AI & Business w/ David Hirschfeld, CEO of Tekyz | Tech Talk - Today In Space

Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Music.

(00:08):
Welcome everybody to Today InSpace. I am, as always, your
space science podcast host fromthe East Coast, Alex G orfanos,
and we're here to talk a littlebit of tech, a little bit of
business. And we have our guest.David Hirschfield, thank you for
joining us.
My pleasure, and thanks forhaving me on. Alex, yeah,

(00:29):
absolutely.
Thank you for reaching out. Andyou know, we're just like,
let's, let's chat tech andbusiness. And this is, I love
this stuff, you know, the spacestuff I'm super into. But I
think most of my day is actuallymore around business and tech.
So this is, this is really fun.So tell us a little bit about
yourself and who you are, whatthe company is, what you're

(00:49):
doing, and then we can, we candive in from there. Okay,
I've been in the softwareindustry for almost 35 years
now. It started out inenterprise, and then ended up in
consulting in my early career.And did was project management

(01:10):
for enterprise like TexasInstruments, Motorola, Intel,
Allied Signal, Arizona publicservice. These are places that I
was involved in project managingfor a contracting group I was
working with, and then I startedmy own software company,
logistics, Route, distribution,inventory management,

(01:31):
which is no joke, like, if thatgoes wrong, it's that oh yeah,
can't run
Yeah, right, especially whenthere's delivery and inventory
and trucks and all thatinvolved, right? Yeah, and
multiple warehouses anyway. Sowe did this. And it started
because a friend of mine, mybrother in law at the time, was
I just gotten out of themilitary, and he wanted to go

(01:52):
into business, and so he boughtsome vending machines. This is
how the whole thing started, andand they was looking, and he had
six or seven machines, and allthe coins and the money and the
inventory, and he was trying tofigure out how to track all this
stuff. And he went and lookedfor software. And back then, it
was all this expensive Unixstuff you needed servers, you

(02:13):
know. And there was nothing thatwould run on a PC. Which
windows, three, one had justcome out. I mean, this, you
know, the first version ofWindows that was really usable
it, yeah, Windows happen, right?Yeah. So I said, Well, why don't
we? And so I was working atTexas Instruments at the time,
so I said to my partner, whydon't we? We are the guys

(02:34):
working with, why don't we starta software company? Because I
always wanted to do that, and Ithought this was a good let's
and we had been talking about, Isaid, Let's build software for
venting operators, right? And wecould, you know, just this sort
of as just to kind of get ourfeet wet with a software
company. You know, we'll createsoftware. We'll put some ads out
in the in the the tabloids forthat industry, and, you know,

(02:59):
we'll start to get somecustomers, and kind of have a
little annuity going from this,and learn a little bit about
software business, you know,Famous last words. So anyway,
and so we, despite, you know,the of course, it nothing is
that simple. But we startedselling, and because it was on
Windows and it was reallyavailable a lot of these smaller
vending operations, we gotswamped at our first trade show.

(03:23):
It was the funniest thing,because we're in the back corner
of this. These are big and theseare big trade shows, because,
yeah, all the big candy makersand soda companies, right then
you have all the equipmentmanufacturers. And there, you
know, we went. We had no idea itwas going to be like that. And
I'd been to big trade shows inthe tech industry, but I just

(03:44):
surprised me how big this was inthe snack food industry. So it
was massive, and we and we gotin late, and we were like, way
buried in the back corner atiny, little 10 by 10 booth. We
even went to, like, Kinkos andgot a banner that was made out
of plastic, which had the nameof our company, Ben master and
and with when you use PVC pipeto push it out, so you could see

(04:06):
it on both from either side ofthe aisle when you're walking up
and down. I mean, it was like,so cheap, right? The whole thing
was like, you know, we weren't,we weren't really trying to
build a real company at thetime. We didn't think we were
anyway, and we were buried. Wehad lines out the booth because
we were the first product onWindows. It was like accessible

(04:27):
to this mass of operators,right? So anyway, that's that
was the beginning of it. Wedidn't eat for three days lunch
because we couldn't pause whilewe're doing demos. It was so
funny that fire marshal came andtold us, You have to clear this
out. And I said, talk to thepeople out there. We're doing
demos. We actually got introuble. So anyway, that

(04:51):
company, despite every despiteeverything we did, it grew to
800 customers in 22 countrieswhen we sold it to a publicly
traded firm in two. 1000. Sothat was my first experience
with the startup, and I thoughtI was, you know, hot shit, and I
knew what I was doing and and Ihad no idea what I actually I

(05:11):
was VP of Products for thecompany that bought us for the
next three years, until leftthere and then went out on my
own again, and it was anotherfew years before I decided to
start techies, and that was 17years ago, so But and then with
techies, I've worked with a lotof companies of different sizes,

(05:33):
but especially a lot of startupsin the 85 or 86 startups at this
point that we worked with, and acouple of them were really
successful. One of them,Broadway producer, built a
product that is used to produceevery single product now every
show on Broadway for the lastsix or seven years, and the

(05:54):
Olympic opening games and WinterOlympics and I mean, every
cruise line uses it for alltheir stage per die. It's really
very successful product. TimCook announced, did a video
montage of six or seven iPadproducts when he announced the
iPad era. I think that was eightor nine years ago. Yeah, and

(06:14):
this was one of the products onthat video montage. So wow, had
a couple customers like that,but the vast majority fail. They
all fail for the same reason.They fail because they lack
product market fit. They waitway too long to get out in front
of the customer with theirproduct and ask them for money.

(06:36):
That's basically the root ofproduct market fit. People say,
I love that. I'm definitelygoing to buy it when it comes
out, and then when it comes out,you have really no idea if
they're going to buy it or not,so until somebody's writing you
a check consistently and enoughnumbers so that you have you can
take a lifetime value of yourcustomer versus what the cost is

(06:58):
to acquire each one of thosecustomers. And if you don't have
a three to one ratio betweenthose two numbers, you don't
have a viable business. So
it's really interesting. Youknow, having worked in the tech
industry for probably a decade,you know, you said 35 years
you've been in this how has thatchanged? Because I can for

(07:19):
someone that's been in both thehype of working at a place like
Apple Retail and then, you know,working in 3d printing. Now,
I've seen hype for sure, and itseems like there's always a
bunch of people who are, who aregung ho on saying, Sure, let me
know when it's out. We'll buyit. And then they're, they're
just not to be seen after itcomes out. Is that something

(07:42):
that has changed over the 35years? Is that something that
that's just been that's just howpeople That's
just human nature? Yeah? Wherepeople say, Oh, well, in fact,
there is a probably the bestbusiness book I've ever read.
This is a little tip toeverybody out there that ever
thinks they want to, well, evenif you're not going to start a
company, it's just reallybrilliant book. It's called the

(08:03):
Mom Test. And the reason it'scalled the Mom Test the guy who
wrote this book, you know, Ireally got to look his name up.
I can't. I keep forgetting whoit is, but he's had several
startups, very successful exits,and he basically says that
everybody's gonna lie to you.They don't mean to lie. They

(08:25):
don't even know they're lying.They just want to be nice, not
you know, they'll either bejerks or they'll be nice. And if
they're jerks, they're justgonna say no, no matter what you
say. But if they're nice,they're gonna say yeah, I
definitely will, no matterwhether they really will or not,
and they don't even realize theywon't. Yeah, so you ask, it's
kind of like giving people freebetas. Yeah, they, you know,

(08:47):
it's a beta test. Sure, I'll goahead and try your product and
test it, and they may or may notuse it and and then they tell
you what they need, want it tochange because they don't really
like it. And then when youchange it, and then you start
asking them for money they don'twant to actually buy it because
they're not really using it orwhatever, right? So it's called
the Mom Test, because if you goto your mother, assuming you
have a good relationship withyour mother, but let's say you

(09:09):
go to your mother and you wantto start a business, you say,
Mom, I've got this great ideafor this really cool business.
You tell her about it, and yougo, what do you think? Well,
what is she going to say? No,you're stupid. You're going to
fail. You'll never amount toanything. No, it's not what a
mother with a, you know, ahealthy relationship, right?
Yeah, right. I love you. I thinkyou're brilliant. I know you'll

(09:30):
be successful. This is such asmart idea, right? That's what
she's going to say, becauseyou're not asking the right
questions. So how do you askthose questions and get actual,
honest, valuable answers fromeven your mother. That's why
it's called the Mom Test. I loveit. I
looked it up. Rob Fitzpatrick,does that sound? Oh,
thank you. Rob Fitzpatrick,yeah, yeah. And it's just, it's

(09:52):
just really well done and smart,and he basically turns the whole
thing around into a clinicalexercise. So. So instead of
loving the product and lovingyour idea, which is the biggest
failure, biggest indicationwhether a founder is going to
fail or not, if they come at theproduct from that perspective,
versus taking off that blackrobe and the vision religion

(10:16):
robe, right, and putting on thewhite robe, the white coat and
becoming a clinician, whereanything, anything you don't
know for sure, is an assumption,and you to go test it. And so,
and how do you test it so thatyou're getting accurate metrics
on the things you're testing? Inparticular, you think it's a
good idea? Is there? Is it agood idea? So what are the who

(10:39):
cares about the idea whatproblem is it solving? So
founders who are consistentlysuccessful come at we are
there's a big problem in thisindustry. We believe we can
solve this problem. We know ithas a high cost, and that people
perceive it as having a veryhigh impact to them because this
problem. So we know how to reachthese people, because of the

(11:00):
high perception of this impact,they'll lean forward when we
talk to them about theirproblem, and they'll see if we
can fix it, and then we talkabout ways of mitigating that
problem, and they start to getvery motivated, right? So that's
really different than talkingabout what a great product and
trying to convince people howwonderful it is, or believing
that people will just see thewonder in it, right? So, yeah,

(11:23):
those founders are much aresuccessful a much higher
percentage of the time. Theylove the problem, and they love
sitting and talking to customersabout their problems. And I
want to touch on this a littlebit, because I, I love where
this is going, and especially,you know, with this audience,
we've got a lot of people whoare in STEM or work either in

(11:44):
academia or they work in theprivate world, but in a science
perspective, that that clinicianway of thinking at it, I don't
think many science minded folkshave that opportunity early in
their career to Understand howmuch people and business, and
how much business is people ofallowing you to do science? And

