Andrej Karpathy, former AI Director of Tesla and former researcher at OpenAI, recently attracted attention when sharing a frank view on artificial intelligence.
He described AI as a very powerful strange tool but given to humans without any clear usage instructions.
Karpathy is a familiar face in the technology world, who has contributed to building Autopilot and Full Self-Driving systems for Tesla.
However, even with a deep professional background, he admits that the rapid development speed of AI often makes him feel behind.
I have never felt so behind as a programmer," said Karpathy.
According to him, this feeling does not only appear in non-industry people, but even AI experts are struggling to understand and effectively exploit new tools.
This reflects a major shift in the technology industry: AI is no longer just a simple support tool, but is becoming a partner in the working process.
In the field of programming, this change is very clear. Instead of spending most of their time writing code from scratch, many developers today focus on guiding AI, checking output results and correcting errors.
In other words, humans are increasingly playing a supervisory and regulatory role, while machines are responsible for the initial generation.
However, Karpathy pointed out that AI is far from traditional software. These systems are often unpredictable, can make mistakes and give different results each time used, even with the same request.
This makes understanding how they operate and completely relying on output much more complicated.
However, Karpathy does not deny the great potential of AI. He believes that these tools promise to bring a leap in productivity, especially for those who know how to adapt.
According to him, those who quickly learn how to work with AI can achieve outstanding efficiency, while those who are slow to adapt are at risk of being left behind.
Karpathy even said that he could work 10 times more efficiently if he could reasonably connect new AI knowledge and tools that have appeared in just the past year.
The problem is that there is no official guidebook for this process. Users are forced to both do and learn, experiment continuously to find the most suitable exploitation method.
Karpathy's point of view shows a notable reality that AI is changing how people work faster than the adaptability of its creators.
In that context, the biggest challenge is not only developing technology, but also learning how to use it effectively and responsibly.
