Specific AI Career Path Strategies
Key Points
- The usual “learn Python in 30 days” or “get a PhD/start a startup” advice is too generic, so you need concrete, role‑specific guidance to break into AI.
- By 2030 AI is projected to add 170 million jobs but also wipe out 92 million, meaning entry‑level positions that traditionally serve as footholds are disappearing.
- Research shows a noticeable drop in employment for young workers (22‑25) in AI‑exposed roles, highlighting the paradox of simultaneous opportunity and automation.
- AI should be viewed as a diverse “career maze” with many distinct paths—much like the various roles in the food industry—rather than a single monolithic field.
- The speaker analyzed 17 current AI career tracks, mapped hiring trends, and offered prompting strategies to help individuals identify and scale the specific AI role that fits their background.
Sections
- Navigating AI Career Paradoxes - The speaker critiques vague AI‑break‑in advice, highlights data showing both massive job creation and loss, and urges concrete, role‑specific strategies to seize emerging AI opportunities while avoiding entry‑level automation.
- Designing Predictive Skill Assessment Prompts - The speaker explains how they craft interview‑style prompts that anticipate meaningful signals—such as production deployments for ML engineers or editorial editing for prompt engineers—to accurately gauge candidates’ true AI and engineering proficiency.
- Evolving Product Management with AI - The speaker outlines a method for merging traditional product management competencies with AI expertise, using an LLM‑driven assessment to generate a personalized, dynamically updated learning roadmap.
- Upstream Prediction for Career Success - The speaker explains an AI‑career assessment that pre‑emptively flags common pitfalls and leverages early‑stage predictive insight—like meme‑driven probability forecasts—to guide users quickly toward viable roles, avoiding costly trial‑and‑error.
- Predictive AI Career Pathways - The speaker argues that the exploding AI job market demands a personalized, data‑driven approach where carefully crafted prompts guide LLMs to predict an individual’s optimal career move by accounting for market timing, transferable skills, and common hiring blind spots.
Full Transcript
# Specific AI Career Path Strategies **Source:** [https://www.youtube.com/watch?v=Z88hYmdxp00](https://www.youtube.com/watch?v=Z88hYmdxp00) **Duration:** 00:17:08 ## Summary - The usual “learn Python in 30 days” or “get a PhD/start a startup” advice is too generic, so you need concrete, role‑specific guidance to break into AI. - By 2030 AI is projected to add 170 million jobs but also wipe out 92 million, meaning entry‑level positions that traditionally serve as footholds are disappearing. - Research shows a noticeable drop in employment for young workers (22‑25) in AI‑exposed roles, highlighting the paradox of simultaneous opportunity and automation. - AI should be viewed as a diverse “career maze” with many distinct paths—much like the various roles in the food industry—rather than a single monolithic field. - The speaker analyzed 17 current AI career tracks, mapped hiring trends, and offered prompting strategies to help individuals identify and scale the specific AI role that fits their background. ## Sections - [00:00:00](https://www.youtube.com/watch?v=Z88hYmdxp00&t=0s) **Navigating AI Career Paradoxes** - The speaker critiques vague AI‑break‑in advice, highlights data showing both massive job creation and loss, and urges concrete, role‑specific strategies to seize emerging AI opportunities while avoiding entry‑level automation. - [00:04:56](https://www.youtube.com/watch?v=Z88hYmdxp00&t=296s) **Designing Predictive Skill Assessment Prompts** - The speaker explains how they craft interview‑style prompts that anticipate meaningful signals—such as production deployments for ML engineers or editorial editing for prompt engineers—to accurately gauge candidates’ true AI and engineering proficiency. - [00:08:25](https://www.youtube.com/watch?v=Z88hYmdxp00&t=505s) **Evolving Product Management with AI** - The speaker outlines a method for merging traditional product management competencies with AI expertise, using an LLM‑driven assessment to generate a personalized, dynamically updated learning roadmap. - [00:11:59](https://www.youtube.com/watch?v=Z88hYmdxp00&t=719s) **Upstream Prediction for Career Success** - The speaker explains an AI‑career assessment that pre‑emptively flags common pitfalls and leverages early‑stage predictive insight—like meme‑driven probability forecasts—to guide users quickly toward viable roles, avoiding costly trial‑and‑error. - [00:15:19](https://www.youtube.com/watch?v=Z88hYmdxp00&t=919s) **Predictive AI Career Pathways** - The speaker argues that the exploding AI job market demands a personalized, data‑driven approach where carefully crafted prompts guide LLMs to predict an individual’s optimal career move by accounting for market timing, transferable skills, and common hiring blind spots. ## Full Transcript
Most of what you read about how to break
into AI is wrong. Not because it's
incorrect, but because it's too vague to
be useful. It's too high level. It's too
general. Learn AI. Don't do that. Get
specific. Get useful. I have spent the
last few days taking apart 17 different
career paths in AI that people are
hiring for right now. I've looked into
the job market, both where we're hiring
and also where jobs are being destroyed.
