Hey Everyone - Welcome back! Completely unrelated, but I got married this past week. Starting off on a really strong note!

Let’s dive in.

As AI systems get better at writing code, analyzing data, and generating insights, something uncomfortable is happening inside technical fields. The bottom of the ladder is quietly disappearing. Tasks that used to define entry level roles are now handled instantly by machines. What remains is judgment, context, and decision making. The gap between beginners and senior practitioners is widening into a wall.

Data science is a clean example. The entry level data scientist used to spend time cleaning data, writing basic models, and generating reports. Today those tasks are increasingly automated or dramatically accelerated. Meanwhile senior practitioners who understand the business, frame the right questions, and know when a model is wrong are more valuable than ever. The field did not flatten. It polarized.

This pattern is not unique to data science. It is happening across knowledge work. AI compresses execution. It does not replace responsibility. That means the middle and top matter more, while the bottom becomes harder to enter.

So the real question is not whether this is fair. The question is how to get on the right side of the wall.

If you are early in your career, the answer is uncomfortable but clear. Do everything you can right now to behave like a senior practitioner. That does not mean pretending you know more than you do. It means obsessing over judgment instead of tasks. Learn how decisions are made, not just how tools work. Ask why a project exists. Ask what happens if the model is wrong. Tie your work to outcomes rather than outputs.

The people who survive this transition are not the ones who wait for permission to level up. They are the ones who take responsibility before they are asked to.

But this creates a deeper problem. If entry level roles disappear, how does anyone gain experience in the first place. Senior talent does not appear out of nowhere. Every expert was once a beginner. If we close the front door, the pipeline collapses.

The answer is that the pipeline changes shape.

Instead of large numbers of people doing low leverage work inside companies, we will see fewer people entering through more demanding paths. Apprenticeships. Project based learning. Working on real problems in public. Building things without being asked. Learning in smaller teams where mistakes are visible and feedback is immediate.

This mirrors what already happens in startups. Early stage companies do not hire for credentials. They hire for ownership. You learn faster because the cost of being wrong is real. AI is pushing more careers in that direction.

The old model promised safety first and responsibility later. The new model flips that order. Responsibility comes early. Safety is something you earn.

So what do you do about it. Stop optimizing for entry level checklists. Start optimizing for decision making. Find problems where failure matters and attach yourself to them. Work in environments where your judgment is tested, not hidden. Seek mentors who will give you real responsibility rather than polished advice.

The wall is real. It is getting taller. Waiting for it to move is a losing strategy.

Get on the other side now, while the path is still open.

In the AI era, skill alone is not enough. Judgment is the moat. Experience is the currency. And responsibility is the fastest way to acquire both.

The last week, I went down a huge rabbit hole with paid ads. This is mostly for growing my newsletters. I used to be super against them, but oh my goodness I was wrong.

As I’ve been building Newsletter hero and my newsletters, I realized that I’m not great at marketing. For the things I’m building, I really don’t have market fit with my audience.

Paid ads cuts through all of that.

In a world ruled by AI, speed is at a premium. Paid ads make it unbelievably fast to get in front of the right audience. They also are highly quantifiable to see if what you’re promoting has merit.

The idea of paying money to reach an audience used to really bother me. This is not the case anymore. When I think about all the time I spent making organic content that may not even resonate, the ROI on paid is actually way higher.

Pretty crazy. I feel like I’m living life with an extra arm now that I’m able to use this tool.

Things I Learned

Content I Made

While there will always be new technical skills worth learning, this week’s survival skill is much harder to outsource: Making decisions with incomplete information.

AI is very good at helping you analyze things once the problem is well defined. Real life rarely works that way. Most important decisions happen when the data is messy, the timeline is tight, and waiting for perfect information is not an option.

Early in your career, it is tempting to hide behind analysis. Run one more model. Pull one more dataset. Ask for one more opinion. It feels responsible, but often it is just fear in a productive costume.

In data science, this shows up constantly. You never have perfect data. You do not know how users will react. You cannot fully model second order effects. At some point, someone has to decide whether to ship, delay, or kill the project. Senior practitioners are the ones willing to make that call and own it.

This shows up outside of work too. Deciding to change jobs before you feel ready. Starting a project without knowing if anyone will care. Hosting a dinner without knowing who will show up. None of these come with clean datasets. They still matter.

A useful way to practice this skill is to shorten your decision horizon. Ask yourself, “What is the smallest decision I can make that moves this forward?” Then make it. You can course correct later. Most people never start.

Another trick is to separate reversible from irreversible decisions. If it is reversible, decide fast and learn. If it is irreversible, slow down, but do not freeze. Progress still beats paralysis.

This is how experience actually accumulates. Not from having all the answers, but from making calls, seeing what happens, and adjusting. This is also why senior talent is scarce. Many people avoid decisions. Very few people seek them out.

If people know you are willing to decide with imperfect information, they will trust you with bigger problems. That trust compounds faster than any technical skill.

AI will keep getting better at filling in gaps. It will not choose for you.

Learn to decide anyway.