00:00:00
if you just look at the rate of change
00:00:02
and the rate of new models coming out
00:00:03
and new techniques yes it doesn't seem
00:00:06
like we've hit the scaling wall and
00:00:07
again if you if you believe that which I
00:00:09
happen to believe and you just project
00:00:11
forwards this stuff is going to get
00:00:13
better and better
00:00:15
[Music]
00:00:17
yeah hello this is Dalton plus Michael
00:00:20
and today we're going to talk about AI
00:00:22
it's officially blowing our minds um AI
00:00:25
is working what have you seen yeah if
00:00:28
you were watching this video you
00:00:30
probably also agree that AI is working
00:00:32
you have the knowledge but are you
00:00:35
leveraging this knowledge like if you
00:00:36
are living in the future yourself what
00:00:39
should you be doing differently knowing
00:00:41
what's coming right how can you prepare
00:00:43
for the future what's interesting is
00:00:45
that we've seen stuff like this happen
00:00:49
before um you know we lived through the
00:00:52
first um static web pages we live
00:00:54
through the dynamic web apps we live
00:00:56
through smartphones happening we live
00:00:58
through Bitcoin happening
00:01:00
knowing that this is not a fad but a
00:01:03
trend early is maybe one of the most
00:01:06
strategic things you can know so like if
00:01:09
you're watching this video
00:01:12
you have a gold mind literal gold mind
00:01:17
um of insight on how the world's going
00:01:19
to change and if you sit on
00:01:23
information you're not going to be able
00:01:25
to take advantage of it if you change
00:01:26
something right now or more than one
00:01:29
thing right now
00:01:30
um you might look back 10 years from now
00:01:32
and say oh wow like like my life worked
00:01:36
out like that was really really great so
00:01:39
um let's go through some of the things
00:01:41
you should do just to kind of give you
00:01:43
some mental context we can look back in
00:01:46
time a little bit and we could say
00:01:48
imagine that you were using the first
00:01:50
web browsers and you were using the
00:01:52
first websites and you like oh this is a
00:01:55
thing imagine most people didn't have
00:01:58
the web no there were there were some
00:02:00
folks that were very early on that stuff
00:02:02
and they had a lot of extra leverage a
00:02:05
lot of Leverage right imagine you were
00:02:07
one of the first users of Gmail and you
00:02:09
were like oh my God like a really
00:02:12
responsive Dynamic web app that's just
00:02:15
like
00:02:15
instantly like oh crap imagine you one
00:02:21
of the first users of iPhone yep imagine
00:02:24
you were one of the people who saw
00:02:28
Bitcoin come out it's first year yep
00:02:31
like these are the kinds of
00:02:34
like like there were ways at the time
00:02:36
with the benefit of hindsight to
00:02:38
leverage that information yes to
00:02:40
profoundly improve your lives or impact
00:02:42
your lives right you were early and you
00:02:44
were like oh this seems like a good
00:02:45
thing and you and you acted on it it can
00:02:48
completely change your life so you are
00:02:50
early let's talk about what you should
00:02:53
do yep right so first if you were ever
00:02:57
interested in starting a company I can't
00:03:00
really imagine a better time yep um I
00:03:04
would argue that like most of the best
00:03:07
companies were created after this kind
00:03:11
of tech sea change yep seeing the SE
00:03:13
change and then dreaming what we could
00:03:15
do with it so like if you are ever
00:03:17
inclined to start a company and people
00:03:19
are saying to you oh this like AI thing
00:03:21
is overhyped or like oh there's too many
00:03:23
it's
00:03:24
like no again you can look to history um
00:03:28
you needed there to be enough users and
00:03:30
the tech to get good enough uh for
00:03:32
internet startups yes um to to really
00:03:35
take off if you started an internet
00:03:36
startup in 94 95 that would have been
00:03:38
too early I think yes but it's much
00:03:39
better doing it later um same with all
00:03:42
these other platform shifts is there's
00:03:43
now enough of an install base there's
00:03:45
now the models are actually good enough
00:03:48
the tools are good enough the INF is
00:03:49
good enough yeah it still is like a good
00:03:51
time yeah you can build applications
00:03:55
that were unimaginable and like you know
00:03:58
we won't go into too many details but
00:03:59
like we literally seen YC companies be
00:04:01
transformed in the last 6 months like
00:04:04
businesses that weren't working like
00:04:07
started working started working and
00:04:10
margins that were underwater are now
00:04:12