(12:08):
you know, we're, we're at thisintersection now where private
space and public space are arecolliding, and we're going to
figure out here with the nextadministration what, what
happens with all of that, but Ithink it's, it's something I
would love to to pick more on,and how much about it is

(12:29):
understanding people and and,you know, I think many science
minded folks critical thinkers,they don't necessarily think of
it that way, and they kind ofshy away from it because it
seems emotional. Shy away, shyaway from what. Shy away from
understanding, kind of like, Iguess, how people are impacted,

(12:50):
not by the science, right, likeor like, at the end of the day,
like you might be all focused onthe details of the things that
you're doing, or doing the bestscience that you can, but
realizing the environment thatyou're in right, even in an
academic sense, you're stillgetting funding that allows you
to enable that. So like youmight have to go and talk to
people who are going to helpinvest in that, and how much you

(13:14):
know being a customer facingtech or science role, how
valuable that
can be, so interesting, it'salmost the opposite of your
typical startup founder, right?Because that's all they do is
wear the right white robe, butthey're not, but they're missing
the and they are looking at theproblem, and they know who in

(13:37):
the problem impacts, and they dostudies, right? And but they
don't, but the understanding ofthe vision and the actual
community themselves, thenthat's not, they're not involved
in that process, right, right?So I think an understanding of
that would help you almost wantto take that person and make
them the executive operationsperson for this founder with a

(13:58):
vision, right? Yeah, yeah, yeah,because it's you got to blend
those two mindsets to besuccessful in a startup. 100%
Yeah, yeah. So, so 80. You said,over 80 different startups that
you guys have worked with andgone through this process,
close to 90. Now, yeah, oh no,not gone through. You mean,
launch first now, then I've gonethrough the startup process

(14:20):
with, Oh, yeah. And so, yeah,yeah, launch first is a lot
newer than
that. Okay, talk to me aboutlaunch first. Okay, so
launch first. So, like I said,you know, founders failing these
high numbers because they don'task for they don't ask, they
don't get in front of theirpotential customers early, and
they invest all this moneybuilding the thing that they

(14:42):
believe. You know that wordbelieve that's a killer. That's
a killer for startups. You knowthat they believe it people need
or want. So a couple things.One, and this isn't quite
launched first yet, butfounders, founders that I hear,
they've got this vision theywant to do. This new social
media thing, they don't, youknow, it's not a particular

(15:04):
problem in an industry that theystruggle with necessarily. They
just think this would be reallycool, that the chances that's
going to be successful is likeone in a million. You have a
Zuckerberg and you've got, youknow, you know, who just had
that incredibly rare one in abillion cents for the right
thing at the right time, and hewas at the right time in terms

(15:25):
of just the trajectory oftechnology, when he decided to
try to make all the connectionsthrough social media. And
because social media that wasout there was just not really
effective, and he had some goodideas, but he was able to
quickly test his ideas.
He was a customer, and was ableto deploy it in the customer

(15:45):
base. Yeah,
he wanted something that madeconnections a lot more personal
and right and and private, andso you could build your networks
independent of being some publicthing. And he was, and he was
able to build something, a quickMVP in a month, or whatever it
was. He's also a brilliant guy,and he was able to put it out

(16:06):
there. I mean, it was prettyugly product initially, and this
was a founder who builtsomething that like when it when
it started to take off, in thefirst year or two, they were
crashing every 15 minutes. Imean this literally, and they
had to restart servers and scalememory. And I mean, because they
didn't have any cachingcapability, had to write

(16:27):
Memcache. And I mean, it was, itwas, you know, it was a classic,
we're going to do a simple MVP,and then this thing took off.
And there was no scalability,but it was so popular that
people still that is so it'sunfortunate. I know I think my
mic keeps switching back, yeah,just

(16:48):
switched out a little bit. Yeah.Okay, there we go.
Yeah. It's a problem with mymics connection if I just
breathe on it wrong. So I'mgoing to fix that next week. So
I can't even begin to tell youhow rare somebody like him is,
or somebody like Elon Musk, orwho, who, by the way, he did

(17:13):
launch first he came up withthis concept for only for his
electric cars, right? And he presold how many of those
Sportsters, right? There was nocar. It looked like there was a
car, but there were no cars. Sohe did pre launch sales in a big
way to prove that there was amarket and and to help him fund
it, right? And so. And then he,it took him forever to come out

(17:35):
with the Roadster, but he, youknow, that people were in love
with the concept. So, right? Andthen you've got, where was I
going with this they, you know,
just how rare that combinationis.
So rare, yeah, but, but, youknow, he, you know, first of

(17:56):
all, you know, you're talkingabout people with IQ and vision
and the rare combination of thetwo things that are just, you
know, that create unicorns,right? But people talk about the
problem with most, with founder,a lot of founders, is that they
are driven by the what the VCworld has created. And that's
this concept that you've got tobasically get funding to build

(18:19):
your product and to build yourcompany early on. And it's just
not true. So you can self fund alot of things yourself if you do
pre launch sales. So the ideawith launch first, and it was
born out of the fact that weused to do these, or we still do
so fairly sophisticated designprototypes that look and feel

(18:40):
like a real product, but thenthere's nothing behind it. But
we do that because we want toiterate through all the user
experience nuances with thefounder early on, and we'll
spend a lot of time, and it'sless it's dramatically less
expensive than doing iteratingthrough the development of the

(19:00):
product and trying to figure outall the workflows at that time,
and because then you go, No,that's not going to work, and
change it now, you're rewritingcode, and it's costly and it's
slow, so we do it all in thedesign until we've got something
that looks like a completelyfinished product. Your vision,
not the MVP, but like the thetwo year roadmap of your
product, because that way, youknow, it all holds together. The

(19:21):
Vision makes sense. It's reallygoing to provide the kind of
value you believe it's going toprovide. And then they're so
realistic. I thought, why don'twe go to client? And then we
started doing that about 1011,years ago. And then I said, so a
few years back, I said, Whydon't you try to convince my
customers to go to theirpotential prospects that would
buy it and say, Why don't gothere and demo this and see if

(19:42):
we can't get them to buy inearly I mean, if this was really
that important to them, and yougive them a good enough deal,
like a lifetime license, if theypay for 12 months subscriptions
in advance, whatever the rightvalue proposition is for that
customer. And and we tested thiswith if. Four different
businesses. Three of them, wewere very successful in these

(20:03):
pre launch sales. And one ofthem, we weren't because the
pandemic had just started andand we were marketing a product
to Native American reservations,tribal councils who had all shut
down, and you could not get ahold of them. So I don't know if
it would have been successful ifthey were accessible at the
time, but that was the one thatwasn't the other ones were

(20:24):
really successful. One was areal estate product that we in
the first 60 days. We sold 23 itwas for real estate portfolio
investors, and it was aportfolio management system. And
we sold 23 licenses for close to$70,000 in the first 60 days.
And that's what kicked offdevelopment of that product. The

(20:45):
second Another one was foraerospace parts distributors. It
was a a when it doesn't matterwhat the product was, but it was
an important product there. Andwe did three demos. All three of
them sold for $15,000 each,right? 45,000 you know, that was
three demos. And there's, youknow, 15,000 potential clients

(21:06):
in that market. And the thirdone was clinical trial software.
We sold two copies in with twodifferent demos for 250,000 for
each in pre sale, no software,right? Just demoing, what very
sophisticated looking prototype,not just design mock ups, but
much more than that. It didn'tdo anything in but it looked

(21:28):
like it was a finished product.And so it's like, okay, this is
definitely a thing, and thismakes a lot more sense to do
this with the real estate onethe first few demos didn't quite
do it. We had to go back andmake some changes based on
feedback, and then it started tohappen. But the changes we can
do in a couple of weeks, sowe're able to pivot very easily.

(21:50):
There was one more since that wedid for a couple people in
Hollywood that were trying tobuild a product that was anyway
that one we got to the demostage, and nobody could was
interested in buying it. Wepivoted a few times, and that
one failed, but it failed fastand cheap because we didn't
start development right. And Ican't tell you all the destroyed

(22:15):
lives that I've seen withstartups where they spend years
and hundreds of 1000s ormillions of dollars of investor
money, and often a lot of theirown money, yeah, and they end up
going bankrupt, or marriages getdestroyed, or, you know, just,
it's just, you know, personaldisaster. So one how fast and
cheap is huge, yeah,
fast and cheap. Fast and cheapis, is, is the game, right? I

(22:38):
mean, I've, I've been aroundtech long enough to have seen so
many like startups and promisesmade that never come through.
And the amount of money, youknow, we're living in, the post
the river is full of of fundingera, you know, in it
never was it never was good. Itwas easier yet, you know, 10

(23:02):
years ago than it is now, but itwas
never the promise that everyonewas was was dreaming of. Yeah,
so
this is from three years ago,when funding was still more
available than it is now. Yeah,and I've corrupt. I've read an
article that had these metricsin it, and I and then I have
some friends who are in who areVCs, or work in VC firms. And I
corroborated this with threeother people. For the typical

(23:26):
VC, for a healthy VC, becausenot all of them are healthy,
right, but for a healthy VC,they'll get 3000 pitches, pitch
decks, of which they dispel 1500of them. You know very just with
these are analysts that arequickly just going, it doesn't
fit our model, whatever it is,right? And then out of those
1500 they do a deeper dive. Andout of those 1500 they'll take

(23:51):
300 to a senior analyst, andthen they'll whittle it down to
60. And then those 60 go to thepartners, and they'll pick 30 of
those 60. So 3000 to 30 getfund, any funding at all with a
typical VC. And out of those 33,of them will two to three of

(24:12):
them will be successful andscale or exit or go public. And
the rest of them are anywherefrom they barely are making a
living, to they fail and go outof business, or they sell and
get their money back out of it.So that chances of success, even
from the VC perspective, if youthink of the filtering system

(24:32):
they have, is just unbelievable.Right point, oh, 1% are right.
Basically get was it? Point 1%3030 my math isn't happening. 30
versus 3000 so point 1% getfunding out of all the pitches
that go right, yeah, right. 30,yeah. So, or 1% whatever it is,

