Learning to Code is Still Worth: This week, Microsoft’s official training and certification division, ‘Microsoft Learn,’ posted a single line on X that directly addresses a question currently causing significant anxiety in the realm of tech education: If you are wondering if learning to code is still a good idea, the answer is: more than ever: no conditions, no asterisks, and no hidden terms.
Depending on one’s perspective, this message from Microsoft could come across as either reassuring or slightly ironic. After all, this is the same company whose CEO stated at Meta’s LlamaCon that 20% to 30% of the code in Microsoft’s own repositories was already “written by software.” It is also the company whose CTO, Kevin Scott, has predicted that 95% of code will be AI-generated by 2030. Thus, the message to students isn’t that nothing has changed; rather, it is that this shift has made the skill even more valuable, not less.
AI now writes a significant portion of code at every major tech company
The data driving this concern is real and constantly evolving. Nadella noted that AI’s contribution is steadily increasing, though results vary by language, with greater progress seen in Python compared to C++. Google has gone even further; Sundar Pichai stated in April of this year that 75% of new code at the company is now AI-generated and approved by engineers, up from 50% the previous autumn.
Anthropic’s own figure exceeds 90%
This is the environment in which every computer science student is viewing the placement season, wondering if they chose the wrong specialisation. Entry-level hiring has become difficult. The “write a simple function” interview question is now outdated and obsolete. Based on these factors, many people online are concluding that this entire career path is closing off.
The nature of the job has shifted from merely typing code to evaluating it
Microsoft’s argument, shared by most practicing engineers, is far more nuanced than the headlines suggest. While AI excels at generating code, it is not as adept at understanding *what* needs to be built, identifying potential points of failure in production, or taking responsibility when things go wrong. Someone still needs to review the output and determine its validity, a task impossible without understanding how the code actually works.
It is also important to consider the volume of work. Projects now involve significantly more code than before, translating to more reviews, more debugging, and more architectural decisions. The bottleneck in the workflow has shifted, but it hasn’t disappeared.
The real value lies in understanding systems, reading unfamiliar code, and effectively utilising AI tools
What has truly changed is the nature of the required skills. Memorising syntax was never particularly engaging, and it has now become virtually useless as a way to stand out. The real value lies in understanding systems, reading unfamiliar code, and effectively utilising AI tools. A Stack Overflow developer survey found that 62% of developers were already using AI coding tools daily, up from 44% the previous year.
Adoption is no longer just a trend; it has become an integral part of the job itself. For its part, Microsoft Learn offers free, structured courses in Python, C#, and cloud development; this makes the post seem both encouraging and promotional of its own courses. Both perspectives can be true. For a student staring at a compiler error at 2 AM, the answer remains the same: keep trying.
Also Read: 87 Million Young Indians Out of School and Work: NITI Aayog’s Plan to Bridge the Gap