suddenly positive customers that were
00:04:14
Loop War are now like you're selling me
00:04:18
magic I would like to buy more of that
00:04:20
so um don't let the haters like like
00:04:25
it's like like imagine the haters that
00:04:27
were like don't buy Bitcoin it's a fat
00:04:28
it's like going to go to zero it's like
00:04:32
like n you know like those were
00:04:33
expensive that was expensive advice to
00:04:35
follow exactly I remember having a
00:04:37
friend being like Blackberry will never
00:04:38
lose to iPhones like oh like you know
00:04:41
like dud like don't listen to those
00:04:44
people um let's say uh and you know of
00:04:47
course we should throw in a plug right
00:04:49
YC applications are open you know
00:04:51
they're four batches a year so um we can
00:04:54
certainly help if you're interested um
00:04:56
let's say you don't want to start a
00:04:57
company probably should be con
00:04:59
considering where you want to work
00:05:01
knowing this knowledge I would think
00:05:03
really hard about where you choose to
00:05:05
work um you want to work somewhere that
00:05:09
is adopting these tools yes and is going
00:05:12
to benefit from the adoption of the
00:05:13
tools and where you could be surrounded
00:05:15
by people that are um very smart at
00:05:19
adopting AI tools that you can learn
00:05:21
from versus the folks that are just
00:05:23
hoping this all goes away or that the AI
00:05:26
good luck with that there a lot of
00:05:27
people that are hop the internet would
00:05:28
go away yeah yeah
00:05:30
a lot easier for business if eCommerce
00:05:32
didn't exist a lot of newspapers that
00:05:34
was the plan you know um anyway you also
00:05:37
want to work somewhere where you could
00:05:38
be mentored by the folks that are really
00:05:40
excellent right if you're just imagining
00:05:42
this is going to be very important
00:05:44
skills in the future and for how we live
00:05:45
our lives um working in a place where
00:05:48
the tools are great and the people using
00:05:49
the tools are also great and can teach
00:05:51
you about them that is really smart
00:05:54
right again imagine it's 2001 and you
00:05:57
could work at a place like Google or
00:05:59
another internet native company that was
00:06:01
on The Cutting Edge of yes building
00:06:03
internet software
00:06:05
versus some insurance company or
00:06:07
something you know even if it's a
00:06:09
programming job there's very difference
00:06:10
on how much you could um learn and
00:06:12
upskill yourself between those types of
00:06:13
jobs and I would argue that you know
00:06:16
startups are adopting these tools faster
00:06:19
than big companies so there's a really
00:06:22
interesting question of like are you
00:06:23
going to get more experience being a
00:06:25
first 10 or first 20 employee at a
00:06:28
startup that's using AI versus working
00:06:30
at that big company where using AI is a
00:06:33
political decision and a PR decision and
00:06:35
a bu bureaucratic nightmare like be be
00:06:39
careful because like the learning that's
00:06:41
happening in startups right now is far
00:06:43
outstripping what's happening in most
00:06:44
large companies um of course they're
00:06:47
exceptions but in most let's talking
00:06:49
about skills right I remember there was
00:06:51
this always this funny joke um within
00:06:54
the first year of the App Store coming
00:06:55
out where some big company was hiring
00:06:57
iOS developers and they were like like
00:07:00
oh you need to have five years of
00:07:01
experience building iOS apps and
00:07:03
everyone was like came out a year
00:07:05
ago like but what's really cool at this
00:07:09
moment is that like if you start
00:07:11
learning today you're almost at Ground
00:07:13
Zero yep and like there's so much
00:07:16
leverage right High School College post
00:07:19
colle um you know you pointed this out
00:07:21
when we were talking before um dropping
00:07:24
everything to become an iOS Developer in
00:07:26
2009 2010 was a strong career smart idea
00:07:30
like if you saw the iPhone come out
00:07:31
you're like oh this is cool I'm going to
00:07:33
go to like boot camp and teach myself to
00:07:36
code and become an iOS programmer yeah
00:07:38
that was a really good career move or if
00:07:40
you were a web developer inside of a
00:07:42
company and they were spinning up a new
00:07:44
iOS team Jo like being SM that was a
00:07:47
very smart move or just building an app
00:07:50
on the side like like a lot of people
00:07:52
built side projects yes and that's how
00:07:54
they learned to become iOS developers
00:07:56
and they benefit in tremendous ways and
00:07:59
I argue
00:08:00
