newsorganizations, you know
information that they're they'republishing.
But then here you can see thatthere there is a big, big area
looks like a desert coveringthat information, and we're not
saying it's good or that it'sbad, but it is notable that it's
(32:56):
missing off of the front page,and so does this extension.
It only scours that particularpage that you're on at that time
, right.
Speaker 2 (33:06):
Yeah, so we have a small list of websites that the
extension is allowed andapproved to work on, and so
that's us telling the Chromestore and the browsers like only
these sites.
And then when we visit, when theChrome extension sees or the
browser extension sees, oh okay,we're on AP news or we're on
(33:28):
Reddit.
If you try it on Reddit, it'sonly oldreddit, not with their
new layout, but AP News, cnnReddit then it will activate, it
will look at all of the contentthat's on the page in front of
the user and then it will sendthat content over to the map to
be understood by the algorithmand then plotted and mapped out
(33:50):
for the user and mapped out forthe user.
And then in that we're able tostill highlight and navigate the
content landscape as we wouldif this was just the bare or
native content landscape or ifwe were just on Associated Press
.
Because again, here's thesmaller, refined vertical feed
(34:12):
of news and I can from there goright back to the page on
Associated Press.
We're not looking to hostcontent ourselves.
We merely just want to be a,not a dictionary, but almost
like yellow pages, we want to beable to say like here, here's
where all the content is thatyou're looking for.
(34:32):
Here's where it lives, and letus get you there and organize
that content spatially for you.
Speaker 1 (34:37):
Yeah, yeah, and I really like the idea of who's
talking about this today.
Speaker 2 (34:43):
Exactly yeah.
Speaker 1 (34:44):
Yeah, that's so fascinating.
So you really didn't have untilyou worked for that crypto
company and you were findingthat there was information that
you needed to be able to accessand you were having a difficult
time being able to access that.
That gave you the idea tocreate this, this mini map.
So have you been doing this onyour own?
(35:06):
Do you have, do you have,people who are helping you?
Like how this seems like it'san incredible amount of work.
To be honest with you, I don'tknow anything.
I don't know jack about coding,okay, but I can imagine this
looks like hella work.
Speaker 2 (35:21):
So I've had help.
I left the company, the cryptocompany DeFi Pulse, back in
December of 2022.
And I've been working onMinimap ever since.
Literally the next day after Ileft, I started working on
(35:42):
Minimap, and at first it wasjust me for several months and
then an old coworker of mineactually a former intern joined
me in, I think, july of 2023.
And so I'd been working on thisfor six months, just going
(36:03):
through and getting the.
Actually, if you give me aquick second, I have a fun
picture book that just shows theevolution of Minimap.
Speaker 1 (36:11):
That's just I love it .
I love it that you have apicture book.
Speaker 2 (36:16):
You know it's a, a it's a visual thing.
We're here to visuallycommunicate things, and so I
feel like it is fun that this isa visual representation of of
the news landscape.
Speaker 1 (36:27):
I think that's why I get so excited about it.
I'm a bit of a nerd that way,but fellow nerd nerds, if you
want to check it out, I highlyrecommend it.
Oh, here you go.
Okay, I see you, yeah.
Speaker 2 (36:40):
So this is the progress, our timeline In
December of 2022,.
I founded Minimap, Iincorporated, created the LLC,
all of that and more, and theoriginal idea was I wanted to be
able to map out search resultsfrom any search bar anywhere
where there's a search bar onthe internet and our algorithm
(37:00):
can understand that content.
That's that's the space that wewant to be, especially for,
like amazon's catalog, which isoverrun with drop sellers, so
there are so many duplicateproducts out there, and so for
us, what that would look like onthe landscape is like that
example earlier, where those twopublishers were making the
(37:21):
identical reports for everysingle precinct in California
and we saw those spikes.
That's what Amazon's catalogwould look like on the map,
where all of those drop sellerswould have hundreds of identical
products.
We don't care how many itemsare in the search results.
