The ChatGPT created by OpenAI has generated several debates on the internet for allowing people to finally have greater access to the world of artificial intelligence. Technology brings many benefits, but it has already been banned from schools in the US, for example. Now Meta’s head of AI said in a speech that “ChatGPT is not revolutionary”.
The statement was given by Yann Lecun, AI lead at Meta during a Zoom conference held last week where he said:
In terms of the underlying techniques, ChatGPT is not particularly innovative. It is nothing revolutionary, although that is how it is perceived in the public. It’s just that, you know, it’s well put together, it’s well done.
Lecun criticized the view that OpenAI is creating something unique, as he said other companies have made similar creations in the past, but have not received the same visibility.
It’s not just Google and Meta, but there are half a dozen startups that basically have very similar technology. I don’t mean to say it’s not rocket science, but it’s really shared, there’s no secret behind it if you will.
In fact, OpenAI’s GPT-3 was created with several components that have been studied for decades, more precisely since the 2000s. As an example, Lecun cited the Transformer architecture, which was created in 2017 and is used by Google in several projects:
You have to realize, ChatGPT uses Transformer architectures that are pre-trained in this self-supervised way. Supervised self-learning is something I’ve been advocating for a long time, even before OpenAI existed.
The scientist even mentions that Transformer was created based on the work of Yoshua Bengio, who presented the first neural language model about 20 years ago. Although that model is considered small today, it can be compared with the invention of the wheel, as it was the forerunner of what we have today.
Cade Metz, journalist for the New York Times, even questioned Lecun if FAIR, Meta’s AI team, will create something advanced that will be known by the public in the same way as OpenAI and the answer was as follows:
Let’s see this from Meta? Yes, let’s see that. And not just text generation, but creation help as well. There are around 12 million stores that advertise on Facebook, and most of them are popular stores and they just don’t have the resources to design a new, well-designed ad. So for them, generative art could help a lot.
Generative art has become widely known recently for creating artistic works from scratch based on existing concepts and Artificial Intelligence. So we should see a lot of news in this sense in Meta and its services while Microsoft should integrate ChatGPT with Bing and other services like Microsoft Office.
