Those familiar with the world of technology and artificial intelligence will undoubtedly have followed the events of OpenAI in recent days. In short: the company decided to fire Sam Altman, the CEO. After being fished out by Microsoft, OpenAI wanted it back. Later, news comes out about a phantom AGI called Project Q*. But is it all true? Personally I think it's an excuse.
OpenAI's move may or may not be a farce
Let's go in order. On November 17, OpenAI management announced the removal of the company's head, citing a decision of the board of directors. Presumably Altman“has not been open and consistent in communicating with board members, which has prevented from fulfilling his duties. Therefore, the board no longer has confidence in its ability to continue to lead OpenAI. Together with him, Greg Brockman left the position of chairman of the board of directors, but at his request. He was also at the origins of the company. Many other leading specialists followed his example.
The temporary CEO, Emmet Shear, co-founder of the Twitch service, supports the slowing the pace of development of artificial intelligence, which is consistent with the vision of Ilya Sutskever, the co-founder of OpenAI, considered one of the instigators of Altman's firing.
Now, I would like to focus on this aspect: the “slowing down of the pace of AI development”. According to the most recent news, it seems that Altman's dismissal was due to the fact that he ran too much. The sources they confirmed that one factor in a long list of complaints from the board that led to Altman's firing was concerns about commercializing advances before understanding their consequences.
In essence it seems that Altman (it is not known whether alone or in company) has developed something that goes beyond artificial intelligence. This something is Project Q*.
But now I would like to make another parenthesis.
Current developments in machine learning and other methods for creating intelligent systems they are not actually artificial intelligence. Even the most advanced large language models like GPT-4 Turbo they predict the next word at the prompts based on statistics, without understanding the essence of what was done. However, the term artificial intelligence (AI) has become entrenched in media culture alongside neural networks and other machine learning models.
Instead, artificial intelligence is broadly divided into two: restricted e general. The restricted one is the one we use every day using ChatGPT, and it is able to solve specific problems: playing chess, generating text or recognizing objects in an image; the general can act like a human brain and make decisions on any topic. An example: Terminator. In short, general AI is the one that comes closest to the so-called AGI.
So Project Q* is AGI?
By the way latest news disclosed, it seems that after Altman's dismissal and return, a letter came out (but there is no trace of it) in which he talks about the aforementioned OpenAI Project Q*. The project includes a system capable of solve mathematical problems that were not included in the algorithm's training set. Basically, the general AI we were talking about poco above would have been reached.
Sam Altman would have been fired for precisely this: the (alleged) desire to market the progress of AGI before understanding its consequences. Indeed, it is no secret that researchers and scholars in this field are time perplexed about the safety of general artificial intelligence.
However, there are doubts about the veracity of the recent reports. Sources like The Verge they question the existence of the letter to the OpenAI board and the link between AI advances and Sam Altman's firing. In fact, many think that the letter is a fake and that Project Q* does not actually exist. The hypothesis is in fact that according to which this project was set up to save OpenAI's economy, given the collapse in shares following Altman's firing.