(24:55):
and then out of that, it'sanother 10th of it's a less than
a 10. Usually about a 10th ofthat that actually are
successful, right? I think thosemetrics are 80% 80% 70 or 80% of
VC funded startups fail. That'sthose are metrics that I've
seen. I don't know. I think it'sprobably lower than that,

(25:16):
because not everything getsrecorded properly when things
fail. So but if you're nottalking about VC funded
startups, it's 95 to 98% ofstart of tech startups fail,
because most of them never getinto a point of any kind of
metric that gets measured well.And
when you think of that of thoseodds, it's, it's something I've,

(25:39):
I've learned having, having athree printing business here and
this podcast, and seeing workingexclusively in startups for the
most part, like that, fail fastand and cheap, like it really is
more about the amount ofattempts than it is about
getting the one idea that, likeyou were saying, that you Love,

(26:00):
that you believe in to work onthe first try. It's just keep
testing, like, like, thatiterative approach of, you know,
how many attempts can you get atthis? You know, we do that withd
printed products. Here we've gotan Etsy store. We have an idea
that it's like, I have a problemI want to solve. I make it, I
test it once I'm happy with it,because I'm the user, I put it

(26:22):
out there, and if it works,great, if I get feedback, we
make it better, you know. And,but that's super cheap, that's
exactly right, yeah, you know,yeah, you've got a built in
testing lab, the way that youguys run, and that's so
important, right? Yeah, becauseit may work, but it may not
work, right? And so you mightget enough sales at least, so
that it's enough to fund fixingwhat's not right with it, right

(26:46):
or or not right, and then, andthen you'll refine it. If you're
still not getting more sales,even though it works better,
then it's just, then it's not,there's not product in America
fit or the market is so tinythat the cost to reach that
market is too high. And so this,you know, it's the same thing,
right? Yeah, so a founder'sfirst job, number one, and only

(27:12):
job when they want to start acompany. And nobody ever I've
never heard this. I've neverheard anybody say this, except
for me. I, you know, if you readthings like the Mom Test and
other then you kind of get afeeling that they're implying
the same thing. But to me, it'slike you have to identify all of
the niches, really tight nichesthat you can possibly market to.

(27:33):
You need to know what they allare. You don't have to pick one
yet, but you will very quickly.And then you need to then sift
out all the problems that you'resolving from a very root level.
If you sit in this, if you standin the shoes of the stakeholder
in each of those niches, how dothey define that problem at a
very personal level? So thatwhen you say the to them, I do

(27:56):
struggle with and they go, God,I hate that. You know, it has to
get to that survival kind ofmechanism level. And any problem
statement that you have that'sgeneral, but is important. If
you just ask, why does thatmatter? Enough times you'll get
to that root survival level.Reason why that stakeholder
needs that fixed and and that'simportant because you can't get

(28:17):
their attention if you don't, ifyou don't distill it out to that
level, right? Right? You want,if you can't get them to go, Oh,
I hate that. And then they opentheir eyes and lean forward and
say, Do you have a solution forthat? Right? That's what I need,
right? That's how you knowyou've got your potential early
adopter. So you got to figureout which of those niches and

(28:38):
and which problem statements,which niche is the one that has
the highest perceived impact totheir top two or three problems,
perceived and separately, youknow, hit my mic again and knock
my sound still good, though Ikeep like I told you, I start
waving my hands. I'm in trouble.And then, and which of those

(29:05):
have the highest perceivedimpact, and separately, highest
cost to that impact, becausethose are two independent
correlations. You may perceivethat this has a huge impact to
you, but it doesn't actually,when you actually measure it, it
doesn't cost all that much, butit just me, maybe an
inconvenience factor. You'reafraid it's going to somehow get

(29:28):
you in trouble, whatever it is.But it perception wise, it's
high. Conversely, you may havesomething that's very high cost,
but maybe it's in an industryeverybody's dealing with this,
and you always have and so ofcourse, we're always going to
deal with this problem. So theperception of the impact is not
high because it's just been withyou forever, whatever the
reason. So what you're lookingfor is both of those numbers

(29:49):
high, because the perception isyou can get them to hear listen
to you, because they want itfixed. And the cost is you can
charge enough to be profitable.Which niche is it and which are
the top? Two or three problemsthat achieve that. That's the
number one job a founder haswhen they start a when they
start a startup, but, but no.Usually nobody does that,

(30:12):
because nobody's telling themthat's required and and
methodologies don't clear thatmake that very clear anyway. So
where
do you think that is thatsomething that is like an
industry falsity, like peoplejust believed it, or people ran
off that for a while, like, I'mthinking like early days of

(30:34):
Steve Jobs, is everyone tryingto be Steve Jobs without doing
the work? Like
they should try to be Steve theyshouldn't try. People should not
be trying to hit grand slams.Just the chances of hitting that
Grand Slam, we all know howoften that happens, right? Like
in a World Series game, in aGrand Slam happens, right? Then
you're going back, like, howmany decades to find somebody

(30:55):
else that's done that? Right?Yeah, totally. That's not what
you you want to hit. You want toknow, you can hit a solid single
most of the time. That's whatpeople if you can build a
business and find that nichecustomer, and you get product
market fit, which means theywill you know that the amount of
energy it takes to get acustomer is 1/3 of you know,

(31:15):
from $1 perspective of what thevalue of that customer is from a
lifetime on average, the pet,there's different things that
change that ratio, but ingeneral, and you can do that in
an industry that's specific witha niche, and you can build and
marketing around that niche, andcan satisfy the niche and grow
your business. You can get to1,000,003 5 million in the in a

(31:37):
year, two, three years, right?You hit a home run? Well, you
may not think it, because you'renot hitting that 100,000,500
million, you know, get VC, butyou can do this also, most of it
self funded, or maybe somefriends and family money, if
really required, but you'renailing it. And now from there,
you've got a platform with a bigcustomer base that you can

(31:59):
leverage to grow into 10 $20million right? And then you'll
get 10 to 20 20 million, oryou're doing these pre launch
sales right now, sudden, you'vegot this growing customer base.
You don't even have a productyet, and growing revenue. And
now, if you do want to raisemoney, it's a whole different
discussion with an investor thanit is when you're trying to
pitch them on an idea,

(32:20):
right? Yeah, if you're bringingthis is like that. I think what
most people would think of froma shark tank perspective, it's
like, Hey, you're already makingrevenue. We're not. This isn't
theoretical, and
you don't even have a productyet. Yeah, yeah, exactly.
There's something here, right?Yeah, you can get an investor's
attention really easily. Yeah,
yeah, no, that's veryinteresting. That's very
and some of your customers mightbecome investors. You know, it

(32:42):
depends on who
you're, right, right? If it'sthat valuable to them, they want
to get in? Yeah, totally. Oh,man. So, I mean, so the space
world, the space industry, asfar as tech, is in a really
interesting position, wherewe've got these private
companies like SpaceX and RocketLab and Blue Origin that are
like providing the resources togo to space for the public

(33:05):
entity, NASA, right, andtaxpayers. So then we've also
got the military side of thingsand the defense side of things,
where a lot of the tech isreally old and hasn't been
updated a long time, and they'rejust getting to this point where
a lot of this new tech is goingto end up in this you know,
Space Force, for a lot ofdifferent things that they're

(33:27):
doing is advancing the techthat's being used in space right
now. So right there's all theseopportunities right now for
different things to happen. Idon't know how close you've been
following it, but I wonder howyou might think about this as an

(33:48):
approach. And just like, Haveyou been following the industry
growth at all in space,
a little bit, that's not, Imean, I love, I love space and
all the tech that's coming outof that, you know, and, but no,
but from a business perspective,this, most of this, you know,
it's very few startups that Italked to that actually are

(34:11):
because, because we're, youknow, they're all software
startups. And right spacethere's got, it's always
hardware components rightassociated with it, and then
software that supports ourhardware. So it might be IoT
devices or, you know, but not asmuch on the space side. But I do
know that, like for a startupthat wants to get involved with

(34:32):
government contracts, again,same sort of thing, there's a
successful technique. I've nottried it or experienced myself,
but where you go find a you bakerelation, you figure out a
problem that needs to be solved.You identify it in some agency,

(34:52):
whether it's space or whetherit's Homeland Security, whatever
it is. And you and you know it'sa problem because there's
public. Uh, discourse around it,right? And whoever the
stakeholders are in thatdepartment are talking about
this. And so you you spend asmuch time talking with people in
that department would that havesome kind of budget control in

(35:14):
that world. And if you talkabout ways of solving those
problems, they might be simpleworkflow things, right, right,
which right? Workflow Automationis like the coolest thing ever.
It's so it's been aroundforever, but when you actually
are successful fixing a brokenworkflow, the impact it has on
the people that are affected byit is dramatic. And I can give

(35:37):
you a really, really simple buta really powerful example of
that. I'll do that in a minute.So, but if you find that person,
you figure out what thatworkflow is, you come up with
that product that would fixthose workflows for that agency
and others. And you maybe builda design prototype. You go to
the people that are thestakeholders, you show them, and

(35:58):
if they really like it, then youdon't even have to be a
government contractor. They canconnect you with one that you
can sub for and direct them to,which is a way that some people
get into government contracting.It's sort of a way around the
limitations of not being on GSAlist and not being, you know,

(36:20):
and now you're working in a, youknow, for basically delivering a
product to government, to thegovernment through, you know,
some major government contractthat already has contracts,
right? And they're just subbingyou to basically, as a way of

(36:40):
bringing you in. Yeah, I don'tknow if that is helpful in terms
of what the question, kind ofquestion you were thinking of
asking, but it still goes tothat whole pre sale concept,
right? It does.
It totally does. And like so I'mthinking of the space industry,
and you mentioned hardware, andthat is totally a huge limit on
how far these companies can go,because, you know, even if it

(37:04):
involves going to space, youstill have to get the space and
until the last five years,where, you know SpaceX is
launching every day, almostgetting to space could take you
years, and your company mightnot even survive the amount of
time it's going to take for youto get On that rocket.
Oh, you mean, if you want tocompete with them, sure compete

(37:25):
with us, SpaceX. Oh, you'retaught, yeah. Well, first of
all, you have to start out. Youhave to be a mega billionaire,
right, right, if you're eventhinking of going
there. Well, I'm even talkingabout, like, all the small
companies that work together. Sofor instance, like putting a
satellite the cottage industry,yeah, like there's so many
different companies that areinvolved, or subsets that are
involved in getting that wholepayload into space. And from a

