ChatGPT, BARD and Generative AI

In last few weeks, the world was taken by storm when ChatGPT was made publicly available by OpenAI . Since then everyone has been amazed about ChatGPT and Generative AI capabilities in general, how it is able to write code, frameworks and plugins from scratch in a matter of minutes and do a lot of other stuff in matter of minutes. It has also been coined as Google Killer by some and since then 谷歌 has announced it's own version called Bard which will be made publicly available in few weeks.?

As an AI enthusiast I too have been exploring the various capabilities of ChatGPT and wanted to share my thoughts.

The coding capabilities of the solution are no doubt amazing owing to the enormous training that it has received from open source repositories, websites and others. It not only generates code within minutes, it also comments it for better understanding and then provides details on why the code snippet was written in a certain format.?

For example, I asked the system to write a python program to upload 100TB of data in S3 bucket in batches.?

It generated the code in less than a minute and went ahead to explain each step and where all configuration changes could be done to make it customizable.?

I tried another program where I asked ChatGPT to generate a Java code for regression. The result was the same again.

I tried out another problem, where I asked ChatGPT to implement a solution to a problem involving regular expression matching using python. This is one of the questions asked in an interview with Google (question is available online in multiple websites), however, the solution present is in Java. So, I wanted to check if ChatGPT could answer the same but do the implementation using Python. And voila! It actually came up with the solution in Python. Although I have not validated it against all the test cases, this would be definitely be a good point to start with. On top of that, it provides a clear explanation of what it has done to implement the solution.

These are just few examples among many of what ChatGPT as well as generative AI in general is capable of.?

The basic question that has been making rounds since ChatGPT was introduced is "Will it replace the developers from a project"? "Bring an end to developers/programmers"? "Result in the end of the Software Engineer profession"??

The answer would be a resounding NO.

However, it will definitely augment and benefit the programmers and organizations along the way where instead of writing from scratch, you may now have a ready to use customizable code snippet/framework/library (of 1000+ LOC) which you can modify, fine tune and utilize thereby saving valuable development time. Also, the coding standards in some cases may not match with the standards/security protocol that an organization maintains at an enterprise level (since these are generated based on open repositories and websites).?So fine tuning and modification will be needed there as well.

However, this will definitely help in shortening the development lifecycle and benefit organizations in the long run if used carefully.

This will also ensure that the need for "Quality" software engineers (professionals with in-depth knowledge of the systems/tools/languages) is prioritized over Quantity only, resulting in better workforce in general.?

It can also prove to be a very useful tool for upskilling and training associates as demonstrated by the way it not only generates code or answers questions, but rather provides step by step explanations along the way to make things easier for the user to understand.

In terms of whether it can replace human positions (in IT and Services world), here are my two cents.?

The real programmers, architects etc. will continue to stay as their expertise will still be required to fine tune and produce actual results that organizations are looking for (The outputs of ChatGPT and other forms of generative API while accurate are still generic in nature and may not cater to the exact needs of the enterprise). Having said that, some of the low skilled sections may get impacted (particularly in the customer support and production support areas) where ChatGPT and similar systems may provide faster and more accurate response and resolution to queries without the need of an executive to assist at all. That is why upskilling of existing associates is so very important.

#openaichatgpt #openai #generativeai #ai

I'll continue to discuss about other aspects of generative AI including ChatGPT (the creative side) along with what the future holds for us (legislations/lawsuits etc.) in my next article.