ChatGPT's code generation will not impact IT industry




 Many of my conversations in the last few days revolved around ChatGPT. This is hardly surprising, given the impact the model from OpenAI has created. In particular, most people seem to think of ChatGPT’s ability to generate code as a big game changer or a big threat. Here I beg to differ.

The importance being given to the code generation capability of Large Language Models such as ChatGPT is a classic example of vividness bias. In short, vividness bias describes the tendency of the human mind to ascribe more importance to phenomena that appear sensational. Airplane accidents attract a lot of attention, but in reality they are responsible for a very small percentage of accidental deaths.

Coding is a very small part of the IT industry. A small percentage of people employed in the IT sector actually do any significant amount of coding. Let me explain how. I will restrict myself to India, as I know this industry.

First of all, almost 80% of the people employed in the IT sector work in the ITES/BPO industry (4 million out of a total of 5 million). Coding is not really a part of their job. Of course, the ITES companies do have development teams, but we can still consider their coding volume as small.

Let’s look at the remaining industry. According to various estimates, 60% to 80% of the spend on software systems happens in the maintenance phase. So it is reasonable to assume that about 70% (taking an average of the estimates) of the IT industry’s revenue comes from software maintenance.

Software maintenance involves front line support, training, configuration and many other activities that do not require coding. Support team members need to refer to code for troubleshooting, not for making any changes. Only the activities of fixing the bugs and servicing change requests require coding. As you can imagine, this is handled by a small percentage of the total workforce engaged in software maintenance.

Of the remaining 30% revenue, a big percentage involves implementing products such as SAP, SalesForce, Dynamics and so on. I do not know how much this fraction is, but I am randomly allocating a 20% part to it. Again, most of the activities in software product implementation are non-coding activities such as requirement analysis, solution design, configuration, testing, documentation, user training, project management and so on. The coding is limited to writing customisation scripts in languages such as ABAP. This is bound to be a small community of developers.

Finally we come to those projects that actually involve writing code, such as new product development. Even in these projects, the proportion of developers is small. Many aspects of product development - product management, requirement generation, design, architecture, user experience, user interface, testing, deployment etc. does not involve coding. But the proportion of coders here is certainly more than that in the other segments that we have seen.

So the biggest group of coders is found in 10% of the section that itself is 20% of the IT industry. Let’s now look at this group a little more closely.

Most of these developers are generously using google for finding code even today. The example code for most common tasks is available on plenty of sites (which is of course why ChatGPT is able to generate it). Many of them are also simply copying the code from some other part of the product they are working on. A lot of software development involves repeating things that are already there and making a few changes.

What will be the proportion of developers that write really innovative code that cannot be found by a google search? If I have to take a wild guess, less than 1 percent of the industry. These are the experts who are behind the new coding techniques that emerge in the industry. I doubt that any generative technology will endanger their jobs.

In summary:

  • Very small proportion of the IT industry workforce does any coding at all.
  • Of those who do, most are already using some readily available code.
  • There are some who do not use readymade code. They work on things that are complex or innovative. These tasks are hard to be done by generative models as these models are trained on things that are already there.
  • The impression that code generation is a big game changer or a huge threat is thus simply blowing things out of proportion.

P. S. I am not saying that coding is not an important activity. In fact, it is the core on which everything else in the industry depends. But it is a small core of a very large industry.

llm chatgpt codegeneration aiindustry

By Devesh Rajadhyax

Co-Founder, Cere Labs


Comments

Post a Comment

Popular posts from this blog

Can language models reason?

Image
  Illustration by Rajashree Rajadhyax Large language models are an impressive technology that excels at providing answers based on vast knowledge of the world. However, they often struggle with reasoning and logic-based questions. To overcome this limitation, the Chain-of-Thought (CoT) method was introduced. This technique emulates how humans solve problems by reasoning step-by-step. Just as we break down complex problems into smaller, manageable parts for analysis and logical resolution, CoT guides language models to adopt a similar approach. By mimicking this natural, structured problem-solving process, CoT enhances the models’ ability to handle logic-driven tasks with greater accuracy and reliability. This approach, however, requires more time and computational effort because it involves processing more tokens . Tokens are the building blocks of text, like words or parts of words, that the model uses to understand and generate responses. Since the Chain-of-Thought method involv...
Read more

AI implementation myths

Image
  Visionary leaders recognize the importance of adopting cutting-edge technologies to enhance their business operations. This is a bold move, especially in the early stages when few businesses have embraced it. Generative AI is currently at such a juncture. Although AI has experienced several setbacks over the years, recent advancements in foundational models like large language models (LLMs) have reignited interest and investment in the field. Generative AI never fails to impress those who see it in action. At Cere Labs, my company, we have developed a Gen AI framework, and whenever I demonstrate it, the feedback is unanimously positive, with many stating it is "useful." However, despite this enthusiasm, there is hesitation when it comes to integrating this technology into business practices. Addressing certain misconceptions and concerns could help leaders more easily adopt Generative AI to enhance their operations. With my 8 years of experience assisting businesses in inco...
Read more

Homework 2.0: The science behind homework and why it still matters in the AI age!

Image
  Illustration by Rajashree Rajadhyax I recently came across an article about how teachers are coming up with methods to distinguish between homework done by students and that done by AI. This effort by teachers is not without cause; a recent case in Massachusetts saw a student penalized for using AI in their history paper, highlighting the need for such measures. I found it intriguing and couldn’t help but reflect on this shifting trend.  Homework! Does that word bring back memories? The race to finish it as fast as possible so you could finally go play! Holidays would’ve been so much better without it, don’t you think? I know I felt that way, and I’m pretty sure you did too. And let’s not forget those classmates with endless excuses for not doing it. The most common one? “I wasn’t feeling well!” 🤒 Some would even throw in a few coughs just to make it believable 😏. I was that kid on the opposite end. Just before a holiday, I’d be sneaking in some homework during lunch break...
Read more