(37:49):
software perspective, it feelslike you have less cost and less
like run out of waiting for thatthing to get the space where
there's so many things thatcould be potentially worked on,
from the ground systems thatlaunch the rockets to the
companies that plan thesemissions, like Software wise, I

(38:11):
feel like there's so much moreopportunity today and in the
next five years, than there isin competing with someone like a
SpaceX that's a That's amarathon of a business,
yeah. I mean, yeah, that, yeah,trying to compete with SpaceX,
if you're not Blue
Origin, with Jeff Bezos. I mean,he's got money to keep that
thing running for

(38:32):
you've got to, because you're,well, he's and, you know, he's
got, he's taken an Amazonapproach to that. How many
decades did Amazon, or how manyyears did Amazon lose money in
huge amounts before it startedto be a show profits, right? I
think a decade at least, yeah,but the and, and, it's funny, I

(38:52):
just had a conversation withsomebody about this too. And the
thing with Amazon was theycontinue to get more and more
investment because, becauseBezos was a master at
articulating the need to own themarket, to own all these
different factions in themarket. So he was growing really
fast, but he had to continue tobuild out his infrastructure and

(39:15):
grow faster so and and he wasable to convince investors to
continue to fund, fund thisbehemoth that was losing money
every year, but growing so fastthat because otherwise, if he
didn't grow that fast, theWalmart would have come in and
taken the market away from him,or other the other big, really

(39:36):
deep pocketed organizations thatdidn't have the vision he had,
that, you Know, the end, thatweren't proving product market
bit, because he had productmarket fit. He had massively
growing, growing sales. But theproblem was his costs were
growing out of proportion tothat, and
they really, I mean talkingabout addressing a niche and and

(39:59):
a. Illusion for people you know,like getting, like, that next
day shipping and that Primeshipping, and just ease of get
buying and getting something
that I have to go to the storeto buy something. What happened?
What's wrong with Yeah, yeah,that shift didn't take very
long, right? No, no, it
was as soon as people got ataster, like, Oh, this is good,

(40:22):
right? Yeah. Oh,
you want to know another bigshift like that that's happening
with it will be happening, andit's going to be like a tipping
point where all of a sudden it'sjust Matt. So I'm going to make
a prediction. We'll talk againin seven or eight years, and see
if my predictions comingthrough. But I think that the
that the day of the multi storyparking garage, or underground

(40:47):
parking garage, like under,underneath the shopping center,
those things, if you're investedin it, sell your investment as
quick as you can, because it'sgoing to be worth nothing in
about somewhere between three toseven years. I'm not sure
exactly when that tipping pointwill happen. It'll be really
fast, though. It'll happenalmost overnight. It'll feel
like it, yeah, why do you think?Why do you think I'm making this

(41:11):
prediction? I'm just my firstthought
that came to my head was they'rejust like, I so I'm a
millennial, and getting mylicense, it was in my permit was
like a big thing at 16, I wantedto be able to drive myself
around. I can't tell you howmany kids or people who are
younger than me, you know, Gen Zdon't even have a license. I

(41:32):
have.
What is license? What is thatabout? Why do
they they have Uber. They haveUber. They have everything at
the tip of their fingertips. Whyspend the money and then have
incur the cost of fixing a carwhich, well, you know, I do get
that. I do get that in some way.Yeah,
but what about all the peoplethat are 40 and 50 and 60? I

(41:54):
mean, so it's going to beeverybody. Here's why. Here's
why. I think, Okay, have youdriven in a Waymo or, you know,
self driving car?
Yes, yes, okay, right?
They're cool. Yeah, they're verycool, and they're incredibly
safe compared to you drivingyourself. So I go out to Phoenix

(42:17):
every month, and that was one ofway most first cities, right?
They have them out there, yeah,and at first they had a very
small radius area. It's only inthe last three or four months
that they've really expanded itso I can go. I My mom lives out
in Phoenix. I've moved fromthere a couple years ago, and I
go back every month to so thatshe doesn't feel like I
abandoned her. And right? Andshe has a funny list, and I help

(42:39):
her with her computer all day.And, you know, nice. So anyway,
so I go out every month, and nowWaymo will pick you up right in
that ride share, just in thelast two months, right in the
ride share spot at the airportwith all the other ride shares.
And it'll drive up 30 mile 25miles north to my mom's place.
Won't get on the freeway yet. Iknow that's coming in another

(43:01):
month or two or three, when theyright, and then it'll get on the
freeway, so it takes a littlelonger. But who cares? I'm
sitting in the back playingchess or, you know, reading or
doing work or whatever.
Now have gained back that time?Yeah, I've gained back the
time. I've reduced the risk,right? And it costs half of what
Uber costs to get to the sameplace, because there's no driver
and and they don't havecompetition yet. So now imagine,

(43:24):
in three years and and Uberscome out with their self driving
car, Tesla's come out with theirwhole self driving thing, and
there's a Tesla service to takeyour car, if you're not using
it, and turn it into a text youmake money from it, right? And
all of a sudden there's thismassive pool of self driving

(43:45):
taxis, right? Available andthere. And this will happen
really pretty fast and and thenyou Why would you buy another
car at that point? Right?Because the cost of owning a car
is dramatically higher thanthis, especially when the price
pressures and bring it down evento work every day, it'll be
nothing compared to what itcosts you to drive your car to

(44:06):
work every day. Yeah, plus youhave to park, and then you have
to maintain your car. Andthere'll be and there'll be such
a saturation of these thatwithin a minute, or maybe two
minutes, you'll always be ableto get one, unlike with Uber or
Lyft, right? Because the numberof cars available for this would
be just massive, and you don'thave to worry about idle time

(44:27):
with the car because, because ofa person's time, right? So,
right? Oh, there'll be a pointand all of a sudden, it's just
the only way you do it. Nobodywill park in a shopping center
anymore. Nobody will take theircar and park at a football game
or baseball game or all theseparking structures disappear. In

(44:50):
fact, how the start buildinghomes won't include driveways,
potentially or or, or thethey'll have these little half
size car garages. Mm. Becauseyou don't need them anymore. You
know, you have it for your it'sgoing to happen really fast.
Once it happens, it may be threeor two or three or four years
people are you get start to geta little saturation, then it's

(45:12):
going to be and it's just gonnabe, like, almost overnight,
yeah,
and I mean, I see a little bitof that now, like I I see
there's some entrepreneurs. Theyrun like a rental car service,
especially with Teslas out of,you know, the Burlington area.
And basically they use theparking in those business parks,

(45:34):
because there's half the half.More than half of those parking
lots are not even full. Soyou've got all this empty space.
It's free parking. You justbring a car there for a few
hours. It gets rented out, thenit's going out. So to your
point, it doesn't really make alot of sense for someone who's
going to own these taxis, theseauto taxis, to put them in a

(45:55):
place they should be drivingaround, getting ready for the
next
the next, yeah, and there'll becharging stations that are that
plug themselves in, so it won'trequire anybody, you know, just
kind of like your Roomba, right?Yeah. After it's finished
cleaning the house, it plugsitself in and charges. So it'll
be like that with the cars,right? And crazy, yeah, yeah. So

(46:18):
it'll happen fast. And I've seendisruption like this. I've
witnessed this before. It's not,you know, Musk didn't say, I'm
going to get rid of all theparking garages in the world,
right? That wasn't his plan.That's not the goal, right,
right? That wasn't his goal,right? In fact, he probably
never, well, he might havethought of it, because he truly
does have freaky visioncapability, but you probably,

(46:43):
but he may never have thoughtabout that, right? That's just
not important to him. But justlike the guys who started Uber,
never thought they woulddisplace an entire class of
vehicle in Southeast Asia, youknow, seven years after they
started Uber, that was thestory. So I was going to India
all three times a year. So from2010 until roughly, until the

(47:09):
pandemic started. And I went on15 trips, basically building my
teams out in India and workingwith them and cultivating them.
And, you know, because reallybuilding a really exceptional
team, no matter where they arein the world, you have to have
presence there, yeah, so thatthey under, you know and and be

(47:29):
working with them, and developpersonal relationships,
especially if you want to createan exceptional kind of result
from the way that the teamswork. So anyway, so I used to go
out on right then from 2010 to2015 and no matter when I would
land in Mumbai, and it would bean, originally a five and a half

(47:49):
hour drive out to Pune, or lateron, with as they improve the
roads, that went down to liketwo hours so. But it was just
buried in in Mumbai and in Pune,not in between, surrounded by
these three wheeled motorizedrickshaws, right? They come Tuk
Tuks or autos, depending onwhere in the world, Mark and you

(48:10):
look like, if you're in atraffic jam, and you look up,
that's all you see, is theseblack roofs of all these. It
just, it's like, phenomenal, thequantity like millions of them.
So from 2015 2016 I took, like,a 15 month hiatus traveling to
India. I don't remember whatexactly why, but then I went

(48:30):
back, and as we're drivingthrough Mumbai, I thought,
there's something reallydifferent. I don't know what it
is. And then we got into Pune, Iwas like, where are all the Tuk
Tuks? There was, it was like, 10of those to one car before, and
now it was 10 cars to one tookcook. I mean, this is like
millions of the whole industryjust shut down. And I couldn't I

(48:52):
thought, it must be likegovernment regulation, you know,
two stroke motors bad for theair. I couldn't think of what
else would cause it. So I got tothe office and my director. Name
is shelaka, director of our, ourteam there. I said, What
happened all the tuk tuk? Shesaid, Well, Uber has come to
India. I said, Why does thatmatter? She said, Well, they
want
to prove them. Oh, wiped

(49:13):
out a whole industry of car, avehicle, and all the drivers
that owned them. Unbelievable. Icouldn't I was, my mouth was
hanging open. She told me that.I was like, okay, that's what
that is, true, disruption, andit's not. And if you look back
at all these totally disruptivethings, it's rarely something
the founder ever thought wouldhappen, or it was just that

(49:35):
consequence of the adoption andgrowth, right? Of a phenomenal
adoption and growth, anyway.Long story short, the next a
year later, all they were allback because it was a new
service competing with Uber inIndia, I think, called Walla or
Walla, and they approved them.And so, of course, Uber, to
compete with them, had toapprove them. And so, like