And I've come out of all of it with some
overall takeaways for you as well as
some approaches to prompting that help
you to figure out where you can career
path yourselves and where you can scale
up if you have a particular job role in
mind. So let's dive into it. First up,
let's get at the overall job situation.
The World Economic Forum, I know you
love to read them, says 170 million new
AI jobs will be created by 2030. Yay,
yay. Break out the party hats. But 92
million jobs will be destroyed in the
same period. So net gain just 78 million
positions projected. Stanford published
research showing 13% unemploy or
employment decline for workers aged 22
to 25 and AI exposed roles since late
2022. That is not news to anybody I know
who is in that age demographic. Look at
my gray hairs. I am not. So AI is
simultaneously creating opportunities
and it's automating entry-level
positions people use to break in. That's
the paradox that we need to wrestle with
and solve. So when you search how to
break into AI in that world, you get one
of two extremes. You get let's learn
Python in 30 days and become an ML
engineer. Lots of lots of that. Or you
get a little bit of gatekeeping like you
need to be a PhD from Stanford or you
need to get into Y Combinator, you need
to found your own startup. Both of these
are suboptimal because they only apply
to a very narrow subsegment of the
overall population. If you look at the
overall number of people interested in
AI roles, not very many of them are
going to be learning Python in 30 days.
If you look at the overall number of
people interested in AI, not very many
of them are getting into Y Combinator
and founding a company. And nobody talks
about it as if AI is not a single career
path but a but a wild maze of career
paths that's developing. And I want to
start talking about it like that because
that's how it actually is. Right now,
when people say, "How do I break into
AI?" It's sort of like a chef hearing,
"How do I break into food?" I mean,
that's a pretty general statement. Do
you do you want to cook? Do you want to
do front of the house? You want to do
back of the house? You want to get into
investing and owning restaurants? Uh, do
you want to just be a taste maker and
hang out like Anthony Bourdain? Like,
who knows? It's Anthony was a chef, but
you get the idea. My point is that we
need to get specific with career
pathing. And that's what this video is
designed to help get you to. First,
let's get into some examples. An AI
research scientist will need a PhD, will
need published papers, and Meta is going
to pay them a lot, right? Like Meta
might pay them half a million dollars.
