00:00:00
Jensen hang co-founder and CEO of Nvidia
00:00:03
the company that is at the heart of AI
00:00:05
Revolution recently made a comment about
00:00:07
what could should learn in the future
00:00:09
and let me spoil it to you it's not
00:00:11
computer science or programming I'm
00:00:13
going to say something and it it's going
00:00:15
to sound completely
00:00:17
opposite um of what people feel you
00:00:19
probably recall uh over the course of
00:00:21
the last 10 years 15 years um almost
00:00:24
everybody who sits on a stage like this
00:00:26
would tell you it is vital that your
00:00:28
children learn computer science
00:00:31
um everybody should learn how to program
00:00:33
and in fact it's almost exactly the
00:00:35
opposite it is our job to create
00:00:38
Computing technology such that nobody
00:00:41
has to program you up and that the
00:00:43
programming language is human at s Ms a
00:00:46
full service agency by a co-founder
00:00:49
we're in a close relationship with
00:00:50
Nvidia and later in this video I will
00:00:52
include a clip of what Jenson said
00:00:54
people should study which is also in
00:00:57
alignment why we expanding our team with
00:00:59
people like Amy zolotov who comes from
00:01:01
Healthcare collectively we trained and
00:01:03
taught close to 20,000 people at this
00:01:05
point from people at all ages and walks
00:01:08
of life who subscribe to my channel to
00:01:11
the amazing teams at we companies like
00:01:12
General electronics and craft Hinds plus
00:01:15
we've been developing AI educational
00:01:17
material both for universities like
00:01:19
Stanford or yiva all the way to
00:01:21
developing courses on up liit I think we
00:01:23
shows that there is Raising awareness
00:01:26
about prompt engineering and the impact
00:01:28
that you can have in your organiz ation
00:01:30
or your team if you just know better how
00:01:33
to work with AI tools and I understand
00:01:35
that when a girl on the internet says
00:01:37
that prompting is the new programming
00:01:39
language that's one thing but I'm really
00:01:41
happy to see that more people kind of
00:01:43
receive a message than it is NVIDIA CEO
00:01:47
on a stage at the world government
00:01:48
Summit saying it out loud everybody in
00:01:51
the world is now a programmer this is
00:01:54
the miracle this is the miracle of
00:01:56
artificial intelligence for the very
00:01:57
first time we have closed the Gap the
00:02:00
technology divide has been completely
00:02:03
closed and this the reason why so many
00:02:05
people can engage artificial
00:02:06
intelligence it is the reason why every
00:02:08
single government every single
00:02:10
industrial conference every single
00:02:12
company is talking about artificial
00:02:14
intelligence today because for the very
00:02:16
first time you can imagine everybody in
00:02:19
your company being a technologist this
00:02:22
is a tremendous time for uh all of you
00:02:25
to realize that the technology divide
00:02:27
has been closed I hope people and
00:02:30
businesses will take it as a wakeup call
00:02:32
and start looking at this quite
00:02:34
seriously also to realize that to engage
00:02:37
AI is a lot easier now than at any time
00:02:40
in the history of computing it is vital
00:02:42
that we upskill everyone as for 90% of
00:02:45
people prompting is going to just become
00:02:47
easier and easier and there is a good
00:02:49
reason for that it's because of smart AI
00:02:51
engineers and prompt Engineers working
00:02:53
in the background optimizing that chat
00:02:55
interface so that a few vague words can
00:02:58
be taken processed by way bigger and
00:03:00
more extensive prompt or the whole
00:03:02
separate AI model by the way all that
00:03:04
optimization gets combined into the
00:03:06
final prompt and sent to the AI model
00:03:08
therefore you get with amazing outcomes
00:03:10
like Del tree or co-pilot it starts with
00:03:13
a prompt from you in an app co-pilot
00:03:15
pre-processes The Prompt through an
00:03:17
approach called grounding put simply
00:03:19
grounding improves the quality of the
00:03:21
prompt so you get answers that are
00:03:23
relevant and actionable one of the most
00:03:25
important parts of grounding is making a
00:03:27
call to the Microsoft graph to retrieve
00:03:29
your business content and context
00:03:31
co-pilot combines this user data from
00:03:33
the graph with other inputs to improve
00:03:36
the prompt it then sends that modified
00:03:38
prompt to the llm copilot takes the
00:03:41
response from the llm and postprocesses
00:03:44
it this postprocessing includes
00:03:46
additional grounding calls to the graph
00:03:48
responsible AI checks security
00:03:51
compliance and privacy reviews and
00:03:53
command generation finally co-pilot
00:03:56
sends a response to the user and
00:03:58
commands back to the apps it's awesome
00:04:00
if you're not specific and you are just
00:04:02
really exploring but it also takes away
