Elon Musk’s artificial intelligence company, xAI, has officially made its AI Chatbot, Grok, open-source. Grok-1’s base model weights and network architecture are freely available on GitHub for anyone to access and modify.
“We are releasing the base model weights and network architecture of Grok-1, our large language model,” the company said in a blog post — adding that Grok-1 is “a 314 billion parameter Mixture-of-Experts model trained from scratch by xAI.”
The company further explains that the released model is from the Grok-1 pre-training phase, from October 2023. While it is trained on a large amount of text data, it is not fine-tuned for any particular task such as dialogue. “This is the raw base model checkpoint from the Grok-1 pre-training phase, which concluded in October 2023,” xAI wrote.
Grok was open-sourced under an Apache License 2.0, and according to VentureBeat, this enables commercial use, modifications, and distribution, “though it cannot be trademarked and there is no liability or warranty that users receive with it.” Venture Beat adds that users “must reproduce the original license and copyright notice, and state the changes they’ve made.”
Grok was first introduced in November last year, with Elon Musk saying that his chatbot is “designed to answer questions with a bit of wit and has a rebellious streak.” He added that “Grok could access real-time data from X, and xAI claimed its real-time access to the X platform gives its model an advantage over OpenAI’s ChatGPT or Google’s Bard.
It should however be noted that besides Elon Musk and his startup’s claims, the model lacks integration with real-time information available on the X platform.
To be able to use Grok, X users have to become X premium subscribers for USD$16 (approx. UGX62,500) per month.
The announcement of Grok going open-source came over a week after Elon Musk sued his chatbot’s competitor, OpenAI, a company he helped co-found with Sam Altman in 2015 and left 3 years later. In the lawsuit filed in February, the billionaire accused the AI research company of no longer being an open-source non-profit organization.
Musk claimed that there was an agreement that OpenAI was supposed to make its technology “freely available” to the public to “benefit humanity”, but the agreement was breached when the company transitioned into a for-profit company and shifted focus to maximizing profits for Microsoft.
OpenAI however refuted Musk’s allegations in a blog post that contained emails from Elon Musk demonstrating that he knew about the company’s plans to ditch open-source and go for-profit. OpenAI also claimed that Musk wanted “full control” of the company.
While open-sourcing technology is believed to help speed up innovations, experts have warned that terrorists could use open-source approaches to create chemical weapons or even develop a conscious super-intelligence beyond human control.