An AI prompt engineer needs strong
writing, doesn't need coding. They're
not going to make as much as the meta
seller. They might make around six
figures, but they also don't need the
same background. A machine learning
engineer, wow, that's showing tremendous
growth year-over-year, but it is a
technical role, and you have to lean in
on the technical chops. Meanwhile, AI
Coach shows even stronger year-over-year
growth at I think almost 60% according
to the research I did, but it doesn't
require the same technical chops as a
machine learning engineer. These are
entirely different roles with different
prerequisites. So, I want to get at ROS
specific qualification assessments that
help get you an idea of where you stand
and that's the project that I've been
working on. So, let me set it up for
you. I built 17 different assessment
prompts. Each one is designed around a
core insight about prediction. There's a
piece that came out over the weekend
from A16Z about how prediction is
replacing postmodernism as the defining
framework of our era. The idea is
simple. Value creation is now about
being predictive of the game rather than
being predicted by the game. And that's
exactly what these career assessments
set out to do. They help you be
predictive of your career path rather
than being predicted by it. And that may
sound like philosophy mumbo jumbo, but
it's really relevant. I find the people
who are able to make progress in AI are
people who are able to accurately show
intent and bridge that intent forward
into the future by making specific bets
on themselves. And what all I'm trying
to do is set up prompts that help you
make that kind of bet on yourself. The
alternative is really simple. It's
wasting 6 months, 12 months, or more
during a really high leveraged period in
human history pursuing goals that don't
fit you well. And that's what happens
when we just generically pursue AI. The
old postmodern career advice was
everyone gets a personalized version of
their career, right? Take this boot
camp, learn this framework, you get a
unique resume, you'll be fine. That's
not really true anymore. And we're
seeing that breakdown. What matters now
is timing relative to the AI revolution.
where you engage and when you engage
with the AI snowball that is rolling and
tearing down the hill. Are you early on
AI governance when the EU AI act makes
it the hottest role next year or are you
late to entry-level software engineering
when those positions are down a bunch of
percent. The assessment tells you where
you stand and that's what I set out to
do because then you can make a bet about
where you want to go. So, how did I
build these prompts? Building the
prompts required me to think about
prediction and anticipation deeply. I
took each prompt and structured it out
as an eight question interview, but the
questions are not random. They're
designed to extract signal about your
qualification level from relatively
limited information. Here's how I think
about it. A good prompt anticipate what
matters and it anticipates what doesn't.
So for an ML engineer assessment, I
don't ask, "Do you know AI?" I ask,
"Rate your Python proficiency 1 to 10
and describe your experience with
TensorFlow or PyTorch. Have you deployed
models to production?" That second part,
the deployment piece, that's heavy
signal. Lots of people will take a
course. Relatively few will deploy to
production with something meaningful.
For an AI prompt engineer, I ask, can
you share an example of heavily editing
AI content to publication quality?
Because that's a real skill. Anyone can
generate the content. I've talked about
this. The value is in the editing. The
prompt anticipates the gap between what
people think the job is and what it
really requires. for AI governance
specialists. Another one I ask about the
EU AI act specifically, not do you know
regulations in general, but the EU AI
act is one of the biggest pieces of
regulation on the planet around AI right
now. Any company anywhere in the
European Union, whether based there or
not, has to deal with it. It's the
hottest governance role in the new year.
And the prompt anticipates the market
reality. So these questions are
structured to predict one of four
outcomes. either you're qualified now
for the role, you're almost qualified,
there's significant gaps, or it will
just say you're straight up not viable.
And I think the honesty is helpful. And
also each category gets to a timeline.
And that's where the prediction piece is
there and it gets a little fuzzy, right?
Like 3 to 6 months for nearly qualified.
Some people speedrun that with AI,
right? 6 to 18 months for significant
gaps. That's fair, but again, people can
sometimes accelerate that if they lean
into AI really hard. One of the things
we forget is that AI is a kind of super
skill. And if you learn how to learn
with AI, you can speedrun some of these
challenges. So the prompt doesn't just
assess current state. It predicts the
path forward based on patterns that I've
seen in real career changes in the data
in 2024 and 2025.
It's not theory, it's this has worked
for people with your background. Let's
look at a pro. Okay, this is an AI
product management prompt. As you can
see, I have the same idea. You're
setting up the interview at the top here
and you just want to ask eight targeted
questions you're looking to get signal
you can predict against. In this case,
we're targeting product management. So,
we ask about overall PM experience. We
ask about a IML concepts and we ask
specific questions there, right? What is
the difference between supervised and
unsupervised learning as an example? I
want to make sure that you're not just
rating yourself 1 to 10 and kind of
guessing or using cluyly to answer that
you're actually able to be honest about
it. We go through, we talk about
stakeholder management, strategic and
analytic skills. We talk about
leadership and communication. One of the
things I want you to notice AI product
management is a great example of a role
that's transforming and evolving. Right?