00:04:04
some controls because sometimes these
00:04:06
chat Bots get so optimized that it gets
00:04:09
horribly hilarious Gemini from Google
00:04:12
went viral for absolutely wrong reasons
00:04:14
refusing to generate white people and
00:04:16
over optimizing for inclusivity and
00:04:19
could say stereotypes but kind of
00:04:21
phenomenal is but it became
00:04:23
stereotypical I'm genuinely curious what
00:04:25
impact it will have that Nvidia CEO uh
00:04:28
had this hot take but but if you listen
00:04:30
between the lines even Sam ultman hinted
00:04:32
at this future I think coding will still
00:04:34
be important in the future it's just
00:04:36
going to change a little bit or a lot
00:04:38
one of the predictions I'm most
00:04:39
confident about the future is it's just
00:04:41
going to happen much faster than the
00:04:42
past did and that you know young people
00:04:45
are always the best at adapting to that
00:04:47
but now it's going to be extra valuable
00:04:49
I think I agree with that framework
00:04:50
incredibly much and I think anytime you
00:04:52
can choose a skill or a class or an
00:04:54
activity to do that you feel aligns with
00:04:57
the learning how to adapt to new
00:04:58
technology pick it for sure but on
00:05:00
specifics I still think that software is
00:05:03
actually a really important skill to
00:05:05
build and some of it is it's not going
00:05:07
to be about the mechanics it's not going
00:05:09
to be about the specifics of the
00:05:10
programming language it's not going to
00:05:11
be about those details but fundamentally
00:05:13
getting computers to do what you intend
00:05:16
like that's the world we're heading to
00:05:17
and so understanding the low-level
00:05:19
details of how it works um that's a
00:05:21
starting point and you know we do see
00:05:22
people who have never coded before who
00:05:24
have built some who built some of the
00:05:26
best AI applications and so that's
00:05:28
actually new right that you can actually
00:05:29
get computers to do things without
00:05:31
having to do it in this like contort
00:05:32
yourself to the machine as it is right
00:05:34
now anyone can work with AI anybody can
00:05:37
instruct AI can you imagine that a
00:05:39
couple years ago this ability to
00:05:40
interact with AI models was available
00:05:42
only to researchers and actual
00:05:44
developers and just because it's
00:05:46
available to everybody and it's getting
00:05:48
easier doesn't mean that people use it
00:05:49
in the best way possible or the most
00:05:51
effectively the way I like to think
00:05:53
about the stuff is like a spectrum on
00:05:55
one side you have like oh my God prompt
00:05:57
engineering is the only thing you need
00:05:59
to know on whatever spectrum is like oh
00:06:01
it doesn't matter programming is going
00:06:02
to be still there and it's a solid skill
00:06:05
and in our team the way we like to look
00:06:07
at things is a bit more leveled yes
00:06:09
prompt engineering is important that
00:06:11
also is going to change if you think
00:06:13
about in a term of 10 years if we all
00:06:15
walk with the neuralink chips in our
00:06:17
head this morning Elon Musk announcing
00:06:19
that for the first time his company
00:06:21
neuralink has implanted a brain ship in
00:06:24
a human everything kind of changed again
00:06:26
and after the statement from Jensen on
00:06:28
the whole of Spectrum there's going to
00:06:30
be a group of people who are like okay
00:06:32
programmers they are not valued they are
00:06:34
not important don't study computer
00:06:36
science and stuff like that and I think
00:06:38
that's also uh wrong because programming
00:06:40
in computer science as Hands-On skill is
00:06:43
one thing but what it teaches people is
00:06:45
problem solving and that is a hard skill
00:06:48
which won't get absolute anytime soon
00:06:50
and I want to say never but we also said
00:06:52
that creativity is probably going to be
00:06:54
the last um affected by AI which turned
00:06:57
out completely opposite so anyway anyway
00:07:00
um taking a part being in the middle
00:07:01
educate yourself about prompt
00:07:03
engineering how to use AI how it works
00:07:05
what are ethical uses of that but if
00:07:06
you're able to also learn about computer
00:07:08
science and how computers work in
00:07:10
general develop deeper problem solving
00:07:12
and critical thinking really skills with
00:07:14
programming that is not going to hurt
00:07:16
you like I myself I talk about prompt
00:07:18
engineering but I also decided to take
00:07:20
cs50 course to kind of bridge the gap I
00:07:23
have when it comes down to programming
00:07:25
funny thing that a month ago before
00:07:27
Jensen's video came out I made a video
00:07:29
where said that domain expertise is
00:07:31
going to be extremely important when
00:07:33
using AI tools and here's one universal
00:07:35
truth when it comes to prompting you as
00:07:38
a user have to have domain expertise and
