Carol Dweck’s research points to a simple claim: ability on its own accomplishes nothing, and effort is the process that turns it into real skill. In the AI age, tools can hand you finished work and quietly remove that effort, which also removes the growth the effort would have produced. The engineers who keep getting better are the ones who deliberately keep the hard practice for themselves and use the tool to check their work rather than to replace it.
Carol Dweck spent her career studying why people with similar talent end up in very different places.
“No matter what your ability is, effort is what ignites that ability and turns it into accomplishment.”
The sentence is easy to agree with and easy to forget. It says that talent, on its own, is only potential. What you can do today is the starting point, not the result. The result comes from the effort you put in between where you are and where you want to be. For engineers and students learning in a time when tools can produce polished work in seconds, this distinction has become more important, not less.
What the research actually says
Dweck’s work separates two beliefs. In a fixed mindset, ability is treated as a fixed quantity you either have or do not have. In a growth mindset, ability is treated as something that develops through practice. The people who improve over years are usually not the ones who started with the most talent. They are the ones who kept working at the edge of what they could do, especially when it was difficult.
Talent is the starting point. Effort is what decides whether that starting point ever becomes ability you can rely on.
This matters for a practical reason. If you believe ability is fixed, every hard problem feels like a test of whether you are good enough, and the safest response is to avoid it. If you believe ability grows, the same hard problem is the exact thing that builds you, so you move toward it. In engineering, where the difficult problems are where the real learning lives, that difference compounds over a career.
Why this matters more in the AI age
An AI assistant is very good at removing effort. Ask for a driver skeleton, a build script, a tricky pointer routine, and you receive a finished answer without the struggle that used to come with it. The convenience is real. The cost is hidden.
When a tool does the effort for you, it also takes the growth that the effort would have produced.
The effort you skip is not wasted time. It is the process during which your understanding is built. When you work through a locking bug by yourself, you are not only fixing that bug; you are building a model of how concurrency behaves that you will use on the next ten bugs. Hand that work to a tool and the bug still gets fixed, but the model never forms. You solved the problem and stayed exactly as capable as you were before.
The reps you skip are the skill you never build.
This is a fine trade for work you have already mastered. If you can write a device-tree binding in your sleep, letting a tool draft it saves time you can spend elsewhere. It is a poor trade for the skills you are still trying to build, because you are removing the very practice that would make you good at them. The judgment is knowing which is which.
How to keep the effort that builds you
Effort is not about refusing to use AI. It is about being deliberate about where you spend your own thinking. A few habits help:
- Attempt first, then check. Before you ask the assistant, write your own answer, even a rough one. Then use the tool to test and correct it. You keep the speed and you keep the part that builds you.
- Keep one core skill fully manual. Pick a skill central to your work, such as reading kernel source, debugging with a tracer, or reasoning about memory ordering, and do it by hand on purpose. Deliberate friction there is your craft.
- Move toward the hard problem. When something feels beyond you, treat it as the place to spend effort, not the place to avoid. That is where ability is actually built.
- Rebuild from memory. After a tool explains something, close it and reconstruct the solution from a blank file. What you can recreate, you understand. The rest you only recognised.
- Give it real time. When stuck, spend ten honest minutes on your own before reaching for help. The effort in that minute before the answer arrives is where the model forms.
None of this is about working harder for its own sake. It is about protecting the specific effort that converts what you can do today into what you will be able to do next year.
Use the tool to check your work, not to replace the work.
This is also why practice sits at the centre of how we teach at TECH VEDA: the understanding that lasts is the kind you build by doing the hard part yourself, with guidance, rather than by watching someone else do it.
Ability is the hand you were dealt. Effort is how you play it. In a time when tools will happily play the hand for you, choosing to do the hard work yourself is no longer the slow path. It is the only path that leaves you more capable than you were.
— Raghu Bharadwaj