We are looking at classic areas of
product management expertise that are
evolving and changing in the age of AI.
And I think part of the way forward is
to map them to specific competencies.
And so we start to look and say how do
we take classic experience with stuff
like product strategy and road mapping
and start to crosscalibrate it with how
you've built stuff with AI capabilities
in the product. And so a lot of the
magic of this assessment is actually
putting it together analyzing the
responses and assessing how you qualify
as an evolved role where you have both
the PM piece and the AI piece. And then
there's a personalized road map at the
end because I want to leave you with
something that's actually actionable,
right? What are some recommendations
that you can take with you? And this is
useful by the way because
you can depend on language models given
this kind of prompt to go and search the
web for the most relevant current
courses and certification. I could have
hardcoded all of these and I chose not
to because the space is moving so fast.
And so instead, given that we have
reasoning models that search the web
effectively with a very strong anchor
prompt like this, when it comes time to
recommend, they are going to look across
your response set and be able to
generate a custom list of courses for
you that reflect your specific answers
in a way that hard coding never could.
And that's a reflection of where we're
at with language models at this point in
2025. It's incredible. like you can get
them to actually think through all of
this and on the-fly recommend courses as
long as you have a solid prompt. The
salary reality check. This is one based
on current market conditions. I could
have also soft coded this and had them
go the the prompt go out and benchmark
the salary. I decided that was not worth
it. I would rather prioritize the
learning piece with the LLM tokens than
the salary piece. And we can always
update that over time, but it is
reflective of where the market is at
this point in the US. you'll need to
discount it some for the EU. Finally, we
get to the assessment criteria. I want
you to understand how I assess this
stuff so it's bold and transparent. I
also want the LLM to be consistent about
how to assess it. And that's it. Begin
now by introducing yourself. So there
that that's the prompt. It's actually
not a magic prompt. I'm a big fan of
showing how I prompt so you don't think
I just do this stuff magically and it
works. I want you to understand how it
works. You know, the hardest part of
building these prompts was building in
anticipation of what people don't know.
They don't know. Most people applying to
AI roles don't realize how transferable
some of their skills are. A compliance
professional doesn't recognize the
transference potential to AI governance
necessarily, and if they do, they don't
know which skills transfer and which
don't. Someone with 5 years of change
management might not realize they're
qualified for an AI coaching role with
tremendous year-over-year growth. So, I
structured questions intended to surface
those realities. I didn't just ask, "Do
you have AI experience?" I asked about
the background, the real background, and
the prompt explains how it maps. You
have compliance background. That's a
scarce resource, right? You can get
there faster than a software engineer
could. This all connects back to that
prediction framework. We create more
order in the universe by contributing
information. That's sort of the central
thesis of the A16 Zpoint. It's been kind
of bouncing around in my head. And I
think that part of what we're doing with
prompts like this is we are collecting
information about what you actually have
and bring to the table. Not what you
think you're missing, but what you
actually possess. And then we predict
the most efficient path to contribute
that information into the AI job market.
That is a path to value. That is a path
to career success. And that's why I
built this. The assessment also
anticipates mistakes. I designed it to
call out common failures, analysis
paralysis, targeting the wrong role,
unrealistic timelines. There's a lot
that potentially derails people and the
prompt strives to get after failure
modes and address them up front. Here's
what I keep coming back to. The A16Z
piece talks about how being good at the
internet became a monetizable career
because prediction is an the entire meta
game. It's the entire game that we're
playing in the economy right now. We
literally share memes with probability
prediction for game results, right? like
I am a Seattle Mariners fan and there
were memes circling about the
probability of the Mariners winning this
past weekend. And that's not the first
time I've seen that. I've seen that for
game after game after game, sports team
after sports team after sports team.