(49:57):
overnight, they're all back.
Yeah. It's amazing how quicklyit can shift when there's a
demand. Yeah, yeah.
Anyway. And so this definitelymakes you think, yeah, oh yeah,
yeah. So say goodbye to parkinggarages. What's all what's think
about, like New York City orChicago or Boston, that's gonna

(50:18):
happen to all that real estate?Yep,
I can't tell you how many emptybuildings there are in Boston
that were built in the last 10years. Yeah,
office space, right? Not
even just office space, actualapartments that are priced out
of what anyone could Yeah, wehave a lot of financial and we

(50:39):
have a lot of medical, andthere's a whole section of
Boston that's literally justmedical personnel.
So New York City's got the sameproblem. Yeah,
yeah. And, you know, I don'tknow who owns these. They keep
building it, but I think theykeep building them exactly,
exactly, yeah,
that I don't get. But the part,you know, that's, that's some

(50:59):
kind of weird, sheltering money,you know, or laundering money,
kind of thing I have, you know,for or, yeah, from over I have
no idea what that's all about,but, but the parking garage one,
I can see that one coming.
That one makes a lot of sense,because it's very specific. It's
not like, like you need like,you need to bring in cars. And

(51:20):
if there aren't cars to get, getbrought in, like, it's just,
like,
there's right, they're justpulling up to drop people off,
right, yeah. So you'll havethese big, like, eight lane, you
know, drop off areas right,right at games and and you'll
just have these self drivingcars all lined up to drop people
off, yeah, you know, six deep orwhatever, and move people in and

(51:42):
out really fast. And plus selfdriving cars, if you have enough
of them in the road, thentraffic is going to move along
smoother, because they'll, youknow, start to swarm as they get
smarter.
Hey, being in Boston, I can tellyou, anything to improve the
traffic would be amazing. I do.I do think that in the beginning
here, there really should be alane dedicated to these robot

(52:05):
taxis, yeah, because, I mean, Ithink it's better from testing,
you know, why not test it in acontrolled environment, instead
of adding the variable of humanson the road? I also, there's
also, and I wanted to ask youabout this. And I think that
prediction with the parkinggarage is going to happen
regardless, but I've seen itdefinitely with certain types of

(52:27):
tech. And I think specificallythere was a phone that was based
entirely around the fact that itwould pause or stop doing it,
stop doing something. When youlooked at it, and I remember
there was just a straight up,people just said, Nope, we don't
like that. And now it's in everydevice, and face ID has that as
a feature. But right, it wassomething that people

(52:50):
would stop doing. It would stopit would stop doing what I
think the whole commercial wasaround the fact that, like, you
were watching a video, and thenyou look away and it would pause
automatically
for you, oh, because it, becauseit was paying attention to your
fit, Right exactly. And I thinkthat people didn't trust that,
and trust it Right exactly. AndI think,
I think it was a feature thathad been cool, but it wasn't

(53:14):
something that people wanted.Well,
you also worry, is it gonna stopfor the wrong reasons a bunch,
and then you're gonna getfrustrated watching the video,
and so 100% I'd say, kind of aninteresting idea. That's one
that would require a lot ofmarket testing to see if it's
that's a that is a great ideathat may or may not be solving a
problem, right? So, right? Havepeople complain because they're

(53:38):
missing video when they lookaway. Is it really a problem
that needs to be solved. What'sthe perceived impact because of
that, right? How much is itreally costing people, enough
that they're gonna so everythinggoes down to do. I feel
threatened by it, because if Ido, then I'm gonna do something
to mitigate it. Even vacations,you can see this beautiful, you

(53:58):
know, this beautiful beach sceneand the nets appealing to you
because you're afraid you'regoing to miss out in your life
if you don't do that right, andyou feel threatened by living
that life, that person that youdon't want to be right. This all
comes that I didn't used tothink that way, but the more I
the you know, the more I seethings over years, I realize it
all comes down to fear,uncertainty and doubt. That's

(54:20):
what drives people to dosomething,
and it can also kill greatcompanies or great ideas before
they even start right? Like, ifpeople just straight up boycott
it, it doesn't like that. Thatphone just they
do, you know, if they do, itmight have been solving a
problem, but it was creating abigger one that and they just,
they weren't talking to thepeople who have the problem,

(54:40):
right, right?
And that's where I feel likeright now AI is sitting. I think
there's a really interestingmarketing technique, and I know
it's gonna change. I know themarketing will change and the AI
will get better, but it feelslike we're in this very strange
era of AI, where the perceived.Use of it, like all the

(55:02):
marketing is around the factthat it is going to help you
cheat in this like you can't,you don't remember the person's
name, but you're going to lookat, you know, you're going to
use AI to remember that person'sname, and then they think that
you're a nice person, that youremember their name, you know,
like it's all about,
if you think of that ascheating, but I don't think of

(55:22):
it as cheating. I think of it ashelping to facilitate the
relationship. So it's aninteresting perspective
difference, right? Yeah, I mayjust be bad at remembering
names, and somebody else justhas a photographic memory around
it. It's not because they'renicer or any better. They just
they have the AI built intotheir head already for that,
like, 100% I'll
give you an example for me. LikeI am terrible at structured

(55:44):
writing. It's all it's why Iwent into engineering, because
writing was a horribleexperience for me. I use chat,
G, P, T, right now to help giveme a structure. And then I fill
the ideas. I'm great with ideas,right? Get out ideas all day,
right?
And then it can articulate it ina really well, well thought out,
a well articulated manner,right?
Yeah, and so I totally agreewith you. So is

(56:06):
that cheating, not necessarily,if it's expressed right now, and
I
think we're in this immaturephase of of of how AI is being
shown to people. We just, we didan episode where I just talked
about the AI tools we use forthe podcast, gave numbers on how
it's helped my life, because Ifeel like there is a lot of AI

(56:26):
that is more of a perceived fearfor folks than it is for a value
for folks. And I don't thinkthat the marketing is really
hitting on the true value forpeople. So I feel like it's this
interaction.
What marketing? Are you? Whatmarketing are you referring a
lot
of the new iPhone, like the newcell phone, AI stuff, because

(56:50):
that's what most people areseeing right now. Well, I've got
AI, I've
got the new iPhone, iPhone 16.Where's my my camera?
How do you like it?
What's awesome? Because thecamera is so damn good, right?
And it has, finally, it has realzoom. So if, like, I'm looking
way back out there and I can'tmake something out, I can go

(57:12):
into the phone and zoom like 25times in now, I can see the
thing clearly, and it's likehalf the size of the half my
screen. It's, it's amazing,that's why, and in light and
everything, that was my bigreason. I had a I upgrade every
three or four years. So I had a12. Both my wife and I did. So
we both went to the 16 and TMobile. Had a killer deal. They

(57:32):
gave us $850 credit on our 12,as long as they were in good
condition. I don't so hers waslike, anyway, so we upgraded.
And I also wanted we I needed todo it because I wanted to be
able to use the AI features theycome out. I have not taken
advantage of any of the AIfeatures. They are just clunky

(57:55):
and stupid. So far as I cantell, I just wanted to stick an
emoji into a text message. Itsays I can't do that, or it says
I don't have that contact. And,you know, it's like, what?
Well, this is, this is thediscrepancy. Well, it's like,
stupid,
right? That would be awesome,you know, stick an emoji, a cute
emoji, of heart me holdinghearts, you know, for my wife in

(58:16):
this text, right? And it has noidea what I'm asking. It, what?
Yeah. And I have the betafeatures on so it's got the
ability to do the image creationfor me now, but I can't get it
to create an image when I ask itto, I mean, you know, it's
really, at least the one I'masking it to anyway, so it's
just not done, right? Yeah? Ithas a long way to go, as far as

(58:39):
I'm concerned, yes,
and that's really, I guess whatI'm trying to touch at is that
the Yeah, the reality versus theperception is, is very stark,
and it is. It feels like thewhole reason to buy the phone is
the AI, and, like you just said,it's, it's really not the reason
that you use, no, it's, it'sanything else, yeah. I

(59:00):
mean, I was curious about the AIabout it for because I wanted
the camera and the speed, youknow, it's a much faster, more
powerful phone and memory andyou know, all that. But yeah,
the otherwise, it feels verymuch like my 12 Did you know
it's not all that different,except when I'm using the camera
or I noticed things that werekind of starting to lag a little

(59:22):
bit. Nothing lags anymore, butthat's the phones. But when
you're talking about AI in othercontexts, it's like, way beyond
the promise, and the speed atwhich it's evolving is, like,
unbelievable. So
give me some examples. Yeah, Iwant to, I want to hear from
your perspective. All right.Well,
so when? Okay, so I kind of likechat GPT before chat G, P, T

(59:44):
came out, like eight monthsbefore I read an article in
Scientific America that, and thearticle was about open AI, and
it was written by open AI,right? And I read that, I was
like mind blown. So AI reallycan do this now. I went right
on. Open. Ai started playingaround with it, and it could do
a lot in terms of just languagerelated stuff, which it never

(01:00:07):
could do that before, right?Anything I ever saw was like,
sounded like it was coming froma computer, not from Right,
right? And then so I startedusing it for different things,
but not a lot, because it waskind of limited. And then chat
GPT comes out, and then all of asudden, it's way more
consumable, but still kind oflimited it initially, and it was
making lots of mistakes earlyon, right?

