You are reading a free post from my newsletter — The Art of Doing Technical Program Management. I write with an aim to demystify the art of technical program management and deliver proven real-world strategies, practical advise and tips, and actionable frameworks to help you level up as a Technical Program Manager.
If you already are a subscriber and found this essay helpful, please share it with your colleagues and network. I am sure others might find it useful as well.
It’s become the default advice for anyone impacted by layoffs, workforce reductions, or uncertainty about the future. I’ve seen it repeated so often that it feels almost unquestioned. In some ways it reminds me of previous waves.
In 2008 it was “go learn how to build apps.” A few years later it became “go learn cloud.” Today, it’s “go learn AI”.
On the surface the advice seems obvious. AI is changing jobs. AI is replacing parts of jobs. AI is creating entirely new ways of working. Heck everyone from CEO to Customer Support is under pressure to “use AI”. So naturally, people should learn AI. But every time I hear this advice, I find myself wondering what it actually means.
What exactly is a laid-off TPM supposed to do on Monday morning?
Do they just:
Open ChatGPT?
Build an agent?
Create a calorie-tracking application?
Watch a few prompt engineering videos.
Start collecting hundreds of prompts in a GitHub repository?
The advice sounds actionable until you ask for specifics. And that’s where I think it starts to break down for me.
Most people giving this advice also can’t clearly explain what role someone should be preparing for. Remember, jobs are going to change and roles will disappear. We are told that many of today’s roles may not even exist in their current form a few years from now. So, we’re asking people to acquire skills for jobs we can’t clearly define. That feels like incomplete advice.
The common rebuttal is that learning AI isn’t about the next role. It’s about transforming your current role. I think that’s directionally correct. The problem is that most hiring processes aren’t evaluating for that. Plus how does someone laid off even do that; they are unfortunately on the outside trying together back in to make that change.
Recruiters rarely ask: “How would you redesign this role if AI could do half the work?”
Most interviews still focus on how candidates performed the role as it exists today. This creates a strange disconnect. We’re telling candidates to reinvent the future while evaluating them against the past.
Then there is the advice to go experiment.
Build apps.
Create agents.
Ship something.
I don’t think that’s bad advice. I just don’t think it’s enough.
Historically, valuable skills were difficult to acquire. That difficulty created separation. Not everyone could become an expert software engineer. Not everyone could become a cloud architect. Not everyone could build distributed systems. AI is different.
The barrier to experimentation has collapsed. Millions of people can build small applications, agents, automations, and prototypes. That’s incredible, don’t get me wrong. But it also means simply building something is no longer a strong signal. Literally everyone is a “builder”.
A recruiter looking at a hundred resumes filled with AI side projects and 100 prompt deep git repos face a different problem: How do you identify who actually stands out?
Maybe the answer isn’t who can use AI. Maybe it’s who understands how AI changes systems. That feels like a much rarer skill. For TPMs, I don’t think the advice should be “go learn AI.” I think it should be “go learn how AI works”.
Go beyond prompting. Learn how models are trained. Understand context windows. Learn where hallucinations happen and why. Understand the tradeoffs between different model providers. Every frontier lab is publishing free resources.
The deeper understanding matters because it informs judgment. And judgment may become the most valuable skill in the room. Not everything needs an agent. Not every workflow should be automated. Not every problem benefits from AI.
The challenge is no longer building solutions. The challenge is deciding which solutions are worth building. That’s systems thinking. That’s process design. That’s organizational design. Ironically, those are many of the skills TPMs have spent years developing.
I also wonder if recruiters need to evolve how they evaluate candidates. “How do you use AI?” is becoming a weak interview question. Everyone has an answer. Most answers sound something like: “I use ChatGPT to summarize Slack threads.”
That’s useful. It’s also not particularly differentiating. A more interesting question might be: “Tell me about a workflow, process, or aspect of your day to day that no longer exists because of something you built with AI.”
Now we’re evaluating leverage. Now we’re evaluating judgment. Now we’re evaluating whether someone fundamentally changed how work gets done.
There’s a big difference between: “I use AI to summarize meetings.”
And:
“I eliminated a recurring meeting because an AI-driven workflow made the meeting unnecessary.”
One is task optimization. The other is system redesign.
I suspect the second category is where the future value lives. When I think about TPMs navigating the next few years, I don’t think the winners will be the people with the most prompts. I think they’ll be the people who understand where work is changing, where coordination is becoming cheaper, where information is becoming abundant, and where human judgment becomes more important.
The future may belong less to AI operators and more to organizational system designers. Maybe that’s the better advice.
Don’t just learn AI.
Learn how AI can change the systems around you the people and processes.
Double down on Core fundamentals of Great Technical Program Management.
Maybe your resume doesn’t need a list of AI courses you have taken but also examples of how you optimized systems.
Maybe use judgement to have a handful prompts in that GitHub repo instead of 30.
I don’t know… this feels to me to be slightly more useful advice than “just go learn AI” or “learn how to build agents”; a little more purpose and directional than, just be a “builder”.
Until next time.
-Aadil
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.