Will Computer Programmers Become Obsolete?

In uncharted territory, human innovation may have the upper hand over AI trained on the past.

Al Jarmon

As of now, you have probably heard of the advances being made in machine learning and AI. With the success of large language models (LLMs), the future portrayed in cartoons like “The Jetsons” from the ’60s and ’80s is closer than ever before. We now have computers that can seemingly answer any question imaginable. Some may argue that AI is close to being sentient. However, it’s important to note that AI, in its current form, is only proficient in hindsight.

From AI that can answer your questions, such as ChatGPT, to AI that can generate images, like DALL-E, they all have one commonality: they derive their training data from past information. This can be extremely useful when AI functions correctly, enabling tasks like trip planning and money-making. However, what AI cannot do is replace human programmers who are constantly seeking new and innovative solutions. This is the primary reason why AI will not render computer programmers obsolete, at least not at this time.

Let me clarify, if you are a seasoned programmer and are familiar with what you are trying to do, you can leverage code snippets available on popular coding forums like Stack overflow to aid you in programming across various languages. However, this assumes that what you are trying to achieve has been done before. Once you venture into uncharted territory, answers become scarcer and less likely to be found in existing AI training data. Let’s take, for example, the release of Panda 2.0.3 on June 28, 2023. If you were to ask ChatGPT the question “What’s new in Pandas 2.0.3,” the response, as of this writing, would be “As of my knowledge cutoff in September 2021, the latest version of Pandas is 1.3.3, and Pandas 2.0.3 does not exist.” This limitation arises because the public release of ChatGPT was trained using data only up until September 2021. If there were significant changes in the 2.x release of Pandas, programmers would have to rely on traditional methods to gather relevant information.

The fundamental problem with current AI is that someone must manually create the training data. Individuals need to experiment with the latest releases, write reports on historical events, and ask questions on platforms like Quora to facilitate AI learning and provide information to the masses. Therefore, computer programmers will not become obsolete unless AI truly becomes sentient.

Reference:

  1. Google Engineer Claims AI Chatbot Is Sentient: Why That Matters – Scientific American
  2. What are AI Hallucinations and Why Are They a Problem? TechTarget
  3. AI can help you plan your next trip—if you know how to ask.
  4. How to Make Money With AI? Using AI Tools for Financial Success

Subscribe

Stay in the loop! Sign up for our newsletter today for getting regular updates.