Elon Musk Advises young people to mix up their studies. This way they will be able to work, according to him, in the world of artificial intelligence and robots. In particular, they must accept “An Education as Broad as Possible: Arts, Sciences, Engineering and Broad General Knowledge”. The key is knowing how to ask questions. The more we know, the better we can tell the machine what we want.
In an interview he was asked about the 20-year-old looking to develop to his full potential. What should you do to keep up with these technological changes? Kasturi is committed to combining art studies with scientific studies. It does not recommend which career to choose or guarantee any job.
“To know what to ask a robot, you have to be able to ask questions”. This is related to so-called prompt engineering (preparing instructions so that the AI understands what it needs to do). Common sense helps us specify our requests.
In a world with many robots and artificial intelligenceWe have machines at our service. We can ask them for things and they will do it immediately. That’s what Elon Musk is hoping for.
However, superficial knowledge is not enough to control complex systems. Kasturi talks about expanding our training. This is not a study plan. And what is good in education is not comparable. What do I need to learn to be able to use them independently?
Already in February 2020, Musk advocated that academic degrees not be a key filter in some Tesla hiring. As we said at the time (searching for programmers and engineers), Musk asked if they knew about artificial intelligence (and could apply it) and that they passed a programming test.
Learn how to evaluate AI responses
In researching the use of AI in the workplace, it is important to know how to judge what the machine gives back to us. (2025) in a Microsoft research study Hao-Ping Li and six other researchers 319 knowledge workers were surveyed. Their work collected 936 examples of the use of generative AI.
For example, participants who were more confident in AI reported doing less critical thinking. On the other hand, the more confident they were, the more critical thinking they reported. However, perception and association are described in this work. Using chatbots can’t be terminated as it impairs intellectual capacity. Verifying information is among the activities where humans are still necessary.
Another specific problem has been measured in education. Hamsa Bastani and his associates A Turkish institute published the results of an experiment on mathematics students. About one thousand students. They compared two different auxiliaries (based on the GPT-4) to a control group (which had no access to these auxiliaries). There are two types of assistants. One is like a simple chat and the other has instructions designed with teachers to protect learning.
With simple chat, a 48% improvement in practice grade was observed compared to the control group. However, test (unassisted) scores were 17% worse than the control group. With a tutor with protection, this negative effect is practically avoided, but in tests without assistants, an improvement is not certified compared to those who studied without AI.
Solving a task does not always mean learning
This difference is collected by Organization for Economic Cooperation and Development (OECD) In its report on digital education (2026). The caveat is that there may be cases where using a generalist AI to assign tasks leads to a better result without implying equivalent learning. It also includes uses that have an educational purpose and that facilitate learning.
For the studying user, it is important to consider more aspects than the results of the answers. For example, if the assistant asks you to think. It is also relevant that you use your efforts to guide you; Thus it helps in detecting errors. The OECD describes tutors who ask questions and adjust their instructions in a dialogue process. To do this, the learner needs to continue training in skills that they will later need to use autonomously.
Picture | With the Xataka version
