Who's Behind OpenAI? – Their History, Achievements, and the Future of AI

I was fortunate to live in an era when "artificial intelligence" was just a buzzword or an occasional movie villain. Back then, AI was science fiction: something that we all knew would come eventually but didn't exactly know when or what's the scale of it.

Well, here it is.

Just as my generation didn't know a world before the internet, Generation Alpha won't experience life without AI — and that's thanks to OpenAI.

In this article, we'll discuss OpenAI, its early years, products and achievements, notable figures, and how they shape the future.

What is OpenAI?

OpenAI is a human-first AI research company based in San Francisco, California. Their primary goal is to build an "artificial general intelligence," or an AGI, smarter than humans — but with a little caveat. OpenAI seeks to create a universally adopted AGI that doesn't strive to succeed humanity but is safe, secure, and beneficial.

True to their mission, the team behind OpenAI are a few of the leading voices in AI governance. They implement internal policies to mitigate the risks of artificial intelligence in the future. One such policy is the careful deployment of products to ensure they're market-ready and provide time for policymakers and the public to adapt.

OpenAI has made significant strides in democratizing artificial intelligence. In the unlikely chance that you haven't heard of their name, you're still likely to know some of their products, such as ChatGPT and DALL-E. 

So, how did it all come together for OpenAI in the first place?

The History of OpenAI

Like other companies in the industry, OpenAI started from the ground up. However, unlike other tech startups, OpenAI was built to succeed. Yoshua Bengio, a thought leader and one of the most respected scientists in the world, created a list of the best researchers in the field of artificial intelligence. A few months later, OpenAI hired nine of them.

But I’m getting ahead of myself. Let’s take a little stroll down memory lane and go back to 2015.

Formation (2015)

It was December 2015. Star Wars was back in theaters, the Paris Agreement had just been ratified, FIFA was in the midst of a corruption case, and, somewhere, a group of world-renowned machine learning and artificial intelligence experts were wrapping up a meeting that's set to change the course of technology forever. 

Among that group were Sam Altman, Greg Brockman, Reid Hoffman, Jessica Livingston, Peter Thiel, Elon Musk, and representatives from Y Combinator, Infosys, and Amazon Web Services. Together, they pledged over $1 billion to build a company that would soon be called OpenAI.

Not-So-Humble Beginnings (2016 to 2018)

The first few years of OpenAI focused on research rather than bringing products to the public. They did, however, release products that gained little-to-no traction around late 2016 to early 2017 like OpenAI Gym (reinforcement learning platform) and Universe (AI performance evaluation platform).

Weirdly enough, their most popular output back then was AI bots in the popular MOBA game Dota 2. At the time, OpenAI's bots were so advanced that even the best human players could not beat them.

During this time, OpenAI also poached the best minds in machine learning to join the team, like Wojciech Zaremba. Apart from the significant funding, one of the primary reasons researchers were eager to join the OpenAI team at that time was that they believed in the development and governance of AGI.

Another significant development during this period was Elon Musk’s departure. Sam Altman, the CEO of OpenAI, asserts that the controversial figure left because of a conflict of interests and because he didn’t believe that OpenAI had what it takes to overtake established companies like Google.

Still, OpenAI persevered even without Musk’s donations, bringing us to…

The GPT Era (2018 to Present)

One of the biggest breakthroughs in artificial intelligence came with the publication of Ashish Vaswani's paper titled "Attention Is All You Need." Without getting too much into the technical details, think of it this way:

Artificial intelligence used to look at a sentence as one single thought. Not only is this highly inefficient but it's also made pre-GPT AIs incapable of providing responses without hallucinating. Ashish Vaswani solved that by introducing attention to the mix. Basically, it breaks down a sentence word-per-word, identifies key language, and focuses computing power towards them.

The first iteration of OpenAI’s Generative Pre-Trained Transformer (GPT) was made using that concept and was released in 2018. OpenAI trained GPT-1 on a limited dataset and still managed to defy expectations. It took less than a year before GPT-2 was released, which was significantly faster and was trained with data from over 8 million websites.

Before GPT-3, OpenAI opened DALL-E to the public, an AI image generation platform. At the time, people were highly critical of its output, mostly because of the surrealism and “uncanny valley” feeling of its output. 