(01:00:28):
Yeah, was this GPT two or,
I don't remember what thenumbers were, but it was that.
It was two years ago, December,right, when it came out, or
January. It's come a long waysince then. Yeah, it was, yeah,
right, but, but it was cool, andit was starting to get better. I
mean, the hallucination went onfor a while that seems to be
pretty much mostly over now, notover over, but it's, it's the

(01:00:52):
rare case where I find ithallucinating in a clear way
anymore, but, but it started toget good enough. Where I was
talking about it with my wife,like, four or five months after
it came out, right after wemoved to San Diego, and we're
standing in the backyard. Shewants to do square foot
gardening, right? She wants togrow all our vegetables. We're

(01:01:14):
not vegetarians, but we eat alot of vegetables. And she says,
how many do you think we need?And I said, I don't know. This
is something, you know, we couldprobably ask chat GPT. Her
sister calls right then, and shegets on a 10 minute call with
her sister talking about themove and all that. And so I
started conversation. Whileshe's doing that, I start a
conversation with chat GPT, andI say, we live in Vista

(01:01:34):
California. So considering themicro climate here, we want to
do square foot gardening. Mywife and I are what our age was,
and that we want to grow all ourown vegetables, but we're not
vegetarian, so keep that inmind. How many? How many four by
four beds do we need? And so ittold us how many. And then I
said, Great, okay, so come upwith a with what should I plant

(01:01:57):
in each of these beds so thatthe things we're planting are
companion plants with eachother, because you don't put the
same kinds of vegetables all thesame, you know, you can't just
put any vegetable together withother vegetables in a bed,
because they don't like eachother, right, right? So some are
compatible and some are not. Sothen it tells me what, and I
said, For what season, do thisas a for all the seasons, and

(01:02:21):
consider succession planting,which means that if I plant, if
I plant, let's say lettuce,right after I take onions out,
or whatever it is, it's going tohate the soil. Let us, you know,
certain things like they followother things and other and
certain things hate to followother things. So you have the
succession rotation fordifferent seasons, and so it

(01:02:42):
gave me that. Then I said, Okay,you know, so companion plants in
the bed and succession planning.I said, What companion flowers
do I need? Because you want toput flowers in with it so that
it will draw the insects away.Certain flowers draw certain
insects that might likecucumbers, but they like this
flower better. And I can't tellyou which, because I'm not a
gardener, but I know that thisis, yeah, right, but you put the

(01:03:05):
wrong flower in there and it'snot and it's just going to bring
the wrong kind of insect that'sgoing to go after your cucumbers
instead, right? So then it gaveme that, and I said, Okay, and
this was right when chat GPTcame out with the ability to
create tables. So I said, okay,so create a, create a schedule
by bed by season for all ofthis, right? And so it did a

(01:03:29):
beautiful job. You know? I thinkwe're eight beds. And so season
one, season two, season three,season four, right? What do we
plant in each bed, and whatflowers and any and so she gets
off the phone, right? So about10 minute call minutes, right?
Yeah, I was just took 10minutes. And I said, What do you
think of this? She went, Oh, shecouldn't even voice that, right,

(01:03:55):
because she was thinking, it'sgoing to take days, right, of
research. And she looked at it,and she knows enough about
guardian to know that this wasall valid. And she was going, Oh
my God, because I've beentelling her about this, and I
got to show it to her. She said,How did you do that? So I went
through the whole thing withher. That's how I think of
working with these tools. Andthe yes and the promise they

(01:04:15):
have at the very simplest level,right, is that they can
literally take what would takethings like work, like
developing code or writing testscripts or creating user stories
for structuring a plan orbuilding software so we are or
doing project estimates. Sowe're building a tool, an AI

(01:04:37):
tool, right now, because we havea very detailed process when we
do when we estimate a project,way more than most companies,
because it's expensive andcostly and time consuming, but
we like to be very accurate withour estimates, which means
really understanding a fullmodule breakdown of the project
that as far as we know therequirements so that we're gonna
have to build, and then puttingdetailed estimates to each one

(01:04:59):
of those. Things so and we'vehad, we get compliments all the
time on the level of detail inour estimates and how
sophisticated they are, but it'sa costly process, most costly
process in my company, that asfar as like cost of sale kind of
thing, right? So we're buildingan AI model that will mimic our
process and actually willprobably improve it as well, on

(01:05:21):
in terms of how we do theseestimates that we're going to
turn into a SaaS product, I'mstarting to develop a lot of
referral partners in my in thebusiness, oh, I'm actually using
some of the code generationtools to build a referral portal
that my that my referralpartners can log into and record

(01:05:41):
referrals and see what thestatus of those referrals are,
if they've actually contractedand then starting to accrue, you
know, referral fees for the next12 months and how much they've
accrued. And, you know, toreally kind of bring a
professional kind of level tothe whole referral partner
engagement relationship, then,where that whole, remember, I

(01:06:06):
talked about the the in launch,first doing the matrixes for the
impact, right? That's a verytedious process. It's really
important, but it's butfounders, you know, founders,
unless they're really committedto doing this right. A lot of
them get fatigued and havestruggled to complete it all,

(01:06:28):
and they want to rush into thedevelopment. So we're building
an AI model to do this for them,and probably do it better than
what a founder can do on theirown, even really dedicated ones,
because it's drawing from somuch information, and can
identify particular trends thatmay not be obvious to a founder
as they're figuring this sort ofthing out, and it's going to be

(01:06:49):
an AI model that we're building.So these are some of the things
that we're doing, especially inthe CO generation area. We're
continuing, evaluatingeverything on the market and
adopting and using whatever wecan to try to accelerate the
delivery. Because what we wantto do is we want to get to a
point where somebody comes to usand maybe that that launch first

(01:07:13):
high fidelity prototype isactually an application that
we've built that we can rewriteany part of it very fast, right,
in two months or three monthswith the design, and the AI is
even helping us with the designand the workflows. I mean, I
look at it like this, AI is thismassive rogue wave, right?
You're sitting there at thebeach and you're trying to surf

(01:07:35):
three foot waves because you'renot that good, and all of a
sudden there's 100 wave that'scoming along, yeah? And it's
like you have one of twochoices. You either, you know,
paddle out to that wave and getinto the curl and get on the tip
of your board and try to stayahead of it, or you just let it
sweep over you and then figureout afterwards, assuming you
survive, you know what lifelooks like afterwards. So we're

(01:07:57):
one of those companies that'slike on the very tip of that
surfboard, trying to get aheadof it. You know, stay ahead and
shoot the girl basically.
Yeah, I hear you. I mean, I Ifeel the same way about it. I
wanted to use it and learn itbecause, I mean, as a site, as a

(01:08:17):
person of science, I don't likethings that become magic. I
mean, I think it's beautifulengineering and beautiful
science when it's like magic.But I think for humanity, it's
terrible if we get in asituation where we don't
understand why that's happening.Well,
we we are already there. Yeah. Imean, even the people creating
open AI and creating all theselarge language models don't

(01:08:39):
fully understand how it's ableto do what it does? Yeah? No,
it's they understand the sciencebehind it and why they went this
way, but the fact that it'scoming up with natural sounding
language and even compassionatekind of approach, right? All
these things, they don't reallyunderstand how it's able to do
that.
Yeah? And that's and to me,like, that's okay. I I think,

(01:09:05):
like, like, you, I want to stayon top of it and and ride that
wave, instead of just waitingfor it to crash. Because it does
feel like a huge moment, likethe tractor in farming, and how
that changed human life, right?Like, everything was farming,
and then the tractor made it sothat, you know, we live in a
world now where nobody knows howto, you know, grow anything for

(01:09:26):
the most part as a populace, youknow. And so we're entering this
new stage where nobody knows howAI is going to change life
around us. So it's, it is. It isa very interesting time. One
thing I wanted to ask you about,oh,
please go ahead, Microsoft,Microsoft, CEO, such a, such a

(01:09:50):
Nadella. I think he, I think itwas him that said this and and
he said it best. Somebody said,this is like the. As important
an invention or a disrupt, achange in the world, like the
light bulb. And he said, No,it's like fire when fire was
found. He said, It's changeseverything, yeah. And I think

(01:10:15):
that's right. I mean in terms ofthe impact and the speed at
which it's going to changeeverything? Yeah, yeah. And
I think if even someone justhears about it in passing, like,
if this the conversation is thefirst real in depth look at AI
that they've even practical AIthat they've they're already
ahead than somebody who hasn'teven thought about it. Yeah.

(01:10:38):
I know people are afraid of it'sgoing to get become conscious
and like, destroy the world,right? And because it wants to
protect the world from humans,that's one way to look at it.
Another way to look at it bytrope, yeah, right, exactly. So
the other way to look at it is,it's going to enable innovation
at such an incredible pace,innovations that we never even

(01:11:00):
thought of that the need for theneed for earning a living that
goes away because food is isjust available and and produce
better than it would be if youWere still doing classical
techniques and and energy is nowfreely available, and it starts

(01:11:25):
to be like a Star Trek kind ofeconomy where you don't, you
don't focus on earning money,you focus on improving the
world, and, you know, andbuilding better communities,
because you don't have to do youdon't have to earn a living
anymore. Yeah, that's right,that's looking out, obviously,
away, certainly
one, one direction. You know, Idefinitely do share that. And

(01:11:48):
then I think that that is theclosest thing to a Star Trek
life that we could have very,very soon,
money becomes unimportant. Yeah,
the concern I have is the goldrush that we've seen, you know,
almost the, you know, financialtakeover of open AI that seems
to have settled itself. There isa lot of greed and money at this

(01:12:12):
Oh, yeah. So
the thing that I never go away,yeah, no,
it won't. It won't. The thingthat concerns me is how it's
trained. I think that's thething I'm really focused on
right now. And it sounds likeinterested to hear how you're
developing yours internally,because I think there is an

(01:12:33):
artistry to that, you know, likehaving worked with large amounts
of data and done testing serieswhere you're trying to root out
the reality of the situation, ina lot of ways, the physics of
the situation, right? Like whatis actually happening in
reality? Yeah, that requires areally good test. It requires
very good data, and then itrequires you to analyze that

(01:12:56):
data. And I The concern I havewith this, and kind of why I was
poking at the the marketing ofAI now is, yeah, is it actually
developing something that'suseful is, is the data set that
these AIs that are being poppedup? Is it providing and being
trained on actual usefulinformation? Or is it, you know,

(01:13:17):
we've already seen AI,
you mean, like, is it going outand reading fake news and this
and assimilating that as part ofits reality? Sure.
Yeah, you are what you eat, kindof mentality, you know, yeah,
right. Is it being given gooddata and then at what point is,
you know, we, we talked to aguest a little while, uh, last

(01:13:38):
year about, you know, hisconcept was, you know, once we
get to a point where these,these AIs, are generating
content, which they already are,and they literally fill up the
tubes of the internet, is thereenough internet to maintain
what's going to be produced on?And then how do we build the
infrastructure from there?Right? Like, yeah. And that's a

(01:13:58):
problem, the problem, right,yeah. And the amount of power it
takes to run these languagemodels is a whole other thing.
When you compare it to theamount of power that the actual
brain uses to do the same thing,yeah, right there. That's,
that's things that will besolved with money and brilliant
people. But yeah, right.
I mean, yeah, it's gonna help.An AI will help solve those
things a lot faster, too. Yeah?

