<p>Aravind Putrevu</p>.<p>Every few years, the technology industry produces an “end of programming” headline. Today, the role is played by generative AI. The reality is more nuanced. AI will not end developer careers. What it will end are parts of the job one could get away with doing slowly. Developers whose value lies in turning tickets into boilerplate are exposed. Those who can turn ambiguity into reliable, shipped outcomes are likely to become more valuable.</p>.<p>What the AI boom actually changes</p>.<p>Coding is getting cheaper. Engineering is not. AI makes writing code dramatically fast, but speed alone does not improve reliability, security, performance, cost efficiency, or user experience. In fact, in many cases it can degrade them, because scale encourages volume over judgment. The bottleneck is no longer whether code can be written, but whether the right thing is being built—and built correctly.</p>.Advent of AI triggers job concerns in ad industry.<p>Junior roles will change, not disappear</p>.<p>A lot of entry-level work used to be small, well-scoped tasks: CRUD screens, simple bug fixes, and repetitive tests. AI can now produce these in minutes. As a result, companies will expect juniors to operate at a higher level: understand requirements, handle edge cases, add instrumentation, and ship with confidence much earlier than before.</p>.<p>The differentiator at the entry level will increasingly be judgment: knowing what to ask, what to verify, and what to ship.</p>.<p>“Prompt engineering” is not a career</p>.<p>Prompts are cheap. Repeatable workflows that produce correct output are not. What will matter is the ability to integrate artificial intelligence into disciplined engineering systems: generate with constraints, validate with tests and linting, run evaluations to catch regressions, and embed guardrails into continuous integration and continuous delivery/deployment (CI/CD) pipelines.</p>.<p>In practice, AI should be treated like a tireless junior developer: helpful and never exempt from review.</p>.<p>When code generation is abundant, the differentiator shifts to those who can review, critique, and improve code quickly, especially code they did not write. The ability to spot failure modes and write tests that protect behaviour becomes a career moat.</p>.<p>AI can draft code. It still struggles to predict what will break under real traffic at inconvenient hours.</p>.<p>Product thinking moves from optional to essential</p>.<p>AI enables smaller teams to ship more. That raises expectations. Developers who succeed will operate like mini product leaders: engaging with users, clarifying constraints, prioritising outcomes, and measuring impact. </p>.<p>This shit is relevant in India, where many tech careers have been shaped by delivery models focused on “hours shipped”. The market is moving to “outcomes shipped”, and AI accelerates that shift. The safest place to stand is closer to the customer problem.</p>.<p>The real new stack is AI, data, and distribution</p>.<p>Most useful AI products are not just ‘LLM calls’. They are systems that combine data pipelines, retrieval and search, caching, privacy controls, latency management, cost control, and observability. What matters is practical literacy: understanding how models fail (hallucinations and prompt injection), how quality is measured through evaluations and human feedback, and how to ship safely through access control, audit logs, and red-teaming).</p>.<p>A polarised market</p>.<p>There will be fewer roles that purely “implement what’s written” and more roles that demand ownership: define the approach, ship it, keep it healthy, and explain trade-offs. Mid-level engineers who don’t level up in judgement, debugging, and product thinking will feel squeezed. Senior engineers who can combine architecture, speed, and accountability will be scarce and well paid.</p>.<p>The new advantage</p>.<p>AI multiplies output. Careers will reward that multiplier effect. The developer who can ship a meaningful feature in a week, with tests, monitoring, and a clear user story, will outperform the developer who can implement ten endpoints without clarity on why. The new currency is not lines of code. It is velocity with accountability.</p>.<p>For developers who adapt, artificial intelligence is an <br>accelerator. Used as a crutch, <br>it leads to stagnation. Used well, it enables faster learning, better systems, and earlier leadership.</p>.<p>(The writer is VP of growth at an AI-powered code review platform)</p>.<p>Disclaimer: The views expressed above are the author's own. They do not necessarily reflect the views of DH.</p>