What we care about is how manyunique items there are, and so,
deduplicating those spatiallyand just stacking all the
(37:43):
similar things up, we wanted tobe able to show the full space
of search results for users.
And so here, like I'm looking atthis one sneaker as an example
and I just want to be able tohighlight where this is, and so
this was the initial idea andthis image I made by hand.
This isn't, and so the firstit's not to scale.
(38:04):
Is that what you're saying?
Yeah, it's not to scale, madeby hand, no algorithms involved.
This is the concept.
And so the first place to startis let's just start with two
things.
And so the first idea was okay,let's see if we can't make a
spectrum between boots andsneakers and have the algorithm
organize things as a gradientbetween those two, and so, on
(38:25):
the right, here we have sneakers.
On the left, here we have boots.
And then the idea was okay,well, can we have somebody, be
it Skechers or Keen Boots,define those polls, those
endpoints, and just that couldbe the business model.
And so, like, okay, brand youcan represent, like this poll
and other brand, that poll, andthen we can organize any and
(38:47):
every piece of content relativeto those two things.
And so that was where westarted.
And then this here is thatfirst implementation, and so
it's a diamond, for reasons thatdoesn't really matter.
What does matter is that as yougo from left to right, it goes
from most boot-like boot to mostsneaker-like sneaker, and at
(39:09):
the very top is the mostboot-like, most sneaker-like
thing.
Speaker 1 (39:13):
That's a nice combination of the two.
Speaker 2 (39:17):
The maximal combination and at the bottom we
had the minimal combination.
Like these aren't boots, thesearen't sneakers, these are cars
and airplanes.
Speaker 1 (39:24):
Yeah, and then it's like what are those doing?
Speaker 2 (39:27):
Yeah, and over here we have a car sneaker, but the
algorithm saw that, hey, this issubstantially sneaker, so we're
going to that belongs overthere and in some capacity it's
right.
And so this is showing us thatwe can organize things onto a
spectrum, a gradient between twounderstood topics, and also
(39:50):
distill a semblance of relevancywhere we, if you're sticking to
this top most border.
These are going to be the mostrelevant boots or sneakers, or
boot sneakers, whatever you wantto make the spectrum out of.
And so this was the firstsuccess that we had in February,
two months after I startedexperimenting with the algorithm
(40:11):
.
Wow, that's amazing.
And so then in March, we turnedthis into a full 3D thing, and
so now it's like, okay, we knowhow the math works for the 2D
scenario, let's make this 3D sowe can define these topics and
then organize several spectrumsat once.
And so what we're looking athere is these diamonds.
(40:35):
Upside down, diamonds representthe topics.
These are the boots, these arethe sneakers, these are anything
that we thought was relevantfor defining the content
landscape at a high level, andthen everything else, these
peaks and valleys and littlespikes.
That's the content, and this iswith Amazon's product catalog,
(40:59):
and then, of course, you can seethe logos, and so these are the
seller logos, and they areplaced based off of the
algorithm's understanding ofeither the seller themselves or
the content that they're selling.
And so behind logo, there's aBarbie thing over there.
There's in the far top leftcorner, there's mighty skins.
(41:19):
They do like skins for yourdrones or your like unibords,
all that kind of stuff.
But this was this was our firstsuccess with 3d mapping.
However, the math needed to beimproved, and so we started
experimenting.
So I started experimenting withdifferent approaches, some
weird spirals and donuts whentrying to organize these high
(41:42):
dimensional point clouds ofinformation figured out.
I figured out how to actuallymake this real, how to
meaningfully organize thecontent around these topics, and
so I can see that there's alandmark behind that valley over
there.
But that's, this is.
(42:03):
This is the first success, andthis is now the algorithm that
we're running with today.
And so then, afterwards, I hadto move on from Amazon.
As I said earlier, they'reexpensive, they're difficult,
they'd only give back 1,000search results and it would take
10 minutes.
