Hang tight while we fetch the video data and transcripts. This only takes a moment.
Connecting to YouTube player…
Fetching transcript data…
We’ll display the transcript, summary, and all view options as soon as everything loads.
Next steps
Loading transcript tools…
Developers Irrational Ai Fear in 2025 | Stefan Mischook | YouTubeToText
YouTube Transcript: Developers Irrational Ai Fear in 2025
Skip watching entire videos - get the full transcript, search for keywords, and copy with one click.
Share:
Video Transcript
Hey guys, how you doing? Uncle Steph
here. So, why shouldn't developers fear
AI? Why is we see a lot of fear? Now,
why is this fear irrational in the
medium and long term? Well, first of
all, AI based development though
productive and useful yada yada yada.
I've talked about this before. It's not
perfect. It's not trivial. You're not
going to have some random business
person, some HR department person
being able to develop
really good functional apps with AI
unless they put a lot of effort into it.
AI will make you more productive just
like React
made front-end web app development much
much easier. Right? VanillaJS,
it's capable, but if you want to do what
React does, it makes you, you know, 10
times more capable. AI is the same
thing. It speeds up the process. It
makes it easier, but it still requires a
lot of work and there's a lot of
planning and there's a lot of complexity
with AI development. Don't fool
yourself. this irrational fear in my
opinion about AI killing all jobs. A lot
of it's hype. A lot of it's because our
lizard brains, our lower brains, link
below, I have a whole course on it. It's
designed to overemphasize, to
artificially magnify
potential threats.
AI is perceived as a potential threat
because it does a lot of stuff, but it
makes a lot of mistake.
If you look at some of my previous
videos, you see people are now being
hired to clean up what is called AI slop
code. It's like messy code.
So, we've seen RAD protest prototyping
tools before. RAD applica rapid rapid
application development. RAD is short
for rapid application development.
We have seen over the years, over the
last couple decades, in fact, several
iterations of that, several versions of
tool sets that allow you to quickly
build applications or more quickly build
applications. Most of them fall short.
That's why you don't hear about them
much these days. Some of them are there.
typically uh when you have rad tools and
I consider AI uh a high level rad tool
they're good at very narrow scopes if
you go outside of that you're going to
have some problems and that's what we're
seeing with AI so AI don't get me wrong
it's much better than previous
iterations previous versions of these
type of tools but it's still that and
the fundamental problem with AI is that
it doesn't think you see AI doesn't
think it doesn't have that capacity to
go, hm, this here doesn't make any
sense. It would just go off on these
tangents. AI makes assumptions. It
doesn't have that ability to reason.
I know they have AIS where they call
reasoning AI and it looks like it might
be reasoning to a certain extent but
it's reasoning level is
at the end of the day based on my
experience working with the latest GPT
I train a an AI fitness coach to
automate some of the uh time consuming
processes in uh when you're when you're
getting healthy. Anyway, so I trained
this thing for about three and a half,
four months now.
And it's like training a small puppy
dog. It really is. You have to give it a
context. You have it to give it a solid
context, but even though you give it a
context, it'll still start dropping
turds all over the place. It's it's it's
it's it's like, why did you do that? You
know, it's like I told you not to. So,
you have to deal with what are called
edge cases. Basically, you give your AI,
you give it a a context, and it does a
bunch of stuff automatically, and you're
going, "Oh my god, I didn't have to
explicitly code this, right? I just
prompted it in."
But then it will just like it will go be
working pretty good and all will go off
on this tangent. You're like, "What the
hell is that?"
It's like having a self-driving car.
It's it's it's to the point where it's
taking it like they've been at it for
what Tesla's been at it for many years
where it's just now getting to the point
where it could drive itself. Driving is
not new as complex as you know building
apps and figuring out business
processes. One of the things I like to
say, if you're halfway there, you're
nowhere, right? So an AI could take you
quickly towards your goal in terms of
developing application, some sort of
workflow, no question. But it's still
you still need to intervene as a human.
You still need to intervene and deal
with stuff. So for example, going back
to my custom AI, I call Brad Fit the AI
fitness coach. Very powerful, very
useful. after three and a half, four
months of training and I still have to
deal with these random edge cases, these
where it just goes off in these little
tangents and uh just ruins it. Still
very useful. So the key to working with
AI from my experience anyway and
speaking to AI experts is you have to
keep the scope of what your chatbot
would do, what your agent would do, you
have to keep it very narrow.