at any stage like I said pre-ol College
00:08:03
post colle you can upskill yourself like
00:08:06
I think another thing that's really cool
00:08:08
is again this is meta you're probably
00:08:10
watching this on YouTube there are
00:08:12
amazing YouTube videos of explaining how
00:08:14
all the stuff worked and in Prior major
00:08:17
platform shifts you had to read books or
00:08:20
you know know people or go to the right
00:08:22
school or something like that you can go
00:08:25
watch people building in public uh
00:08:27
building models explaining how
00:08:29
Transformers work you name it there's a
00:08:31
lot of stuff that's open source and so
00:08:33
if you are technical or or aspiring to
00:08:37
be technical which you should be go
00:08:39
watch the videos and learn how all of
00:08:41
this stuff works from scratch it will
00:08:43
profoundly help you for the rest of your
00:08:46
life I can't think of a downside to
00:08:48
knowing how all this stuff works no
00:08:49
downside seriously and I think what's so
00:08:51
funny is like all of the trends Tech
00:08:53
trends that we mentioned or break those
00:08:54
we mentioned before all of them happen
00:08:56
in public this isn't happening in some
00:08:58
back room you know have access to Y the
00:09:01
issue is people being skeptical it's not
00:09:03
that it's not available for you to learn
00:09:06
so no I think that's that's incredibly
00:09:10
preent yeah there's no Gatekeepers to
00:09:11
learn this stuff that's what I'm trying
00:09:13
to say the only gatekeeper is
00:09:14
ambivalence yes or or cynicism or
00:09:17
cynicism or hearing you know negativity
00:09:21
and buying into it and negative people
00:09:24
don't tend to get lucky nope you know
00:09:27
they don't tend to have um really great
00:09:29
stuff happened to them on accident so
00:09:32
you know maybe lean into positivity on
00:09:35
this one yes we're giving you the hint I
00:09:37
would say next become a user right like
00:09:41
many of these products are available and
00:09:44
you know I remember previous waves of
00:09:47
Technology where people were like oh
00:09:51
well it's not ready yet oh I'll wait
00:09:53
like oh I don't need to switch and I
00:09:57
think this is different you should be
00:09:59
actively looking for better products
00:10:02
like I mean we remember when when Google
00:10:04
came out and it was fundamentally a
00:10:06
better web search product and like there
00:10:09
are a whole bunch of people like oh web
00:10:10
search blah blah blah blah blah and it's
00:10:11
like no people had access to Google got
00:10:13
more done and so if you've been
00:10:17
waiting sign up and and it's worth the
00:10:20
money yep like it's worth paying money
00:10:22
for I mean I when I was um a Founder
00:10:24
working on consumer stuff I would
00:10:26
download every consumer app that came
00:10:28
out literally and I would study every
00:10:30
onboarding flow and I would study
00:10:32
everything about the service I would
00:10:34
study the viral loops and just
00:10:36
constantly be trying to challenge myself
00:10:38
to make sure I knew more than anyone
00:10:40
else about what the new products were
00:10:43
yeah and so and what they could do yeah
00:10:45
and so there's again I don't know why
00:10:46
you wouldn't be looking at every AI
00:10:48
coding tool that comes out or every
00:10:50
workflow tool that comes out and signing
00:10:52
up and trying to earnestly use them and
00:10:54
try to try to use them to get your job
00:10:56
done not just to research them as
00:10:58
startups yeah um be a real user be a
00:11:00
real user and you're going to learn
00:11:03
again you will be living in the future
00:11:04
yeah and and I would say this like there
00:11:07
is a bit of a learning curve like I
00:11:08
remember teaching my dad how to use
00:11:10
Google and it was
00:11:12
counterintuitive that every question he
00:11:15
had for me on the phone all I would do
00:11:18
is type that question in Google and
00:11:19
produce an answer like that wasn't you
00:11:21
know that wasn't intuitive to him there
00:11:23
are some learning CS like you know I was
00:11:25
I was having this like hilarious back
00:11:28
and forth with an Elm agent and I was
00:11:30
realizing I am using this incredibly
00:11:33
different than I use Google like the
00:11:35
things that I type how I respond to
00:11:37
answers like all of it is different and
00:11:40
I got to be comfortable just kind of
00:11:42
going it's a small learning curve with
00:11:44
it's a learning curve and then maybe
00:11:46
last but not least don't do nothing like
00:11:50
I think that's the like maybe the meta
00:11:52
point of this entire video change
00:11:56