They limit us any of theirvendors, through their regular
(42:26):
APIs, to two requests a second,and if they only give 1,000
requests or results in theirsearches, it would take 10
minutes to act for us toactually get a user's real time
search, and so that wasn't goingto be tenable for us.
And so what you're looking athere is the social media
(42:46):
platform, blue sky, and so nowwe've we, I turned to blue sky
to actually create content maps,and so what you're looking at
here on the right or, sorry, onthe left is the major topics by
the counts of documents and thensome relevancy scores.
(43:08):
This was my proof that I couldsearch for crypto content and
not just highlight all of thecontent that's on crypto island,
but all of the content that'soff of crypto island, because
there is content that isfinancial or not financial.
Let's space, let's say spacefocused, and it's like we
(43:28):
figured out how to useblockchain to do a thing on the
ISS, space, space, space, and sothat's a space dominant topic,
but that's crypto over in space,and so you otherwise probably
would have never been able tosee that exactly discovered it,
and so, like now we can.
I can see all of the places thatcrypto content is popping up
(43:51):
throughout the content landscapethat's outside of a crypto
dominant context, and so thiswas a big success.
And then this is the time whereone of our other software
engineers came on the platform.
We started experimenting withhighlighting by date,
highlighting by sentimentanalysis, refining the actual UI
, and so now it's getting alittle bit more sophisticated.
(44:12):
Refining the actual UI, and sonow it's getting a little bit
more sophisticated.
We're bringing on landmarksfrom from companies on Twitter.
So what I did here is I wentthrough and collected 20
companies and took their toptweets and gave them to the
algorithm, and now all of theselogos that you're seeing are
(44:33):
organized by how the algorithmunderstands their posts.
And one of the reallyinteresting findings here is
NVIDIA.
I thought we had a bug.
I thought that NVIDIA is atech-focused company.
They're AI-focused, they arehuge in the tech space and so
unfortunately, I didn't capturethe topic that this little
(44:56):
island represents.
But this is corporate andcorporate news,
corporate-focused stuff,business, also adjacent to
influencers and social media,and I was like why is NVIDIA
over in the corporate space?
Why is NVIDIA?
Why aren't they over in tech orAI?
And it turned out it's becausetheir top 10 tweets actually to
(45:20):
this day on LinkedIn, on Twitter, across all of their socials
nine out of 10 of their postsare about their CEO or their
executive teams or theirleadership and what they're
doing to further NVIDIA as acompany, and so they are talking
about themselves at a corporateor executive level always, and
so, as far as the algorithmunderstands it, they're not a
(45:42):
tech company, they're aenterprise company.
They're a social mediainfluencer.
They're a enterprise company,they're a social media
influencer.
They talk about themselvesversus the tech that they make,
and so this is an example of amisalignment between what the
company actually does and howthey portray themselves out on
the social landscape.
Speaker 1 (46:04):
If you've read my book cracking the rich code, you
know it is chock full offantastic advice from top
thought leaders and supersuccessful entrepreneurs from
around the world.
How would you like to be amember of an exclusive community
that provides that same how-tocontent from those same leaders?
What if you were able to attendmember-only live events and
(46:27):
interact with them?
I'd like to invite you to jointhe Rich Code Club.
It will change the way youthink about yourself, your money
and your life.
It's the only social mediaplatform purely focused on
helping you grow.
Join the Rich Code Club todayfor free by clicking on the have
(46:50):
.
In the current model, how manycompanies would you be able to
look up specific companies tosee the disbursement of topics
that they talk about or tweetabout or post about in all of
their different platforms?
Speaker 2 (47:05):
Yeah.
So right now, as I showed, weare able to highlight different
companies out on the socialmedia scape and whether or not
there's brand misalignment.
That's still pretty difficult.
What we have to do is then readlike, let's go back to Us
Weekly here.
Oh, are we going to get it?
(47:25):
It should be loading.
And so what it really means isthat just understanding the you
know what they're actuallytalking about versus where the
algorithm is placing them.
