Microsoft acquires exclusive license for Elon Musk’s largest GPT-3 AI language model
- Microsoft is expanding its partnership with billionaire Elon Musk’s OpenAI as it acquires an exclusive license for GPT-3.
- GPT-3 is the largest and most sophisticated AI language model on the market right now.
- According to Microsoft CTO Kevin Scott, the company will be utilising the capabilities of GPT-3 in its own products and services.
AdvertisementMicrosoft already had a partnership going with tech billionaire Elon Musk’s artificial intelligence (AI) research company OpenAI. Now, it’s taking that collaboration a step further with a new exclusive license for the largest AI language model ever created — GPT-3.
The tech behemoth founded by Bill Gates was already lending its Azure cloud computing resources to OpenAI for training its many models. Microsoft made a $1 billion investment to become the company’s exclusive cloud provider.
“While we’ll be hard at work utilizing the capabilities of GPT-3 in our own products, services and experiences to benefit our customers, we’ll also continue to work with OpenAI to keep looking forward,” said Microsoft’s CTO and executive vice president Kevin Scott in a statement.
The new license, which gives Microsoft rights over GPT-3, is another signal of the company’s confidence in OpenAI’s research.
“The deal has no impact on continued access to the GPT-3 model through OpenAI’s API, and existing and future users of it will continue building applications with our API as usual,” said the Musk-owned enterprise.
What is GPT-3?
Released in July, GPT-3 is the third iteration of the company's growing language model.
Despite sparking debate over the ethics of AI and how it's used, GPT-3 has been lauded for being the most sophisticated AI-language model in the industry.
GPT-3 can predict what you want to say before you even ask by running nearly 45 terabytes of text data through 175 billion parameters. It looks at patterns in data to make suggestions. To provide a glimpse of the scale at which GPT-3 operates, Wikipedia’s 6 million articles only make up 0.6% of the entire training dataset.
|Number of parameters
Facebook India Head goes to Supreme Court against Delhi Assembly Panel's summons on 'hate speech'
EXCLUSIVE: YES Bank CEO Prashant Kumar clarifies that the forensic investigation isn't restricted to the top 10 defaulters
EXCLUSIVE: Dailyhunt is curating an entire family of 'Bharat' apps — claims it can push a new mobile app every three days
Popular on BI
- A college is removing its vending machines after a student discovered they were using facial-recognition technology
- 11 states pay more in federal taxes than they get back - here's how every state fares
- Steve Jobs once said the best managers are 'individual contributors' who aren't interested in managing people
- Netflix announces series on 1999 Kandahar hijack, Anubhav Sinha to direct show
- Great to reunite with Sanjay Leela Bhansali after 25 years: Manisha Koirala
- Farhan Akhtar, Shibani Dandekar team up for Netflix series 'Dabba Cartel'
- 10 Fruits that help increase stamina
- DGCA fines Air India ₹30 lakh after elderly passenger's death at Mumbai airport