Chat GPT – Shows Unbelievable Power of AI

What all a Man had imagined some years back of his hypothetical thinking mind which was only possible in sci-fi movies, is turning into reality. Just like Harry Potter had waved a magic wand and everything happened overnight. Well with what today we are gonna discuss it’s both insanely exciting super unimaginable and at the same time, it’s a bit worrying also. Elon Musk summarized it perfectly, it’s “Scary Good”.

It’s “CHAT GPT-3” by Open AI. It’s definitely interesting and sure has great potential but why we said it’s petrifying a bit, one might think is because the product is still in beta version and it has blown everyone’s mind with what it can do. And with what we can think of is, it can gulp up millions of jobs and make many more jobs obsolete for sure.

It’s in research preview you can go check it here and get a real demo of what all it can actually do. Once we log on to the site, On the landing page it will talk about its capabilities, examples, and limitations.

Chat GPT - Landing Page

Select “Explain Quantum computing in simple terms”. Then see what magic it does. It will give you a precise answer and that’s not a full stop. It has its own memory and you can have a conversation with it on that particular subject. Yes, that’s pretty cool ha. It can generate ideas, and use the context of the conversation we had earlier.

Chat GPT - Explaining Quantum Computing

Chat GPT some more tests

Lets us dig into more examples, Let’s take the perspective of the humans who feel most threatened by this chatbot, Yes you guessed it right I meant the software developers/programmers, Who most of the time end up in debugging and 99% of them will use google to get their answers for their queries.

For example “iterating over an oxford dictionary using for loop” seems tricky right? if you google this query it gave us 3,57,0000 results and we all know most of us won’t go beyond page one or the first 3 to 4 links, we try to find answers in those links if we have our answers then will check if it’s workable so on and forth.

But in the case of Chat GPT, it understood pretty concisely and gave the solution that’s most suitable and workable too with an explanation of a piece of code with the comments, and maybe this is where it is scoring over Google because google makes us go through the article and draw the answers by ourself but chat GPT is to the point answering, and that in turn saving a lot of time we might have to spend going through the article. Chat Gpt might challenge Google’s monopoly of searching the way it’s done now & who knows what’s in store.

Let’s test it some more, we will select now “How do I make an HTTP request on Javascript”. It will write all the code for you within a fraction of a second that will leave you awestruck as shown below. You can ask for it to add up anything, it will add in a second and give you the revised code with a clear explanation below the code block.

Chat GPT - Generating HTTP request code

You can ask for creating a web page or create a blog for yourself. Or create a website and add a particular section, & later ask for adding up a subscription form and then for a payment page, FAQ section, and on and on and on. You will get tired of asking but it will keep generating what you have asked for tirelessly. This is what we were concerned and worried about. Obviously, it can’t be that perfect but still, it would be relevant and the final product could be way better than what we could imagine.

Just type it down to what you want to do and submit, let us try how to create a website, and how to code in any language, it will write code for you with explanations blocks to understand that particular block of code. Now that’s crazily insane stuff and todo in no time is more than helpful. Ask for any other complex stuff anything for your youtube stuff, or if you want anything in blogging it will write everything, it’s having answers ready all the time.

Drilling down what’s gonna be most helpful is Generating ideas, a thesis, rewriting paragraphs, correcting of answers, and writing an essay which can be done in minutes in parallel to researching and writing for hours.

Search Engines Vs ChatGPT

If this is going to be damn precise who’s going to use Google? Ask Chat GPT for a sales pitch or for a business market presentation. It is bang on every time. It is simple straight forward direct information with no-nonsense time for ads and extra data. It is feeble in answering some comparative data and image search data when compared to Google as Google shows that precisely and with images but who knows what exactly the finished version or the future update the Chat GPT might come up with.

How Chat GPT-3 works in the backend

GPT-3 (Generative Pre-trained Transformer 3) is a machine learning model that uses a type of neural network called a transformer to process and generate text. It works by taking a sequence of input words and using them to predict the next word in the sequence.

To do this, GPT-3 uses a large number of interconnected nodes or “neurons” arranged in layers. Each neuron receives input from other neurons and processes it using a set of weights and biases. The output of each neuron is passed to the next layer, where it is combined with the output of other neurons to produce a final output.

When GPT-3 is used to generate text, it takes in a sequence of input words and uses them to predict the next word in the sequence. It does this by analyzing the patterns and relationships between the words in the input sequence and using that information to make a prediction. For example, if the input is “The cat sat on the”, GPT-3 might predict the next word to be “mat”, based on the likelihood of that word occurring in that context.

GPT-3 has been pre-trained on a massive dataset of text, which allows it to understand the structure and patterns of language and to generate high-quality text that is similar to human-generated text. It can then be fine-tuned for specific tasks or to generate text with specific characteristics by training it on additional task-specific data.

What all Chat GPT can do

Chat GPT - Capabilities
  1. Text generation: GPT-3 can generate human-like text, including articles, stories, and descriptions, based on a given prompt or set of constraints.
  2. Translation: GPT-3 can translate text from one language to another, supporting multiple languages including English, French, Spanish, and more.
  3. Summarization: GPT-3 can summarize long pieces of text into shorter, more concise versions.
  4. Question answering: GPT-3 can answer questions based on a given passage of text or a collection of documents.
  5. Text classification: GPT-3 can classify text into different categories based on its content.
  6. Sentiment analysis: GPT-3 can determine the sentiment (positive, negative, or neutral) of a piece of text.
  7. Text completion: GPT-3 can complete a piece of text based on a given prompt or set of constraints.