<p>Aravind Putrevu</p>.<p>Every few years, the technology industry produces an “end of programming” headline. Today, the role is played by generative AI. The reality is more nuanced. AI will not end developer careers. What it will end are parts of the job one could get away with doing slowly. Developers whose value lies in turning tickets into boilerplate are exposed. Those who can turn ambiguity into reliable, shipped outcomes are likely to become more valuable.</p>.<p>What the AI boom actually changes</p>.<p>Coding is getting cheaper. Engineering is not. AI makes writing code dramatically fast, but speed alone does not improve reliability, security, performance, cost efficiency, or user experience. In fact, in many cases it can degrade them, because scale encourages volume over judgment. The bottleneck is no longer whether code can be written, but whether the right thing is being built—and built correctly.</p>.Advent of AI triggers job concerns in ad industry.<p>Junior roles will change, not disappear</p>.<p>A lot of entry-level work used to be small, well-scoped tasks: CRUD screens, simple bug fixes, and repetitive tests. AI can now produce these in minutes. As a result, companies will expect juniors to operate at a higher level: understand requirements, handle edge cases, add instrumentation, and ship with confidence much earlier than before.</p>.<p>The differentiator at the entry level will increasingly be judgment: knowing what to ask, what to verify, and what to ship.</p>.<p>“Prompt engineering” is not a career</p>.<p>Prompts are cheap. Repeatable workflows that produce correct output are not. What will matter is the ability to integrate artificial intelligence into disciplined engineering systems: generate with constraints, validate with tests and linting, run evaluations to catch regressions, and embed guardrails into continuous integration and continuous delivery/deployment (CI/CD) pipelines.</p>.<p>In practice, AI should be treated like a tireless junior developer: helpful and never exempt from review.</p>.<p>When code generation is abundant, the differentiator shifts to those who can review, critique, and improve code quickly, especially code they did not write. The ability to spot failure modes and write tests that protect behaviour becomes a career moat.</p>.<p>AI can draft code. It still struggles to predict what will break under real traffic at inconvenient hours.</p>.<p>Product thinking moves from optional to essential</p>.<p>AI enables smaller teams to ship more. That raises expectations. Developers who succeed will operate like mini product leaders: engaging with users, clarifying constraints, prioritising outcomes, and measuring impact. </p>.<p>This shit is relevant in India, where many tech careers have been shaped by delivery models focused on “hours shipped”. The market is moving to “outcomes shipped”, and AI accelerates that shift. The safest place to stand is closer to the customer problem.</p>.<p>The real new stack is AI, data, and distribution</p>.<p>Most useful AI products are not just ‘LLM calls’. They are systems that combine data pipelines, retrieval and search, caching, privacy controls, latency management, cost control, and observability. What matters is practical literacy: understanding how models fail (hallucinations and prompt injection), how quality is measured through evaluations and human feedback, and how to ship safely through access control, audit logs, and red-teaming).</p>.<p>A polarised market</p>.<p>There will be fewer roles that purely “implement what’s written” and more roles that demand ownership: define the approach, ship it, keep it healthy, and explain trade-offs. Mid-level engineers who don’t level up in judgement, debugging, and product thinking will feel squeezed. Senior engineers who can combine architecture, speed, and accountability will be scarce and well paid.</p>.<p>The new advantage</p>.<p>AI multiplies output. Careers will reward that multiplier effect. The developer who can ship a meaningful feature in a week, with tests, monitoring, and a clear user story, will outperform the developer who can implement ten endpoints without clarity on why. The new currency is not lines of code. It is velocity with accountability.</p>.<p>For developers who adapt, artificial intelligence is an <br>accelerator. Used as a crutch, <br>it leads to stagnation. Used well, it enables faster learning, better systems, and earlier leadership.</p>.<p>(The writer is VP of growth at an AI-powered code review platform)</p>.<p>Disclaimer: The views expressed above are the author's own. They do not necessarily reflect the views of DH.</p>
