There’s so much happening with artificial intelligence that it’s staggering. That’s not to say that Skynet is going to take place soon, but its development has gone leaps and bounds. And it could get even better from here, especially with OpenAI leading the charge.
For those of you unfamiliar, this organization is one of the biggest out there when it comes to artificial intelligence. Its research laboratory has been doing great things over the years, attracting a number of high-end partners. These include Elon Musk, Sam Altman, and Microsoft, who recently invested a billion dollars into them. But, what is OpenAI all about? Let’s take a closer look.
What Is OpenAI, Anyway?
This research lab first opened in December 2015, which means it’s been operational for over six years. It’s located in the Pioneer Building in San Francisco, where it continues to call home. And it’s here where the company makes major advancements.
OpenAI actually consists of two different corporations. The first is a for-profit group known under the name OpenAI LP. The second is the initial parent company, a non-profit called OpenAI Inc. Yes, its possible to operate a for-profit and non-profit under one roof, and contribute for the greater good.
The company specializes in researching artificial intelligence and its many tools. However, instead of creating privatized technology for corporate giants, it focuses on the “greater good” of humanity. Its research primarily digs into developing friendly AI systems.
The company initially got its start with investors like Musk (who has since departed but remains a donor) and Altman, along with others. Since then, it’s vowed to “freely collaborate” with a variety of institutions in the hopes of making AI more accessible. This is through a number of patents, as well as research that’s easily accessible to the public circle. This is important, considering what the company has in mind with its artificial intelligence developments.
But What Does OpenAI Research?
Now that we have a pretty good idea of who OpenAI is, the real question is – what the heck do they do? Well, quite a bit of research, and not necessarily in one small area.
While the overall goal is to help humanity overall, OpenAI has branched out in many ways when doing so. This includes establishing a program called OpenAI gym, which launched back in 2016. With this, it created reinforcement learning research for many to access. But this was merely the start.
In 2017, the company launched a software platform under the code name Universe. This actually helps gather general artificial intelligence across games and other platforms. This may explain how Microsoft became involved two years later, investing a whopping $1 billion into the company. As a result, Microsoft now serves as a commercial partner, hoping to implement AI technology into its products. (Who knows, we might just be playing an AI-driven Xbox simulator one day.)
Perhaps the latest advancement that people will take interest in is GPT-3. This serves as an advanced language model that actually uses trillions (yes, trillions) of words from the Internet. With this, OpenAI hopes to use the program as a way to answer questions through natural language, without human involvement. Not only that, but it would provide translation tools, as well as the ability to improvise text within certain pieces of technology.
It’s all still well in the works, but it shows how much Altman (still with the company as CEO) and his team are pushing forward with it.
Next Up, DALL-E
As for what OpenAI is currently working on, it’s tackling an advanced Transformer model. (No, we don’t mean the robot.) That model is DALL-E, not to be confused with Disney’s WALL-E.
It serves as a means to build images based on textual descriptions, with a 12-billion parameter version of GPT-3 lending a hand with providing natural language inputs.
On top if that, it also has another program called CLIP, and it does kind of the other way around. It actually creates the description for an image that’s provided for it. This allows the two to work hand in hand with an advanced AI model. That’s something that doesn’t just stand to help businesses, but also potential college students and up-and-comers in the AI world.
DALL-E recently got an update this past April with the aptly-named DALL-E 2. It acts generally the same, but serves up more realistic means with its images. More than likely, OpenAI will continue to work on the program. We could even see a DALL-E 3 somewhere down the road, likely by 2023.
So, in short, OpenAI has a lot on its plate. It’s moving and shaking in the world of artificial intelligence, and it’s not showing any signs of slowing down anytime soon. Especially with Microsoft on board. Don’t be surprised if they cook up something cool by OpenAI in the future.