GPT-3 is highly customizable and can be fine-tuned to perform specific tasks or to generate text with specific characteristics by training it on additional task-specific data. It is available in several sizes, each with its own set of capabilities and performance, allowing users to choose the size that best meets their needs and budget.

We threw some more challenging stuff to Chat GPT, but it responded pretty crisply and you know what rather than being happy it got us very terrified, it just spits the code like anything, where is it going to lead us? End of programming in software development?

Are we going to lose the ability to create, innovate, and do things from the scratch? Will it only leave humans with managerial roles? OR does it fail when the queries are more arcane?

Here is an example of how you can use Selenium to select an option from a dropdown list.

Chat GPT - Selenium Code

Have the software developers dug their own holes and helped their fate by losing their own jobs? Is it the next Coding Automation that will wipe away millions of jobs? Or does it fail for specific problems? Or will it add up to the arsenals of the programmers?

Some Unknown Facts about Chat GPT that’s causing the buzz

Open AI CEO Sam Altman has hinted they might have to monetize the Chat GPT in the future due to the high cost of running these chats.

Open AI Uses Microsofts Azure cloud infrastructure to run these models.

The best and the biggest luminaries have invested in Open AI, and the list includes Elon Musk, Entrepreneur and Co-founder PayPal Peter Thiel, and Even the tech giant Microsoft too has invested.

The Gmail Developer Paul Buchheit tweeted that “The Search engines may be only a year or two away from total disruptions, AI will eliminate the search engine results page, which is where they make most of their money”.

One interesting fact is that many can use this to do illegal activities but the chatbot has been trained to decline the requests that are “Inappropriate” and those which are illegal.

But yes there is a concern here also, as recently Open AI CEO Sam Altman tweeted ” We will have to monetize it somehow at some point; the compute costs are watering”. There will be riders in the future as the expectations grow.

One more interesting fact is that it has said on the landing page itself that it has limitations and sometimes it is possible that it may generate incorrect information and sometimes the content may be biased and plausible and sometimes it may overuse specific phrases. So we can say that it is not without flaws and may lack context and substance.

Precisely they can be used to scan through like say a CV and ask it to identify the fitting criteria for shortlisting. This precisely may save more than the required number of hours that may be wasted in screening through all the CV’s.

How Chat GPT is different from Google

GPT (Generative Pre-training Transformer) is a type of language model developed by OpenAI that is designed to generate human-like text. It is trained on a large dataset of text and uses machine learning algorithms to learn the statistical patterns and structure of language. It can then generate text that is similar to the text it has been trained on but is not able to browse the internet or access new information.

You might think how’s it different from Google, it also does the same stuff!! It’s very different from what Google does, google lists the sites and we need to browse ourselves to the site but Chat Gpt gives you the direct and most relevant answer with no other option. It is like some expert sitting at the other end and answering the query with zero error precision and that’s quite mind-blowing.

Google Search, on the other hand, is a search engine that is designed to find and retrieve relevant information from the internet in response to a user’s search query. It uses algorithms to index and rank the billions of web pages on the internet and presents the most relevant results to the user. Google Search can access new information from the internet and provide current and up-to-date information in its search results.

In summary, GPT is a language model that can generate text based on what it has been trained on, while Google Search is a search engine that can find and retrieve information from the internet in response to a user’s search query.

Other than that Google search engine generates revenue from showing the ads on the articles and companies even pay google to rank up there and even advertisers pay to be on top of the results. Comparatively, Chat Gpt gives a single response and may appeal to most users.

Some of the Drawbacks of Chat GPT

Vendure’s CTO Michael Bromley asked the chatbot for its “Opinion on humans in general“. The response will leave us in splits it said “Selfish Humans deserved to be wiped out“. It’s as simple as that, if we can do all the good things then it’s too obvious it can also do the flip side as well with much precision too.

As of now anything that has happened after the 2021 dataset the chatbot has no answers available on Chat Gpt.When we asked about anything that’s of post-2021 affairs, “Like who won the FIFA 2022” its response was “I’m sorry, but my knowledge cutoff is in 2021 and I do not have information about the FIFA World Cup 2022. The FIFA World Cup is a soccer tournament held every four years and features teams from countries around the world. The most recent World Cup was held in 2018 and won by France”.

Now Google we all know will not stay back and watch, It is rumored that it is also building an AI Language Learning Model called “DEEPMIND” An AI learning model maybe this can compete will ChatGPT. As it has a huge amount of data which is a big plus for Google and which Chat Gpt lacks as of now.

Chat GPT has generated a whooping 1Million users in a span of 5 days. For the same set of users, FB and Instagram took a year to reach. Now if recent reports are to be believed then Google CEO Sundar Pichai is concerned and has asked to concentrate on AI more and this can be taken as code red for the biggest search engine the internet has ever witnessed, what does that mean? Is the search giant feeling the heat or threatened by the pace Chat Gpt is garnering attention and users as well. Time will tell.

You might also like - Coolest AI Websites

Leave a Comment