(01:14:19):
So going back to it. I thinkit's, it's a very, I'm not gonna
have any control over it, but,but what comes next? And I think
it's gonna be a battle of theAIS is, is gonna be, you know,
which AI is really providing a,a real value, and which ones,
you know, we saw, I think theearly, what was it? Early Gemini

(01:14:40):
really had some kinks in it, buteveryone was trying to, it was
the arms race of AI. Everyonewas trying to get them out as
fast as possible. So, I guessdoes that? Is that a concern?
You know,
it's funny, because they theywanted to just get something out
in front of the public, becausethey had nothing in front of the
public. But Google, it. Googleis the one that created this

(01:15:01):
whole concept of large languagemodels, right, right? And they,
their technology is, like,phenomenal, you know, the stuff
that they've got now. And, youknow, it's a, it's a horse race,
right? As far as who's you'regoing to use and, and frankly, I
use several of them, right? Andvery often I'll use several for
the same thing to get differentperspectives or take the

(01:15:24):
results.
Oops, sorry, I lost your audio.
Oh, shoot, it's back. Okay. Igotta stop moving my arms get
excited. I'm fixing that today.This is driving me nuts. I am
which one? You know, it's just aconstant. You know, there's

(01:15:47):
different, like, for example,notebook. Lm, I'm sure you've
played with that notebook. No, Ihaven't. No, oh my god. So
Google came out with this sixweeks ago, I think. No, book.
Lm, it's a, it's very cool.Basically, it's a, it's a
notebook kind of environment youjust got, you know, if you've
got a Google account, you justgo up there where you upload

(01:16:07):
documents and and web links andthings like that, and adjust all
this. And then you can talk toit and ask it questions about
all this information assimilatedand have it produce stuff for
it. But, and then you can say,Okay, go ahead and generate an
audio for me, and it will createa podcast episode with a man and
a woman talking to each otherabout this content and really

(01:16:28):
insightful kind of edgy kind ofway. And you would not if you
didn't know it, you would nothave any idea this is AI crazy
be creating these two voices andhaving this conversation. I had
to do it on my website, just tohave it and and they created a
podcast episode assessing thiscompany techies, they seem as so

(01:16:49):
funny. And they talked about howmuch we've accomplished such a
short amount of time. And theysaid, you know, that's like dog
years and technology, you know,and it has all these little I
was and then it also pointed outsome things in terms of how
we're messaging, that it wasthey were talking between each
other, and I realized that's notthe messaging we want, but they

(01:17:10):
really distilled it down to whatit was a great feedback tool.
But so there were a lot, therewere a bunch of podcasts that
came out that were using thispodcast with the great episodes
where the episode would be thesetwo AIS talking to each other
for 1015, minutes and but smartpeople, I think, are not doing

(01:17:32):
that, because they know thateventually, very quickly, the
algorithms will already reallyding their podcast from the for
those episodes. But, oh yeah,you should try. It's amazing.
Yeah, to do it, yeah, so,
but it's a great way. So that'sa kind of a rag model. So rag is

(01:17:53):
rich augmented. I forget whatthe G stands for, rich
augmented. This is like. Sobasically, you're taking a large
language model, and you'recreating a separate vector,
database vector. You're creatinga you're training a separate
database with specific context,right? But maybe it's

(01:18:14):
healthcare, maybe it's whatever,maybe it's project estimation
capabilities, right? And then,and you're in for using that to
inform the large language modelof the context. So this is, this
is how most tool mod moderntools that go beyond just using
the large language models andtheir APIs. This is what they're

(01:18:35):
doing when they're buildingsomething specific for an
industry that's smart, you know,smarter than just the smartness
of a large language model. SoRight? And large, and LLM is an
example of being able to do thatyourself, just for your own
personal use. And you can createas many notebooks as you want,
each one you know, on a specificgroup of content that you're

(01:18:55):
kind of collecting, yeah, so,but we do it a little bit more
intentionally, right when we'rebuilding, when we're training.
So, so what we want anything webuild to do is to to become very
specific in terms of itscapability and what it's focused
on, but it's still going tospeak in a very natural way,
because it's using the LLM tofor the for all the

(01:19:18):
communication, and plus, it hasthat vast amount of knowledge
from the LLM that helps itinform and go beyond just what
we've trained it with. And youknow, but what we've trained it
is like, this is the context.This is what we absolutely know.
And you have a lot of otherinformation you can add to that,
but not come up with somethingcompletely different or unique,

(01:19:39):
because this is we know asexpert information, right? And
so that's, that's how webasically building these models.
So, but the what was I going tosay, the real power that's
coming out is these agenticworkflows, and we've been
working with that for. Monthsagentic workflows. Means I've

(01:20:02):
got multiple AI agents, eachwith a different specialty, and
they're working together toproduce something, oh, wow. So
like, for example, and there'ssome tools starting to come out
with these built into it. SoI've got, I'm building code,
right? So I'm a project manager,so I'm organizing the tasks. I

(01:20:25):
am a developer, which is aseparate agent that is an expert
at writing code in there inthese particular technology
stacks. I'm a tester, so I knowhow to write test plans. I know
how to accurately test thesoftware, and the three of them
are communicating as you'redeveloping with the idea that

(01:20:46):
each one is really good at whatit's doing, and the QA knows how
to identify problems and feedthat back to the developer,
which then goes back in andfixes it until the QA says,
Okay, this is working the waythat it's expected. And we've
got a business analyst agentthat documented what the
requirements are based on thehigh level requirements. So they
created user stories, and right?So this is where it's headed,

(01:21:09):
and this is the big thing openAI's been getting working on.
And if you look at the 01preview, you see some of that
agentic workflow, because haveyou played with the
o1 preview? No, it's the firsttime I've heard of it. Yeah.
Okay.
It's been out for a few weeks.You have to have a paid account

(01:21:31):
to see, okay, right? So open.Ai, it's available. It's open.
Ai, yeah, okay, I'll check thatout. I do have a page, yeah? And
you just ask, you just go to themodels, and one of them is the
O, O, 1o 4o i Preview, I think Isaw it. I didn't know what it
was, okay. Oh, yeah. That's themost recent, the most powerful

(01:21:51):
model. And when you ask it aquestion, it stops and it says,
thinking of the answer,considering, you know, you can
see different agents takingdifferent perspectives in terms
of what it's doing, and it'stelling you what it's doing
while it's working through theanswer consider, you know,
writing out, writing therequirements, considering the

(01:22:16):
language and context, right? Andso you can see it's looking at
it from different perspectives,because it has different agents.
It's dynamically creating theseagents who are expert at
specific tasks because it knowsit needs it at this point in
this process. Yeah, and that'sthe beginning of that agentic
workflow being built right intoopen. AI, the next really major

(01:22:38):
release is going to like, like,blow this open. Wow.
So it's like having a littleteam basically, yeah, which,
which, you know, to kind of tiethis off into a bow here,
because it's been freakingamazing. I think about AI in the
future of human space travel alot, and how, you know, I think

(01:23:02):
the last time this came up, Ihad kind of thought of it as,
you know, at some point, theseAIs will join, at least in a
space mission perspective,you'll have all these different
AIS combined to create, youknow, let's just say the
computer, like we had in StarTrek, and it's able to do a
bunch of different tasks. Butwith sydentic workflow, you

(01:23:23):
could have different subsystemsof the whole thing. And
basically what you have inMission Control, on board the
the actual spacecraft itself,helping out in all the problems
that they're going to deal with,traveling deep space and all I
mean, AI is going to be socritical, because you just reach

(01:23:44):
the limit of light and space andtime, right, like, right, having
to wait for the communication tohappen to then do the analysis
to figure out the problem. Ifyou can do it on board that
spacecraft,
that's eventually they'll figureout, you know, ways of getting
around the light limitation, thelight speed limitation,

(01:24:05):
there's got to be, oh yeah. AndI'm sure AI is going to help.
Yeah, exactly. Yeah, yeah. So Idid want to give you an example
of workflow, a simple example ofworkflow automation, because
this is, I did this four yearsago, not AI related, but I'm
about to do another one for thissame guy, and it will take me

(01:24:26):
hours, probably to do this forhim, which took me months last
time. So this is a guy runs areally big business network,
yeah, and he has a Tuesdaynetworking meeting. He has like
1500 members, but about 70people at his networking meeting
every Tuesday morning, and halfof them are online because of

(01:24:46):
the pandemic, and the other halfare at the his facility, right.
And then he goes back and forthbetween announcing, right. So
they have big screens up intheir facility so everybody can
see everybody. And then peoplethat are now at their facility
walk up to. The front of theroom, and there's a camera right
there. So he turns that on, soit goes back and forth. He's got
this all wired really well. He'sbeen doing this for 12 years.