And so right now, we're placingthem based off the content that
they produce, and so what wecould do to actually highlight
this is there is a descriptionthat's going to load shortly
(47:47):
about what Us Weekly is, andthen there's all the content
that we have, and we can look atthe difference between how the
algorithm understands theirdescription versus how the
algorithm understands thecontent that they produced, and
that's how we could identify amisalignment or incongruency
between the brand and theircontent.
Speaker 1 (48:10):
Right.
So if I'm an organization, Iclaim to focus my content on
sports, but then I look on themap here and I see, wow, there
is an awful lot of informationthey're producing on politics.
That could be an indication ofa misalignment there.
Speaker 2 (48:29):
Yeah, that actually could be a very fun layer that
we could add to the map to showwhere these types of
misalignments are happening.
Yeah, it would be a fun sideproject for us.
Speaker 1 (48:43):
I'm good for that.
I'm good for for side projects.
Hey, can you?
You know it'd be cool.
Speaker 2 (48:50):
Yeah, but it's it's, it's real, it's impactful and
there are a lot of brands outthere, Like, for instance,
Wendy's food company, but theway they and I'll scroll- back
here.
Speaker 1 (49:01):
Yeah, I mean they're.
They're so famous for their,their tweets.
Speaker 2 (49:04):
Exactly, and so you can see out in the distance.
Here there's five guys justtowards the top of the screen,
but Wendy's isn't there, butfive guys in Wendy's burger
chains fast food.
Speaker 1 (49:15):
So why?
Speaker 2 (49:15):
aren't they next to each other?
And that's because Wendy'sportrays themselves as more of a
meme-y social media account.
They don't really push theirburgers as much as they push
just the idea of Wendy's andself-promotion, and that kind of
stuff, and so it evenhighlights the difference in
media strategies that thesecompanies are taking.
And so for users, again, thisgets back to being able to
(49:38):
understand these large entitiesthat have vast amounts of
content at a glance.
And there's no way that you cando that with any number of
metrics or chat descriptions.
It just doesn't work.
I like to give the example oflike I could draw a squiggly
circle and I could try todescribe to you that circle in
(50:00):
as many metrics as I want, andthere is no way that you could
recreate that circle.
It just it doesn't work that way, and so how do we expect to
understand the dynamics of thecontent landscape when all we
have are metrics, chatdescriptions or vertical scrolls
?
And so that's why I think wehave like something's going to
(50:21):
give, it's going to.
Things are going to change atsome point, whether it's by our
hand or somebody else, but justto continue on.
You know, we kept evolving, wekept iterating, like here's just
showing how similar storiesgroup together and occupy like
very close, like proximity onthe map, and then in September
(50:43):
and October that's when wereleased our beta the landscape
looks very different.
Now we had to change because wehad to make it easier to like
see things.
When you move your applicationfrom the 2D space to a 3D space,
everything goes out the windowin terms of UI design and color
dynamics and all of that, and soit's a tough challenge.
(51:04):
I was not prepared for thelevel of color science needed to
think the light and yeah, andso we just kept iterating.
We filed patents for themapping algorithm that we have,
and that brings us to today withthe content landscape as you
see it.
Speaker 1 (51:22):
Yeah, that's so fascinating.
So where do you see this goingnext?
Speaker 2 (51:28):
We are trying to raise on Kickstarter right now
to get ourselves to market andin front of users.
What we want to be is a mediumbetween users who need to
understand the content landscape.
We're at a point where that'sjust a must now, because brands
(51:48):
have big tools for this kind ofstuff.
There's Muckrak, there's Cision, there's all of these social
listening tools that can breakdown the content landscape into
any number of metrics and nuancy, descriptions and chat.
But it's that's that's stilllike a problem that we're trying
to solve.
Like again, I don't think thatwe can understand the content
(52:11):
landscape through just metricsand descriptions, but the
problem is these socialcompanies.
They have the tools needed tounderstand the landscape as it's
relevant to them, but theirunderstanding of the content
landscape is not the same ascontent consumers.