something yeah in response to this yeah
00:11:58
you know something that other people
00:12:00
don't know yes capitalize on that
00:12:03
knowledge don't sit on it like it's the
00:12:06
knowledge isn't going to have any
00:12:07
interest you got to do something with
00:12:08
the knowledge and man the earlier you do
00:12:11
it the the more leverage you're going to
00:12:14
have now a lot of people talk about the
00:12:17
idea that maybe in this world being
00:12:21
technical being a programmer isn't
00:12:22
strategic what do you think about that
00:12:24
yeah I mean this is an opinion we we all
00:12:26
don't know but um I think all of the
00:12:29
tools getting better means that some
00:12:31
aspects of being a programmer um like
00:12:34
memor having all the apis memorized is
00:12:36
maybe okay that part is less valuable
00:12:38
but I would argue knowing how all of
00:12:41
this stuff works is going to be just
00:12:44
more and more of a premium in modern
00:12:46
society yes cuz someone has to fix this
00:12:48
stuff when it breaks and someone has to
00:12:50
understand yes what the heck the AI is
00:12:53
doing like to the extent um all of your
00:12:56
technical knowledge atrophies and you
00:12:57
have no idea how of this stuff works
00:12:59
it's just like a magic box no that is
00:13:01
not a great place to be and so in my
00:13:04
opinion really understanding how this
00:13:06
stuff works it a lowlevel will um
00:13:09
increase the value of you being a
00:13:11
technical person yes massive because no
00:13:13
one else will know how to debug it when
00:13:15
things go wrong yes and just because you
00:13:17
start using cursor to program does not
00:13:20
mean that uh people that have never
00:13:22
programmed before are going to be
00:13:23
programming just as well as you and so
00:13:25
again I know there's different opinions
00:13:26
maybe some people I've heard people
00:13:28
predicting seen the death of programming
00:13:30
as a career for many many many years our
00:13:32
whole lives before we before we were
00:13:35
born um again may maybe again maybe I'm
00:13:38
wrong uh I just I wouldn't make that bet
00:13:40
I would bet on being extremely technical
00:13:44
just having a premium that will increase
00:13:47
over time with the proliferation of
00:13:48
these tools and that has been again
00:13:50
backwards looking that has always been
00:13:51
the case yeah and if you're not
00:13:53
technical I can't imagine I mean it's
00:13:56
always been strategic to work with
00:13:58
someone who as a programmer it's still
00:14:01
incredibly strategic never has it been
00:14:03
more important to have a technical
00:14:04
co-founder um if you're not technical
00:14:07
last Point here right we are seeing YC
00:14:10
companies doing amazing things with the
00:14:13
current models we're kind of giving you
00:14:15
a straightforward observation not a
00:14:17
prediction but there is this question is
00:14:21
there another step change Improvement
00:14:23
coming and you know let's push the issue
00:14:25
of AGI aside I think there's a big gap
00:14:27
between step change Improvement and an
00:14:28
AI but
00:14:30
like what is your prediction How likely
00:14:33
do you think it is that we have another
00:14:34
step change that'll be here in the next
00:14:36
year or two years it
00:14:39
seems like we are not near the wall no
00:14:43
and if you just look at the rate of
00:14:44
change and the rate of new models coming
00:14:46
out and new techniques yes um it doesn't
00:14:50
seem like we've hit the scaling wall and
00:14:52
again if you if you believe that which I
00:14:53
happen to believe and you just project
00:14:55
forwards this stuff is going to get
00:14:58
better and better yeah and it already is
00:15:01
getting it already works yeah so yeah
00:15:05
hold on to your hat um yeah I think this
00:15:07
is one of those moments we've seen it
00:15:09
with internet adoption or smartphone
00:15:10
adoption where it's like when the
00:15:12
products get good they go through this
00:15:16
s-curve where they start
00:15:18
getting good and I think that um smart
00:15:21
Founders would assume that these
00:15:24
products or the tools they have access
00:15:26
to will get better over time and I I I
00:15:28
hate describing where that limit is I
00:15:30
think it's irrelevant really where that
00:15:31
limit is but like it's just an argument
00:15:34
for why it's even more strategic to get
00:15:35
involved today because like you're
00:15:38
riding on a wave and almost everyone we
00:15:40
know those who is successful found the
00:15:42
right wave and rode it and this is a
00:15:47
good one all right great chat sounds
00:15:49
good thanks
00:15:51
[Music]