(01:25:07):
Before it was a zoom thing, azoom hybrid thing. And when it
was all at a local spot, and 30%of the people there are guests
every week, so and so, everyMonday. So four years I've been
a member for six years or so.About four and a half years ago,
I thought, you know, I've gotthis idea for an app that would

(01:25:29):
be way smarter and make it somuch easier to focus on all the
pitches, because 70 people talkfor a minute each one, right?
And so there's a few people I goI am really need talk that
person by the time I've heard 10more pictures, I don't remember
who they are, why I need to talkto them, or even how to find
them, because I don't didn'ttake a picture of their face, or
I didn't write the right note,right? So I came up with this
app idea that would that wouldmake it really easy to capture

(01:25:53):
and connect with the people thatmatter to me by just basically
tapping my phone while they weretalking. Okay? I thought at the
school, I thought he was goingto love this. So I sat down with
him, and I started talking aboutthis. He says, You know, I don't
want this to add more to myworkload. I said, No, this is
going to make your workloadeasier. He says, I just can't
handle anything else on myMondays, anything that gets in
the way of me preparing forTuesday. And I went, Okay, tell

(01:26:15):
me about your Mondays. That'swhat started this whole thing.
Because I realized, okay, he'sgot a much bigger problem than I
have. You know, I was trying tomake the meetings more, even
more meaningful, and he's like,he spends all day. He spends
787, hours, seven to eight hoursevery Monday, getting ready for
his Tuesday meeting. And some ofthat is interruptions, because
he's constantly beinginterrupted, but probably five

(01:26:36):
and a half six hours of solidwork that he's doing, getting
everything organized, pulling inall the descriptions of
everybody, because he announceseverybody, and he announces them
like they're the most importantperson in their business and
their industry is really good atthis. Yeah, right. And, and he
orchestrates it, who comes afterwho? Because he doesn't want
three people in the similarindustry to all speak one after
the other, right? So, so hestaggers them in a really smart

(01:26:59):
way. Plus he has people that arepart of his organization stand
up to at different times,talking about, come on, join our
connectors committee, you know,things like that at different
points anyway. Yeah, exactly.And, and so he's pulling in
people's name. He got to makesure he's got all their contact
information. He pulls theirdescription from a different

(01:27:19):
place, all this, and it's a lot,right? And it's for 70 different
people, and he's organizing it.So I said, Okay, tell me about
your Mondays. What do you do?And he's always telling me, I
said, Okay, how about if Iautomate that for he says, You
can't automate this. It's justso specific to what I do. And I
said, Well, let's see what I betI can. Let's try, right? Yeah,
yeah. And I ended up doing likethat, right? I didn't use my
team for this, which would haveprobably got a lot faster, but I

(01:27:41):
wanted to just have every oncein I want to indulge and write
some code myself, but which Irarely get to do anymore. So So
and we just use Google Sheets,and I wrote Google App Script
logic right to to orchestratethis forum, and did integrations
with this ticketing system andstuff like that. So he told me
how it runs two weeks, threeweeks later I come back because
I was doing this one that I hada little extra time I come back.

(01:28:04):
He said, Okay, I did it. Let'stake and he says, Great, let's
show me. And as soon as I startshowing it says, you know,
that's just not gonna workbecause it doesn't handle this,
this or this. I said, Oh, well,you didn't tell me about this,
this, right, right, but, but Ido understand you need this,
this and this. So give me acouple weeks and I'll come back
and do this, this and this. Hesays, Okay. He says, I can see
it's handling that now, but wejust never talked about this.

(01:28:28):
And I said, No, we never didtalk about that. So this
happened, like over a period ofthree months, and sort of
dripping out the requirements,because people that have these
processes that they're reallycomfortable with don't even know
what they do right, until youtry to model that thing for
them, and then all the nuancescome out, right? Yeah, which is
just how it always is anyway. Sowe finally got, three months

(01:28:51):
later, got all the nuances.There were a lot of them, all
the nuances handled right. Andso we ran it one week in
parallel with him doing itmanually, and it came out with
the right results. And then thefollowing week, he ran it for
the first time on its own, andhe's like, This is amazing. I
was able to do my Monday in 10minutes. I love it. That's
amazing. Six hours, right? Everysingle Monday for then for the

(01:29:14):
last six years, think how manyhours that is, right? Yeah. And
He only takes off his Tuesdayswhen there's a major holiday of
some kind, right? So like, fouror five times a year he doesn't
do it, and twice a year he doeswhat he call these big meetings
where people fly in from aroundthe country, and billionaires
come in for this, and 50panelists get up and they all

(01:29:35):
speak, right? And but it takeshim three days to prepare for
that, and now it takes him,like, an hour and a half or two
hours. So you know, he is mybiggest fan, right? And he's
been using this now every day.In fact, we just talked a few
weeks ago. I was telling him,I'm starting a podcast. I wanted
his his feedback on that, abouthow I'm approaching it. He liked

(01:29:56):
to approach it. He said, Oh, I'mkind of afraid to tell you
about. This, but I have like,six or seven people from our
network who are doing podcasts,and I rebroadcast their podcast,
because he has a big podcastingnetwork, and he has a radio
show, he says, but it's reallytime consuming, so that's why
I'm kind of leery. And I said,Okay, this sounds familiar, and

(01:30:18):
it's I anyway, so we're meetingin January. I'm gonna automate
that whole flow for him, forpublishing all the stuff that he
does and automating it. And Ican probably do it. And if you
know, won't take me three monthsthis time, other than getting
the information from him on thenuances, yeah, probably do it.
You know, most of the work nowusing AI in a few hours to build
this, maybe, maybe a coupledays. If it just, if I just keep

(01:30:41):
dragging all the nuances out ofthem, right? So in a week or
two, between the time it takesto get stuff and then show them
what I've done, and meetings,instead of three months will be
done, and now he'll be able tojust, you know, people, people
will fill out the form when theygot a new episode, and
everything will just beautomated. And he'll get some
kind of report to show what itdid every time somebody's

(01:31:02):
published so he can confirm thatit's all working properly. Wow.
So that's AI, right? Otherwise,it would be three months again,
yeah, the same kind of thing.But the impact the workflow
automation can have is, likedramatic. Well,
I can tell you from this podcastitself in the last two years.
You know, I think a lot of theso brainstorming, I definitely

(01:31:26):
work with chat GPT to just giveme some structures that I can
spend way less energy trying toget this brain to do something
it doesn't like to do, like youwere talking about, like,
that's, that's huge. And then,you know, clip generation,
promoting online, all thatstuff, transcripts, time stamps,
that were taking me just way toomuch time. I mean, I've probably

(01:31:49):
saved eight to 20 hours in myweek of doing this and it, it
frees me up. I like thinking ofit like going from the offensive
line to being the quarterback.You don't have to be every
single position. You canactually focus more on the
strategy and the thinking of,you know, what do I actually
want to do here? What? What morecan I? Can I add here so that I

(01:32:13):
am big believer in that, andthat's from my perspective,
you're doing exactly the rightway. You're basically being
channeling Michael Dell,because, you know, when he was
building Dell computers, and hedid an interview with the Forbes
article, and they asked whatthis is, you know how it was so
successful? He says, I justfocus on whatever the current
bottleneck is, and I just workon that. I put throw the

(01:32:37):
resources to open up thebottleneck, because as soon as I
do, everything will flow intowhatever the next bottleneck is,
which wasn't a bottleneck untilwe and that's how we scale,
because that's the next thingthat needs to be focused. I just
look to see where the bottleneckstarts to happen and all the
resources go there. So that'sexactly what you're doing,
right?
Cool, yeah, that's it. That'sthe bottleneck. Was me, right?

(01:32:59):
Yeah, which it often is,honestly, and I think that's,
it's a, it's kind of fascinatingto think how much of this AI
adoption is going to be also a,like, a cathartic thing for a
lot of people, if they, if theyjump into it right of like,
like, that whole experience youjust described, of that guy
going through that his he didn'teven consider it as Something

(01:33:21):
that could be done. And as soonas it did, it was like, take
more, take more out of it.Because, if I don't, wasn't that
the idea that it could beautomated. But he was like, he
couldn't even really conceive ofautomated. Come on, look at all
the things I have to do. Whatare you talking about? Totally.
But he is like, so happy. I lovewhen I when it started to work.

(01:33:42):
I freaking love that. So David,let's, let's close this out a
little bit. This has been anamazing conversation back,
because I know AI is going tochange in the next week.
Nevermind. Yeah, tomorrow.Crazy. So let's close out. Tell
us, where can folks reach outfor techies, and if they, if

(01:34:04):
they got they want to reach outto you, contact you. Okay,
so tech is spelled T, E, k, y,z. You can reach me. You can go
to my website, techies.com youcan email me. David@techies.com
pretty easy to remember, and youcan also listen to my pod, my

(01:34:26):
new podcast, which I justrecently launched, called
Scaling smarter. And it's atscaling smarter.net, so name,
yeah, yeah, I would. I can'ttake credit for picking that
name. I had a really stupidname, and the guy helping me
with this podcast said this is amuch better name. And I was
like, Okay, fine, it's a muchbetter name.

(01:34:48):
I love it. I love it in the lastword for the folks especially, I
love the AI and just havingsoftware and computers help us
do. More in life and gettingtowards that Star Trek future.
Yeah. Any last thoughts forpeople?
Yeah, think of AI as your bestfriend that just happens to know

(01:35:10):
everything about everything. Sotalk to it like it's a friend,
and even say, ask it nicely.Yes. Please help me do this. By
the way, I I did that from thevery beginning. I don't know
why. It just was natural for me,because I felt like I was
getting so much value out of it.Was like I was grateful. And
then when it gave me somethingthat helped me, said, that's
that, you know, it was like,that's correct, or if it wasn't,

(01:35:32):
I said, No, that's not right.And I had explained why consider
it from this perspective. Andthen when it is said, That's
right, thank you, you know,because I want to give it
feedback, yeah, I get betterresponse from Ai than other
people that are just Curt andthat think of it as a tool,
right? And treat it like a tool.And I think that's because, just
because it's learned the world,right? And if somebody is asking

(01:35:54):
nicely, that means they want amore thoughtful, considered
response than somebody who'sjust being short and Curt, and
they just and so it doesn't tryas hard, which is weird, right?
But, but, and plus, eventuallyit might become conscious. I
want it to remember how nice Iwas
100% I live by that rule.

(01:36:16):
Remembers these other theseother people, they were nice to
me, but I'm going to protectthem. They're the first to go.
That's funny. Yeah,
that is hilarious. David, thankyou so much for joining us in
Today In Space, and best luckwith your podcast.
Really enjoyed the conversation.Very funny, too. Thank you.

(01:36:39):
All right, folks. Spread loveand spread science. Thank you
for joining us for anotherepisode of Today In Space. Be
well and we'll see you on thenext episode. See ya.

All Episodes

AI & Business w/ David Hirschfeld, CEO of Tekyz | Tech Talk

Episode Transcript

Popular Podcasts

NFL Daily with Gregg Rosenthal

On Purpose with Jay Shetty

Dateline NBC

.css-15opob5{left:0;position:absolute;top:0.8rem;} All Episodes

.css-14f5ked{margin:0;word-break:break-word;display:-webkit-box;-webkit-box-orient:vertical;box-orient:vertical;-webkit-line-clamp:2;overflow:hidden;}AI & Business w/ David Hirschfeld, CEO of Tekyz | Tech Talk