AI Makes Coding a Core Skill Beyond Engineering
Andrew Ng, the Stanford adjunct professor and founder of DeepLearning.AI, argues that generative AI makes coding more worth learning, not less. He says professionals who can specify what they want a computer to do—and use AI to turn that instruction into custom software—are already gaining a productivity advantage across functions such as marketing and recruiting. In Ng’s view, the scarce skill is not writing code by hand but defining the task, outcome and criteria precisely enough for a computer to execute.
AI makes coding more worth learning, not less
? andrew-ng rejects the argument that professionals can ignore coding because AI will automate it. His view is the reverse: AI assistance has made coding easy enough that learning it is now worthwhile for everyone.
Ng is not urging people to write code by hand. He says he almost never does that himself. His point is that code remains “the language of computers,” and that a valuable professional skill is being able to state precisely what a computer should do.
For the foreseeable future, one of the most important skills will be the ability to tell a computer exactly what you want it to do.
His disagreement is specifically with senior business leaders who advise people not to learn to code because AI will automate the work. In Ng’s account, AI-assisted coding changes the calculation in the other direction: it makes it far easier to get code written and makes custom software more practical across professional roles.
People who can build do not have to wait for engineering
? andrew-ng says he is already seeing a productivity gap between professionals who can code and build custom software and those who cannot. The divide, he says, is not confined to software engineering. People in other functions can use AI-assisted coding to build for themselves rather than wait for an engineering handoff.
His marketers are one example. Ng says his best marketers do not need to wait for an engineer to build a website. When they have an idea, they can build the site themselves. The source renders that sequence simply as idea → build → live.
Recruiting provides a second example. Ng says his best recruiters do not read large numbers of résumés by eye or by hand; they write code to screen them. The source illustrates the workflow as:
resumes[0→500] → filter(defined_criteria) → hire(best)
The operational distinction matters. AI may help produce the implementation, but the recruiter still has to determine the criteria to be applied. Someone who can express those criteria in code can have a computer screen a large set of résumés rather than manually read each one.
The on-screen pseudocode is schematic, but it captures Ng’s practical claim: professionals can turn a defined task and its criteria into a computer-executed process.
The scarce skill is defining what the computer should do
? andrew-ng does not frame the opportunity as becoming a conventional programmer. His argument is that AI has made it easier to get code written, while leaving people responsible for deciding what they want the computer to accomplish.
That requires more than a broad instruction. A professional needs to identify the desired outcome, describe the work to be done, and specify the criteria that should guide it. Code is the language through which that request becomes executable; AI assistance can help translate the request into working software.
People that can do that will be much more powerful and much more effective.
For Ng, easier code generation does not remove the value of coding. It makes the ability to direct a computer—and to build custom software around a specific need—available to more people. Those who can do it may have a productivity advantage over colleagues who cannot.