Recently, I saw an interesting perspective from Naval Ravikant about the true state of AI and software engineering jobs. He is the founder of AngelList and an early investor in Uber and Twitter, so he knows well about technology and the market.

So the story goes—many people are now panicking over crazy predictions about AI. Sam Altman said AI will take over 95% of jobs, then CEO of Anthropic said software engineers will disappear in 6-12 months. Everyone already believes "programmer careers are dead" and there's a severe resilience crisis. But Naval Ravikant feels all this is exaggerated.

He has two main arguments. First, even though AI is advanced, it will still make mistakes. If Claude or other AI tools write code for you, their output won't be perfect. There will be bugs, architecture issues, errors. People who truly understand the underlying logic can quickly patch these gaps. So if you want to build a solid app with high performance and good error handling, you still need a background in engineering.

Second, many problems in software engineering are beyond what AI can handle. Usually because those problems are outside the training data range. For example, sorting or linked list reversal—AI has seen thousands of examples, so it's an expert. But if you venture into new territory—high-performance code, architectures that don't exist yet, solving problems that have never been solved—you still need manual coding. This situation will continue until enough cases exist to train new models or until AI can do higher-level abstraction reasoning.

But I like one point that Naval Ravikant highlights. The market only wants the best. If there's a better application in a specific segment, people won't settle for mediocrity. The winner-takes-most market means you have to be the best in your field. But good news—what field you can become top in is unlimited. You can keep redefining what you do until you become a leading expert in that subfield.

So the takeaway from Naval Ravikant: don't fear AI. Software engineers who truly understand their craft and can leverage AI tools smartly will remain valuable. Like in other fields— as long as you master your domain and become a top expert, AI won't replace you. This isn't about fighting technology, but about evolving and staying ahead.