Prediction is a meme. The farther
upstream you are in that information
flow, the more likely you will be
predictive of the game as opposed to
being predicted by the game. And that's
where these what these assessments do
for AI careers. They put you upstream in
the information flow. Instead of
spending 6 months learning Python and
then realizing ML engineer isn't viable
with your background, you know it very
quickly in like 15 minutes. Instead of
applying to a 100 jobs and getting
nowhere, you know more of which door to
try first based on real market data and
an AI customized conversation. So maybe
you're going after the 41% growth in ML
engineering roles or even the 134%
growth in AI content creator roles.
That's insane. You're being predictive
instead of being predictive. You're
contributing information about your
actual qualifications to the market
rather than getting lost in generic
advice that doesn't account for who you
are. And this matters more now because
of that entry-level job displacement
that I talked about at the top of this
video. The old path was get your
entry-level job, learn on the job,
advance that is disappearing. The new
path requires you to be strategic about
which door you try first, and that
requires prediction. But there haven't
been very many resources about it, and
that's why I created these prompts. I
identified four overall pathways, if you
want to ladder all of this up, that work
in late 2025 headed into 2026. There's a
technical bridge for people with a
coding background into AI. It's a
timeline of 6 to 18 months, depending on
how fast you learn. There's a
non-technical entry role for writers and
creatives. That one's faster. It tends
to be 1 to six months if you really lean
in. There is a domain expert pivot, and
that one really varies by your
experience level with AI and the exact
domain you're in. It's tough to predict,
but from what I've seen, 3 to 12 months
is pretty reasonable. And then there's a
whole burgeoning exploding governance
and compliance uh route. And that one is
fairly linear. It's about 3 to nine
months if you have some prior experience
in any of the compliance roles. The key
is those pathways have different timing.
That's the prediction element. If you're
a compliance professional, you can
become an AI governance specialist a
whole lot faster than you would become
an ML engineer. And so the prompts help
you predict the pathway that fits your
background. and then give you a little
bit of a sense of the timeline. And they
might not give you what you want to
hear, right? I wrote them to be honest,
not to be comforting, because I think
that's what we need in this moment. I
wrote them to be true based on what we
see in actual 2025 hiring requirements.
The market is real. Amazon posted
hundreds and hundreds of AI positions
just in the first few quarters of 2025.
Apple ditto, Tik Tok ditto. Everywhere
you look, the AI related roles are
exploding. But everyone's talking about
AI like a singular thing. You need a
personalized approach that's based on
the real market data in 2025. Something
that allows you to start to be
predictive rather than the one being
predicted. You know, every prompt I
write and every prompt you write is
fundamentally about prediction. Given
this chain of information, predict the
next good thing. That's what good LLMs
do. But prompts anticipate what the
model needs to predict. as well. They
structure the information flow. They
extract signal. They account for what's
missing. That's what these career
assessments do. Given this person's
background, predict their qualification
level. Predict their optimal path. And I
had to anticipate what information
mattered. As I was constructing the
prompt, the prompt acted as a bridge to
help you put your intent together and
predict across the future where you
should go from a role perspective. I had
to account for what PE people typically
miss, the transferable skills, the
market timing, the entry-level
displacement, and I had to package all
of that into a prompt so that the LLM
could predict the future in a way that
was useful for you. Building good
prompts is really about understanding
what you are trying to predict, working
backwards to figure out what information
you need to predict it accurately, and
then building that bridge. That is the
craft of prompting, and I wish more
people understood it. But TLDDR, I built
17 different role prediction prompts so
that you aren't asking about AI as a
whole. You're asking about specific
roles. So if you want to learn more, you
can head over to the Substack. I have
all the prompts there. If you wanted to
just dig into that prompt and dig into
how I think about prediction and how I
think about getting into the job market,
I hope this video has been helpful as