GPT-3 was, in every aspect, a significantly better version of GPT-2. However, OpenAI was formally introduced to the rest of the world with the release of ChatGPT, an AI chatbot based on a fine-tuned version of GPT-3. A few months later, GPT-4 was integrated into the ChatGPT platform.

And now, we’re here in a world completely changed by OpenAI and GPT. That’s the complete history of OpenAI.

OpenAI’s Products and Achievements

From 2018 to today, OpenAI has rolled out several game-changing AI software. These include the following:

GPT-n (2018)

One of OpenAI's most significant accomplishments to date is the Generative Pre-Trained Transformer or GPT. It's a large language model that uses sophisticated deep-learning algorithms to analyze and generate text.

GPT is trained on a massive amount of data, which increases with each iteration.

• GPT-1: 7,000 books.
• GPT-2: 7,000 books and 8 million websites.
• GPT-3 and GPT-3.5: 499 billion tokens (words, punctuation marks, emojis).
• GPT-4: No official information, but it’s estimated to be five times bigger than GPT-3.

Today, GPT is used by not only ChatGPT (which we’ll get to in a minute) but also other popular products like Jasper, Writesonic, Frase, Rytr, and more.

DALL-E (2021)

DALL-E is an AI image generation platform first announced in 2021. The first version of DALL-E was only available in closed beta, and so was DALL-E 2 initially. However, the latter was eventually released for public consumption in September 2022.

Currently, OpenAI just announced the third iteration of DALL-E, which promises to have better nuance than DALL-E 2 and competitors like Midjourney and Stable Diffusion.

Codex (2021)

With text and images both covered by GPT and DALL-E, Codex is OpenAI’s solution to AI-powered programming. However, it doesn’t exist as a separate platform but as the model behind the GitHub CoPilot.

Codex is adept at twelve programming languages, including Python and Java. It’s a Swiss army knife of development that can analyze code blocks, anticipate the next line, and generate lines of code.

ChatGPT (2022)

Perhaps their most popular product, ChatGPT is a prompt-based AI chatbot that can generate texts based on an instruction from a user.

ChatGPT is also their most controversial software, besides being the most popular. The chatbot's accessibility and knowledge led to widespread academic dishonesty that's still rampant today.

Whisper

Lastly, we have Whisper: OpenAI's speech-to-text system trained in more than 50 languages. It's designed to make transcription easy and efficient. Moreover, Whisper is also open-source, with OpenAI hoping that experts in the field can refine its algorithms for widespread adoption.

Notable Figures Behind OpenAI

This article wouldn't be complete without mentioning the brilliant minds behind OpenAI. These are the people that are defining the future of artificial intelligence:

Management

• Chief Operating Officer, Sam Altman. Former president of Y-Combinator, former CEO of Reddit, and one of Time’s Most Influential People in the World for 2023.
• President, Greg Brockman. Former CTO and founding member of Stripe.
• Chief Technology Officer, Mira Murati. A respected thought leader on AI governance.
• Chief Scientist, Ilya Sutskever. Co-founder of AlexNet, major contributor to TensorFlow, and one of the best minds in deep learning. 
• Founding Member and Codex Lead, Wojciech Zaremba. Former mathematics prodigy and Google Brain contributor.

Board of Directors

• Adam D’Angelo. CEO of Quora.
• Tasha McCauley. CEO of GeoSim Systems and Fellow Robots co-founder.
• Helen Toner. Georgetown’s Director of Strategy and Foundational Research Grants.

Investors

• Reid Hoffman. Co-founder of LinkedIn.
• Jessica Livingston. Co-founder of Y Combinator.
• Peter Thiel. Co-founder of PayPal and one of Facebook’s earliest investors.

The Future of OpenAI

With the release of GPT-4 last March and DALL-E 3 this month, many are left wondering: what else can OpenAI do?

According to their blog, their short-term goals include releasing more robust AI-powered systems, so it's reasonable to assume that, at least, GPT-5 and DALL-E 4 are coming in the next few years. But what's after that?

OpenAI's ultimate goal has always been to create an AGI that benefits humanity significantly. This doesn't stop at R&D — it extends to investing in ethics research and influence in policy-making. Despite being an industry titan, it's no secret that OpenAI advocates for AI oversight.

This is the world that OpenAI is building.