Decoding Project Q*: The 'humanity-threatening' AI breakthrough likely behind Sam Altman's sacking saga
Shortly before Sam Altman's dismissal, a group of OpenAI employees purportedly submitted a letter to the board expressing apprehensions about a technical advancement within the realm of artificial intelligence that has the potential to 'threaten humanity'.
The OpenAI board's sudden decision to dismiss CEO Sam Altman earlier this week may have been prompted by a revelation from a group of staff researchers who warned about a potential breakthrough with the capability to 'threaten humanity'. The researchers, in a letter to the board, reportedly highlighted the significance of the breakthrough, named Q* or Q-Star, suggesting it could enable Artificial Intelligence (AI) to "surpass humans in most economically valuable tasks." This revelation came after apparent failure on Altman's part to communicate the development to the board.
The story so far
On Friday, OpenAI made an announcement that four board members had terminated CEO Altman and removed President Greg Brockman from the board. The reasons provided were vague, citing Altman's alleged lack of consistent candor with the board.
The tech community rallied behind Altman, and major OpenAI investors sought his reinstatement. Despite negotiations between the board and senior executives, a resolution was not reached.
Microsoft extended job offers to Altman, Brockman, and other OpenAI team members. In response, almost the entire staff issued an open letter to the board, threatening mass resignation unless Altman was reinstated and the board was replaced.
Initially, OpenAI named CTO Mira Murati as interim CEO, but within 48 hours, Twitch co-founder Emmett Shear was appointed as her replacement on an interim basis.
Subsequent negotiations ensued, leading to the removal of all but one board member and the reinstatement of Sam Altman as CEO, though with certain compromises.
What is Project Q*?
Presently, if you prompt ChatGPT to tackle a math problem, it relies on its advanced predictive-text approach, drawing answers from a vast text database and making word-by-word decisions based on human-like responses. This method may yield correct or incorrect answers, but the AI lacks inherent mathematical skills.
OpenAI seems to have achieved a breakthrough in this domain, successfully empowering an AI model to genuinely solve novel mathematical problems. This advancement is referred to as Q*. Leveraging substantial computing resources, the new model successfully tackled specific mathematical problems.
Does Project Q* threaten humanity?
Mathematics plays a crucial role in the advancement of generative AI. While current generative AI excels in tasks like writing and language translation through statistical predictions, the ability to perform math, where a singular correct answer exists, indicates enhanced reasoning capabilities akin to human intelligence. This could prove valuable in novel scientific research. Project Q* could serve as a guide toward the coveted goal of AI: Artificial General Intelligence (AGI).
According to a Reuters report, some people at OpenAI believe Q* 'could be a breakthrough in the startup's search for what's known as AGI'. OpenAI defines AGI as autonomous systems that surpass humans in most economically valuable tasks. Unlike calculators limited to specific operations, AGI possesses the capacity to generalize, learn, and comprehend.
"Given vast computing resources, the new model was able to solve certain mathematical problems. Though only performing math on the level of grade-school students, acing such tests made researchers very optimistic about Q*’s future success," the report added, quoting a source from OpenAI.
However, the group of staff researchers, in their communication to the board, reportedly emphasized the potential risks associated with the capabilities of AGI. The realm of computer science has long grappled with the debate surrounding the dangers posed by super-intelligent machines. AGI is an AI system sufficiently intelligent to execute any task within the human capability. The realization of this goal could potentially result in the widespread replacement of human jobs by AI.
However, a report in The Verge citing a source refuting the Reuters story about Project Q*. "A person familiar with the matter told The Verge that the board never received a letter about such a breakthrough and that the company’s research progress didn’t play a role in Altman’s sudden firing."