Reading Time: 3 minutes

ChatGPT is the hot topic in recent weeks. The internet is being flooded with the social media posts, blogs and videos.

What is ChatGPT?

ChatGPT is a trained natural language model developed by OpenAI. It interacts in a human type conversational way and can drive the conversation by understanding follow-up questions, admitting its mistakes, challenging incorrect premises, and by rejecting inappropriate requests. It can learn on the fly from ongoing or past conversations and improve its future responses.

Beta Launch

OpenAI has launched a beta version of ChatGPT. The ultimate goal of this launch is not to provide actual service, but to receive the feedback from users regarding pros and cons; limitations; issues; collecting data; learning AI capability and its behavior; potential threats; and to improve it further.

People from all communities (like science, developer, technology enthusiasts etc) have shown a great interest. They are interacting with ChatGPT in their own way and with their own set of questions.

The Feedback

ChatGPT is receiving a lot of positive feedback where people are comparing it with tech giant Google and considering its as next generation search engine which could pose a potential threat to Google’s hegemony in search domain. There are a lot of negative feedbacks as well where people are literally making fun of ChatGPT. There are a lot of social media posts, blogs and videos where people are highlighting the challenges ChatGPT is facing and provided wrong answers while answering basic/simple questions.

The rush to provide feedback and judgement

As we can see that people are literally bombarding internet with immature feedbacks. You can easily sense a FOMO where people just want to catch this ChatGPT train. They don’t want to make the same mistake they made when they missed BitCoin or Tesla stock. We are living in an era where everyone is trying to be an influencer, although most of them are good at nothing. Most of the immature feedback is being rushed by these pseudo-influencers who just want to post something to feed their followers.

Although OpenAI did not yet launch fully ChatGPT but people have already claimed the expertise in this domain and have already passed their judgement on its success or failure. Some of them have gone a step further where they have already launched a crash courses (like ‘how to learn, query or use ChatGPT’ or ‘N cool things you can do with ChatGPT’) on ChatGPT.

We need to understand that ChatGPT has just launched in beta and is not mature enough to compete with existing search engines or other AI models. Reaching on a conclusion based on initial launch is just imprudent.

My Opinion

I have been trying to for quite some time and have tried various kinds of queries related to philosophy, science, biology, programming, politics, opinions and other general queries. Here are a few pros and cons I have noticed.

1. Human type conversational way: It’s impressing to see how beautifully OpenAI has designed ChatGPT to adopt human type conversational way, as well as the ability to drive the conversation by understanding follow-up or past questions.
2. Content size and its representation: The length of the output content is almost perfect. ChatGPT can decide the length based on the number of questions in the query and the type of question. If the uses is asking multiple questions/information in a single query, it’s able to maintain the human type continuity in multiple answers. ChatGPT can also understand if there is any relation between multiple queries or they are completely different.
3. Ability to admit mistake and understanding follow-up questions
4. Quality of content: ChatGPT has shown outstanding performance in understanding and answering almost all kinds of questions, from all kinds of domains. It has shown phenomenal performance and accuracy in answering software engineering questions where it’s able to not only provide correct answers but also to produce code, as well as to debug the piece of code. ChatGPT can provide correct answer to almost all questions which you can find on software engineering sites like StackOverFlow.
5. Throwing diplomatic answers instead of correct one: For political, or controversial topics, ChatGPT refrains from providing correct answers, instead. It takes the route of diplomatic answer which may or may not be correct.
6. Inability to understand simple mathematical problem: Almost all users can see the inability of ChatGPT in understanding simple aptitude math problem. ChatGPT cannot rectify the errors even after pointing out by the users.
7. Inability to provide current information: Although this information has already been put in to public domain by OpenAI that ChatGPT is wouldn’t be able to provide information in current affairs and doesn’t understand world after 2021, but it should still be able to provide basic information like current affairs, world position, games score etc.
8. No Comparison with existing search engines: It’s too early to compare it with other giant search engines. We should allow some years to ChatGPT to fix ongoing issues and become mature.