Content consumers are in thedark, and so we first want to be
(52:35):
a tool for the content consumerjust to get their bearings on
what's going on in the worldaround them, to be able to take
their random piece of socialcontent that they discovered on
reddit or whatever platform and,okay, show me everything else
that's like this, yeah.
And then, second, we want to bea medium that allows brands to
(52:57):
represent themselves brands,communities, anyone, those
landmarks.
Now there's an opportunity forbrands to represent themselves
in.
Now there's an opportunity forbrands to represent themselves
in contextually, in the spacesthat are relevant to the content
that they produce, to thespaces that that consumers are
looking for them.
And so it's either you know yousee walmart and then you see
(53:18):
walmart on the map, you're gonnabe like okay.
Or starbucks food you see theirlogo on the map, be like okay, I
, I kind of know what this isbecause I know what starbucks is
and so I have a sense of whatI'm going to find next.
Or vice versa the user is infood, the user sees a logo that
they've never seen before, butthat logo is in food and it's
next to starbucks and it's likeokay, this is probably going to
(53:40):
be a coffee shop and and so,just like Google Maps allows
users to look at a city at ahigh level and see logos and
landmarks and the little spoonand fork icons or museum icons
and understand like, oh, okay,food districts, entertainment
district, all of these things.
We want to be able to giveusers the ability to
(54:03):
contextualize content, landscapein that way and give brands the
ability to express themselvesand show themselves to users in
the same view and just have aunified view.
Companies and massive brandsview themselves, and how users
understand the world around themis vast.
There's nothing out therethat's attempting to bridge that
(54:34):
gap or even cater to the basicneeds of users with these
problems that we talked abouttoday.
Speaker 1 (54:40):
Yeah, that's fantastic, just from a brand
awareness, brand managementperspective, any organization,
it doesn't matter the industrythat you're in, I mean, you are
a brand, so being able to have amechanism in place to try to
figure out, okay, your consumersdo they have a perception of
(55:02):
you that matches what you hopeit is?
You know one thing that I say alot you know as a communication
person is the message you sendmay not necessarily be the
message received, and I thinkthis is a perfect example of
that.
If you are a brand, whateveryour, you know whatever
organization, wherever you are,whatever it is that you're
selling, or you know the contentthat you're producing, it's
(55:25):
really important to be able tokeep that in mind.
Is the message we're sending,the message that's being
received?
And if there's a misalignmentthere, we need to be aware of
that so that we can figure outwhat changes we might need to
make.
You know in our messaging andhow we're getting it out, where
we're getting it out.
You know, yeah, that'sfantastic.
Speaker 2 (55:45):
Yeah, absolutely.
I firmly believe that there is.
There's a lot to be gained fromjust a shared understanding of
the content landscape and thenfrom there just building out.
Speaker 1 (55:58):
Yeah, absolutely so, if anyone is interested in
playing around with the map orthey want to get in contact with
you or they want to help insome way, because I know that
you have a Kickstarter campaigngoing on right now, so give us,
give us, all the informationthat we need for that.
Speaker 2 (56:17):
Absolutely.
We're looking for feedback.
We're trying to get users onthe platform to understand how
they engage and what their firstimpressions and experience are
with the map.
They're the first contact andso getting feedback.
Join our Discord, give us thatfeedback and, if you like what
you see and you get the visionand you understand where we're
(56:39):
trying to go with this, becausewe want to be multi-domain,
multi-modal in the sense thatit's not just news, but news,
social media, e-commerce,anything that the algorithm is
capable of understanding.
That's what we want to be ableto put onto a map.
My goal.
It's nowhere within reach rightnow, but I'd love to have a map
like this of music.
(56:59):
I'd like to take a snippet of asong that just I really jam to
like.
Sometimes they're just thesesmall 10 seconds.
I'm like that.
Find everything with that vibeand give it to me.
Speaker 1 (57:09):
That would be fun, that would be great yeah.
Speaker 2 (57:11):
Yeah, exactly, and that's where we want to go with,
again, anything that thealgorithm is capable of
understanding.
