Who is OpenAI, the company behind Dall-E, ChatGPT and the all-out development of artificial intelligence?

“We believe AI should be an extension of individual human will and, in the spirit of freedom, distributed as widely and evenly as possible.” Reading today the birth certificate of OpenAI written in 2015, it still sounds as revolutionary as it did then.

Create a “friendly AI”

In a blog post At its most modest, OpenAI detailed its mission, both simple and terribly complex: to build a “general” artificial intelligence, that is to say capable of understanding and learning in the same way as a human, or almost. “It’s hard to imagine how much human-grade AI could benefit society. And it’s just as hard to imagine how much it could harm this one if built or used incorrectly”, warned the young shoot. The structure was then designed as a non-profit organization in order to “break free from financial constraints and produce long-term beneficial impact through research”.

Behind these half-reassuring, half-terrifying declarations, we find at the time Greg Brockman, former engineer of the financial firm Stripe; Ilya Sutskever, a machine learning specialist who cut his teeth at Google; and Sam Altman, CEO of venture capital firm Y Combinator. The three men repeat over and over wanting to create a “Friendly AI”, that is, an artificial intelligence that would have a beneficial rather than a negative effect on humanity. This fear of seeing the human race overtaken by machines, the authors shared it at the time with another engineer and businessman: Elon Musk.

Advertising, your content continues below

Silicon Valley Star Parade

Long before taking control of Twitter and laying off half of its employees, the multi-billionaire spoke at length about the dangers of artificial intelligence. In 2014 at MIThe stated in particular that AIs could represent “the biggest existential risk” never encountered by mankind and that you had to be “extremely careful in the development of artificial intelligences”. It is therefore quite naturally that the entrepreneur found himself financing the beginnings of OpenAI, adhering to the need to build a friendly AI.

But the boss of Tesla was not alone in the adventure. Amazon Web Services and the sulphurous founder of PayPal, peter thiel, had also agreed to feed the coffers of OpenAI. It must be said that behind the wheel were respected Silicon Valley brains from, among others, the once very popular company Leap-Motion. A few years later, in July 2019, Microsoft got its hands on the wallet to inject $1 billion into OpenAI and secure privileged access to the firm’s technologies.

OpenAI stands out as the company to follow

At the time, OpenAI had already presented two relatively stunning versions of its GPT language model, which would become the “brain” of many other projects, such as Dall-E-2 and ChatGPT, both well known today. In 2019, OpenAI Five (bots trained by the company via techniques reinforcement learning) also had the luxury of beating the world champion of Dota 2a very popular MOBA.

Advertising, your content continues below

In 2020 came the tool that will explode the popularity of OpenAI with the general public: GPT-3, the most complex natural language processing (NLP) model produced with 175 billion data points considered. The amount of documents digested by GPT-3 to improve comprehension and text production is also dizzying. To get an idea, the entire English Wikipedia represents 0.6% of the volume of data analyzed by the AI.

In the future, AI will be able to understand language better because it can see what words and sentences mean.

A year later, Dall-E, a cousin of GPT-3, again surprised the small world of technology thanks to its relatively fine understanding of language and its ability to create images combining several concepts such as “a baby radish in a tutu walking a dog”. It was then the turn of Dall-E-2, the first version accessible to the general public, which met with phenomenal success on social networks. “We live in a visual world, explains Ilya Sutskever. In the future, you will have templates that include both text and images. The AI ​​will be able to better understand language because it can see what words and sentences mean.”

Growth and some contradictions

In seven short years, OpenAI has therefore become the company behind the biggest industrial successes in the artificial intelligence sector, not without some controversy. In 2019, the nonprofit transformed into a “capped profit” company to meet significant financial needs. Under this new status, the company agrees to limit its earnings to a ceiling set at 100 times the amount of an investment. A generous amount which, according to some, would take OpenAI away from its historic mission, namely to have “fiduciary responsibilities only to humanity itself”.

Advertising, your content continues below

The company also only has 25% of women or non-binary people among its employees and 30% among its management team. Not enough to reassure when racist and sexist biases are debated. A fairly marked inter-self that promotes a very uniform culture, where research on AI is considered more as a way of life than just a job. According to the MIT Technology Reviewpart of the employee’s pay is also indexed to “assimilation” of corporate culture. OpenAI is also accused of having a very personal definition of transparency, cultivating intense secrecy about its research before publishing (sometimes belatedly) the results.

Still, OpenAI’s concerns haven’t changed, according to its boss. In an interview from 2021Sam Altman repeats that the goal of the company is to create general artificial intelligence “moving in the right direction, not the wrong”. By way of example, the young CEO tries to explain the social utility of a model like GPT-3: “You can imagine private teachers in the form of AI who could teach you everything you want to know […] AI life coaches, therapists or medical advisors”lists the Californian, which nevertheless specifies that uses like these “will still require a lot of time and a lot of work”.

With such ambitions and the risks they represent, it is hardly surprising that OpenAI seeks at all costs to tame AI.

Advertising, your content continues below

We would like to say thanks to the writer of this post for this awesome web content

Who is OpenAI, the company behind Dall-E, ChatGPT and the all-out development of artificial intelligence?

Check out our social media profiles and other pages that are related to them.https://yaroos.com/related-pages/