AngelList Co-founder Naval: AI productivity is overly mythologized, software engineers will never disappear

Author: Naval Ravikant

Translation: Felix, PANews

In the current environment of rapid AI large model iterations, the global market is filled with deep pessimism and anxiety. On one side, OpenAI CEO Sam Altman predicts that “AI will take over 95% of programmers’ jobs”; on the other, Anthropic CEO forecasts that “AI will fully take over software engineering roles within 6-12 months.” The statement that “programming careers are dead” seems to have become a global consensus, facing the most severe “survival crisis” since the birth of the internet.

However, this fear of job disappearance stems from a misunderstanding of the underlying technological logic. Naval Ravikant, co-founder of AngelList and early investor in Uber and Twitter, believes that recent claims about AI boosting productivity may be exaggerated. No matter how advanced AI becomes, it will always make mistakes, and software engineers remain an indispensable profession.

No matter what field you’re in, even in the smallest niche, as long as you specialize and excel to become a top expert, you won’t have to worry about being replaced by AI.

Below are Naval Ravikant’s latest insights.

“Does AI mean that traditional software engineering is dead?” Of course not. Software engineers — even those not directly involved in training or fine-tuning AI models — are still among the most valued groups worldwide. Of course, those responsible for training and tuning models are even more highly regarded because they build the tools that software engineers use.

But software engineers still hold two major advantages. First, they think in code, so they truly understand the underlying mechanisms. All abstractions have vulnerabilities. Therefore, when computers write programs for you (using Claude Code or similar tools), they will always make mistakes.

They will produce bugs, have imperfect architectures, and never be entirely correct. Those who understand the underlying logic can fix these vulnerabilities promptly.

So, if you want to build a well-architected application, if you want the ability to define a good architecture, if you want your programs to run efficiently, perform at their best, and catch bugs early, you still need a software engineering background.

Traditional software engineers can better leverage these AI tools. Moreover, many problems in software engineering remain unsolvable by AI programs. The simplest way to understand this is: these problems are outside their data distribution.

For example, AI is very good at tasks like binary sorting or reversing linked lists because it has seen countless examples. But when you step outside their familiar domain — such as writing ultra-high-performance code, operating on entirely new architectures, or creating entirely new things and solving novel problems — you still need to manually write code yourself.

This situation will persist until enough new training data becomes available, or until these models can perform sufficient reasoning at higher levels of abstraction and independently solve difficult problems.

Remember: The market has no demand for mediocrity. As long as a niche already has a better application, no one wants mediocre ones. Better applications will almost always capture 100% of the market share. There might be a tiny fraction of market share going to the second-best app, simply because it excels in a niche feature or offers a lower price, and so on.

But overall, people only want the best. The bad news is that competing for second or third place is pointless — like Alec Baldwin’s famous line in the movie Glengarry Glen Ross: “First prize is a Cadillac, second prize is a set of steak knives, third prize is you’re fired.”

In today’s winner-takes-all market, this is absolute truth. The bad news is: if you want to succeed, you must be the best in your field.

However, the fields where you can be the best are endless. You can always find a niche suited to you and become a top performer there. This reminds me of a tweet I once posted: “Strive to be the top talent in your field. Continuously redefine what you do until your dreams come true.”

I believe this principle still applies in the AI era.

Related reading: A memo from 2028: If AI wins, what will we lose?