The more narrow and specialized the
agent, more narrow and specialized the
uh chatbot, your AI, the better off it's
going to be, the less likely it's going
to do some crazy stupid stuff that's
going to ruin everything.
If you look at Open AI, so they came out
with GPT4 and I trained my my agent, not
my agent, I changed I trained my fitness
coach AI based on GPT4 because Grock
didn't have capability of a GPT had
gemini didn't have capability of a GPT
head and it's very good on four. I got
it working after three and a half months
or so. So then five comes out JPT5. Now,
the companies that own these AI, they
they're trying to tell us, "Oh my god,
AGI is around the corner. Oh my god,
Mimosa, they come out with five and it
breaks Brad fit.
It breaks months of work. Not horribly,
but it breaks it pretty good." And I'm
like, damn. So, what does that tell you?
That tells you that it's not a perfect process.
process.
several years ago when I was looking at
AI a little bit. So I picked up a book
on it. The chapter I was concerned about
was a chapter on training models
and the author of a book was apparently
a top AI guy. He was saying and I've
gotten confirmation from other people
saying that AI training model training
is more of an art than it is a science.
That's a big red flag. When people start
telling you something is an art,
more art than science, that tells you
the process is at best imperfect. So,
when you saw they went from GPT4 to five
and they busted poor Brad Fit, my AI
fitness coach, the world's greatest
living fitness coach. That's an AI. Um,
that tells you they got problems. Now,
not just me. I have a friend of mine and
some of his uh some of his u AI which
was based on GPT4 it got busted the hell
out of it in GPT5 and they had to roll
back and it's not so it's not just me
anyway. So there you go. It's complex.
We have different levels of complexity
uh or different types of complexity
shall I say with AI than you have with
software development.
And it's just another tool in your tool
set as a developer, just like React is,
just like native JavaScript is, just
like uh WordPress is. If you're a
developer, you may client wants a a vlog
or a blogging site, if you will. Makes
sense to use WordPress or some other
content management system,
etc., and so on. So, yeah, don't fear
the AI. It is not a threat to developers
in any way and in fact it's a huge
opportunity in my opinion. So if I were
you, you learn your foundations of code
which uh I would recommend the web stack
with that will give you the most
opportunity. But it doesn't really
matter if you do C or C++ or Java or you
do JavaScript, Python, PHP, my
recommendation, HTML, CSS. You learn
that, learn to work with AI, become an
AI expert. You got to look at, you know,
some experienced developers will say,
okay, I got these tools. One tool is
called JavaScript.
Another tool is called Python. Another
tool is called SQL.
Another tool is called, I don't know,
C#, whatever. Right? These are all tools.
tools.
React as a tool. WordPress is a tool. I
would argue Wix that uh build their
website is a tool. So are AIs. Gemini is
a tool. Grock is a tool. Anthropic code
is a tool. GPT is a tool. They're all
tools. You just got to learn to use
them. Back in the 90s when the web was
new and I would go in to see clients and
they would say what is a website Steph I
would explain to them what a website is
and I get some gigs when I get other g
anyway adopt my point is adopt the new
adopt the new so learn the web
start working with the AIs understand
what Gemini could do for you what GPT
could do for you what code could do for
you what Grock could do for you
understand the tools and uh also
understand the language
So for example with my fitness app AI
allows me to do things I could not do
with code and it's allowing me to do
things that I could do with code but it
does it much faster. But on the other
hand there are still things that I need
to know the web stack and coding and
onboarding and all this kind of stuff
and architecture
databases theor database theory. There's
these things they still need to know to
make this thing work with its ultimate
vision. With its ultimate vision. So
yeah, AI is not a threat to developers.
It's just a new tool set. Leverage it.
Those who embrace the change are going
to do fantastic.
I'm old compared to you probably. I'm
ancient. I'm old enough to know that you
should never use Ruby. And
I'm embracing the change. You should too.
too. [Music]
Click on any text or timestamp to jump to that moment in the video
Share:
Most transcripts ready in under 5 seconds
One-Click Copy125+ LanguagesSearch ContentJump to Timestamps
Paste YouTube URL
Enter any YouTube video link to get the full transcript
Transcript Extraction Form
Most transcripts ready in under 5 seconds
Get Our Chrome Extension
Get transcripts instantly without leaving YouTube. Install our Chrome extension for one-click access to any video's transcript directly on the watch page.