And so right now we have ourproof of concept, our minimal
viable product out there thatdemonstrates the concepts at
play here, and what we need isto get the funding to bring this
to market, to get stablerevenue and get to the point
(57:34):
where we can expand thealgorithm to our algorithms to
encompass social media,e-commerce and even images and
video things that are out ofdomain of just text, but
multi-domain, multimodal and,yeah, maybe even at a point
where we can do music and audio.
But that's far in the future,and so right now we're just
(57:56):
looking to get funded to getthis crowdsource campaign off
the ground.
Speaker 1 (58:02):
Okay, yeah, so how long is that campaign going to
last?
Speaker 2 (58:05):
So we have 16 more days of our campaign.
All right, we need 23rd.
Speaker 1 (58:09):
So got to get cracking on that.
Speaker 2 (58:12):
Oh, I know, yep, I'm watching the countdown.
Speaker 1 (58:16):
Yeah Well, how can we get in touch with you, any
listeners who might want tocontact you directly?
You mentioned Discord.
Are you on any other socialmedia platforms?
Speaker 2 (58:25):
You mentioned Discord .
Are you on any other socialmedia platforms?
I'm on Discord and LinkedIn.
I don't go on Twitter too much,it's just I have not enjoyed
their experience, for obviousreasons, and so yeah, and if
anyone wants to contact medirectly, it's just Clayton at
Minimapai and I always read myemails, and so you know I'd love
(58:46):
to.
Speaker 1 (58:46):
I'd love to have a conversation.
Yeah, that's fantastic, andI'll make sure to have the link
to the website in the show notesas well.
Well, my friend, it has beenexciting seeing this journey
since the first time you and Ihad a conversation talking about
the media landscape and youintroduced what it was that you
were working on, and just to seesome of the changes even
between then and now, I'm superexcited to see where this is
(59:10):
going.
So we're definitely going tohave to have another
conversation in the future sothat we can think back to this
moment in time and do the wholeremember when kind of thing,
kind of you know conversation,because I think it's fantastic.
So many applications and justfrom a consumer perspective, a
(59:30):
news consumer perspective Ithink it is filling an important
need that exists out there, andthe fact that you're coming at
it from a visual perspective, Ithink it provides something that
all of the other metrics cannotprovide at the moment, and so I
(59:51):
think more information isalways better.
So I just appreciate whatyou're doing and I appreciate
that you're out there makingsure that we can get as informed
as possible, because it's socritical.
We need to have a society, weneed to have a community that is
as well informed as possible.
So yeah, yeah.
Speaker 2 (01:00:13):
Thank you for doing this.
No, thank you.
I appreciate that it'smotivating.
Speaker 1 (01:00:18):
Oh yeah, I'm a fan, so for the listeners more to
come, make sure that you jointhis Kickstarter campaign.
Let's make sure that the thiskeeps growing and the usability,
I think when you get out thereand you start playing around in
there and, like Clayton said,they're looking for people to go
to the website and and giveyour feedback.
(01:00:39):
So that's really important.
Be a part of it now.
How exciting is that you can?
Speaker 2 (01:00:43):
be a part of it and Join us day one.
It's exciting, it really is.
Thank you, jen, I appreciate it.
Speaker 1 (01:00:51):
Absolutely All right.
You have a great rest of yourday and, listeners, I'll have a
great rest of your day as well,and we'll see you next time.
Take care now.
Thanks for listening.
If you enjoyed this episode andyou'd like to help support the
(01:01:11):
podcast, please share it withothers, post about it on social
media or leave a rating and areview.
Popular Podcasts
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
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Therapy Gecko
An unlicensed lizard psychologist travels the universe talking to strangers about absolutely nothing. TO CALL THE GECKO: follow me on https://www.twitch.tv/lyleforever to get a notification for when I am taking calls. I am usually live Mondays, Wednesdays, and Fridays but lately a lot of other times too. I am a gecko.