By: Abhishek Pughazhendhi, RIG Inc Intern Researcher


Mankind has always found AI fascinating. Although traces of AI development can be found as early as the 1900s (The Ferranti Mark 1 machine, a checkers playing program), mainstream AI that is driving a lot of our current tools and technology began after the 2000s.

Apple’s Siri, Google’s Google Now, and Microsoft’s Cortana were released one after the other between the years 2011 – 2014. It is safe to consider that these were the first AI programs that people found fascinating and made them believe in the future of AI. Since their inception, AI programs that use Natural Language to conjure what the users are looking for grew in popularity. Although, have you ever wondered how far we’ve come?

Siri has been answering our questions since 2011 but what can the most recent AI inventions do? Have they gotten more intelligent? Let’s find out.


GPT-3 (Brings Your Thoughts To Text, Articles & Code)

Imagine sharing a few lines about that article you’ve always wanted to write and with the press of a button, you now have a 10-page blog that’s ready with subtitles, punctuation, and proper grammar. This is just one of several things that the Generative Pre-trained Transformer (GPT-3) can do.

GPT-3 Coding An Entire Game With A Line Of NL Input

This third-generation Generative Pre-trained Transformer is the brain-child of Open-AI. Programmed with over 175 billion machine learning parameters, GPT-3 is now the largest neural network that mankind has ever created. Being considered an epitome of not just Natural Language Processing but also Natural Language Generation, GPT-3 can just accept a few lines from the user reflecting his/her ideas and convert them into poetry, blogs, dialogues, and workable code that runs without error.


DALL-E 2 (Your Personal Digital Artist)

Need someone who can put your wildest dreams onto a canvas? OpenAI’s DALL.E 2 does exactly that. This AI program can create hyper-realistic and artistic images from just a single line of natural language input. Can you imagine an astronaut riding a horse in space? DALL.E 2 can and here’s how. DALL.E uses all the existing images that are available on the open internet and modifies them according to a said natural language caption. It creates variations to the original based on the user’s request.

The Legendary “Astronaut Riding A Horse In Space” Art By DALL.E 2

Through the use of “Diffusion”, caption matching, and photorealism, the DALL.E project can do exactly what a digital illustrator does in seconds. Most tech-enthusiasts would agree that the DAL


L.E project is one of the most revolutionary AI research projects to have ever been developed.


ThisPersonDoesNotExist (Hyper-Realistic Facial Photo Generator)

These Faces Do Not Exist

thispersondoesnotexist is something that may frighten most people. A visit to the site would make you understand what exactly makes it scary. This random face generator generates hyper-realistic photos of people who’ve never existed and is impossible for the human eye to detect. Utilizing the capability of Generative Adversarial Networks (GANs) to its fullest, this AI algorithm can generate new faces (which are completely fake) every time you refresh the page. Although some might consider this sinister, GANs and other similar AI superimposing algorithms are going to drive the future virtual world.



The traditional “Is AI good or bad?” question has now expired. In the current day and age, AI is essential and crucial in moving the human race further into the future. This article is just a glimpse of how far we’ve come in terms of artificial intelligence.

In the words of Eliezer Yudkowsky, “The greatest danger of Artificial Intelligence is that people conclude too early that they understand it”. The line between “human-generated” and “machine-generated” is getting thin with every passing year and we can only assume it’s for the good.







[1] GPT3 –

[2] DALL-E 2 –

[3] ThisPersonDoesNotExist –



[1] Generative Pre-Trained Transformers for Biologically Inspired Design – Qihao Zhu, Xinyu Zhang, Jianxi Luo

[2] Realistic Image Generation from Text by Using BERT-Based Embedding – Sanghyuck Na, Mirae Do, Kyeonah Yu, Juntae Kim

[3] Generative adversarial network: An overview of theory and applications – Alankrita Aggarwal, Mamta Mittal, Gopi Battineni