00:07:40
use specific language to condition the
00:07:43
AI model and steer its responses to the
00:07:45
right direction because if you know your
00:07:47
stuff very well not only you will be
00:07:48
able to explain it well and make clear
00:07:50
instructions but you will also able to
00:07:52
detect hallucinations the countries the
00:07:54
people that understand how to solve a
00:07:57
domain problem in digital bio biology or
00:08:00
in education of young people or in
00:08:03
manufacturing or in farming those people
00:08:06
who understand domain expertise now can
00:08:09
utilize technology that is readily
00:08:12
available to you you now have a computer
00:08:14
that will do what you tell it to do to
00:08:16
help automate your work to amplify your
00:08:18
productivity to make you more efficient
00:08:21
and so I think that this is just a
00:08:22
tremendous time and going back to that
00:08:24
question what kids should study in the
00:08:26
future this is what Jensen said if I
00:08:28
were starting all over again um I would
00:08:31
realize uh one thing that one of the
00:08:34
most complex fields of science is the
00:08:37
understanding of biology human biology
00:08:39
not only is it complicated because it's
00:08:41
so diverse so complicated so hard to
00:08:44
understand living and breathing it is
00:08:45
also incredibly impactful complicated
00:08:48
technology complicated science
00:08:50
incredibly impactful and remember we
00:08:52
call this field life sciences and we
00:08:54
call drug Discovery Discovery as if you
00:08:57
wander around the universe and I sudden
00:08:59
hey look what I discovered nobody in
00:09:01
computer science nobody in computers and
00:09:04
nobody in the traditional industries
00:09:06
that are very large today nobody says
00:09:08
car Discovery we don't say computer
00:09:11
Discovery we don't say software
00:09:13
Discovery we don't go home and say hey
00:09:14
honey look what I found today this piece
00:09:16
of software we call it engineering and
00:09:19
every single year our computer science
00:09:21
our software becomes better and better
00:09:23
than the than the year before every
00:09:25
single year our chips get better every
00:09:26
single year our infrastructure gets
00:09:28
better however however Life Sciences is
00:09:30
sporadic if I were to do it over again
00:09:32
right now I would realize that the
00:09:34
technology to turn life science to life
00:09:36
engineering is upon us and that digital
00:09:39
biology will be a field of engineering
00:09:41
not a field of science it will continue
00:09:44
to have science of course but not a
00:09:46
field just of Science in the future and
00:09:48
so I I hope that that this is going to
00:09:51
start a whole generation the Albert
00:09:53
Einstein College of Medicine will be
00:09:56
tuition
00:09:58
free of people who enjoy working with
00:10:01
proteins and chemicals and and enzymes
00:10:03
and um materials and and they're
00:10:05
engineering these amazing things that
00:10:07
are more energy efficient that are
00:10:09
lighter weight that are stronger that
00:10:11
are more sustainable all of these
00:10:13
inventions in the future are going to be
00:10:15
part of engineering not scientific
00:10:18
discovery but if studying is not
00:10:19
applicable to you and you want to get
00:10:21
into AI or learn how to use it my
00:10:23
recommendation to you is to start
00:10:25
exploring together with other people so
00:10:27
you can exchange DS and tools there are
00:10:30
so many resources online that you can
00:10:33
definitely dive into uh there's also a
00:10:35
lot of noise and what I learned from
00:10:36
mounder Joe who is Educator from his
00:10:38
background that there is a lot of
00:10:40
science how humans learn and this is why
00:10:42
we really enjoy the partnership with the
00:10:43
pits because one thing it's life uh you
00:10:46
have accountability you have to do real
00:10:48
projects for your work for your business
00:10:50
for yourself um and then there is this
00:10:52
the whole pure Community as well make
00:10:54
sure to check out the prit courses we
00:10:57
have for image for data analytics
00:10:59
um we have ai inbt for everyone and also
00:11:02
the advanced course yeah and we are
00:11:04
working on many many more normally I
00:11:05
would tell you to go watch another video
00:11:07
of mine so that algorithm is happy but
00:11:09
actually this time we really want to ask
00:11:11
you a question how do you feel about the
00:11:13
whole programming prompt engineering
00:11:15
where you stand what has been your
00:11:16
experience and journey in 2023 and what
00:11:19
is your approach in 2024 and as always
00:11:22
thank you for watching and sticking
00:11:23
around I've been going through some
00:11:25
personal stuff but I want to make a
00:11:28
promise that I will be up uploading
00:11:29
videos more consistently so if you are
00:11:31
new on this channel it would mean a lot
00:11:33
if you would subscribe
