Musk is never absent from major events related to OpenAI, even as a bystander. Just now, Musk claimed that he received a letter about OpenAI and "the issues (mentioned in it) seem worthy of investigation."
But soon, the letter was deleted.
However, according to relevant reports, this letter was originally written by a former OpenAI employee to the board of directors. In the letter, they stated their dissatisfaction with Sam Altman and Greg Brockman and asked the board to "thoroughly investigate these allegations and take appropriate action." This is in sharp contrast to the move of "more than 700 employees signed a joint letter asking the board to reinstate Altman."
Although the original letter has been deleted, an X platform netizen posted a summary of Grok (a ChatGPT competitor created by Musk):
The letter was addressed to OpenAl’s board of directors and expressed concerns about recent allegations against the company’s CEO, Sam Altman. The former employees urged the board to expand the scope of the ongoing investigation and request private statements from employees who left the company during the turmoil. They claim that many employees were laid off to facilitate the transition to a profit model, and that Sam Altman and Greg Brockman displayed a pattern of deception and manipulation in their pursuit of AGl. They cited examples of such behavior, including discrimination, exploitation of resources and lack of transparency. These former employees are calling on the board to stand up against these unethical practices and launch an independent investigation into the conduct of Sam and Greg.
What was written in the letter from Github404?
The letter comes as some investors in OpenAI were trying to convince Sam Altman to give up his job at Microsoft and return to leading the startup. Others are considering filing a lawsuit against the board over the manner in which Sam Altman was fired. Microsoft CEO Nadella also expressed conditional agreement for Sam Altman to return to OpenAI.
However, according to these anonymous former employees who wrote the letter, OpenAI's governance structure is flawed: "The OpenAI governance structure specifically designed by Sam and Greg intentionally insulates employees from oversight of for-profit operations precisely due to their inherent conflicts of interest. This opaque structure allows Sam and Greg to operate with impunity and immunity from accountability."
"Despite the mounting evidence of misconduct by Sam Altman and Greg Brockman, those who remain at OpenAI continue to blindly follow their leadership, even at great personal cost. This unswerving loyalty stems both from fear of retaliation and the temptation to obtain potential financial gains from participating in OpenAI's profits." (Note: Before SamAltman was fired, OpenAI was planning to sell employee stock at a valuation of $86 billion. This precious realization opportunity is full of variables due to SamAltman's departure)
Former OpenAI employees urged the board to:
Expand the scope of Emmett’s investigation to include a review of Sam Altman’s conduct since OpenAI began transitioning from a non-profit entity to a for-profit entity in August 2018; (Note: Newly appointed OpenAI Emmett Shear promised to investigate Altman’s firing and produce a written report)
Open call for private statements from former OpenAI employees who resigned, took medical leave, or were fired during this period;
Protect the identities of those who come forward and ensure they are not subject to reprisal or other forms of harm.
They claim that a large number of OpenAI employees were driven out of the company simply to allow the company to better transition to a profit-making model, and these accusations can be proven by the fact that the company's employee turnover rate (employees leaving due to unpredictable or uncontrollable reasons) was approximately 50% between January 2018 and July 2020.
The former employees also noted that they "witnessed a disturbing pattern of deception and manipulation" by Sam Altman and Greg Brockman during their time at OpenAl, where they were completely immersed in their pursuit of achieving AGI.
"Many of us were hopeful about OpenAI's success and viewed Sam and Greg with suspicion. The duo's behavior was increasingly concerning, and anyone who raised objections or disagreed with them was either silenced or driven away. This systematic authoritarian oppression within the organization resulted in an environment of fear that effectively stifled any discussion of the ethical implications of OpenAI research."
The authors also provided detailed examples of these accusations:
Sam Altman asked researchers to delay reporting on the progress of a specific "secret" research project, only to then cancel it on the grounds that results were not being produced soon enough. Anyone who questioned this practice was deemed a "culture misfit" and even fired, some as recently as Thanksgiving 2019.
Greg Brockman used discriminatory language against a team member who was transitioning. Despite repeated promises to address the issue, Brockman failed to take any meaningful action and simply avoided any communication with the affected employee, thereby creating a hostile work environment. The team member was eventually fired after being accused of poor performance.
Sam Altman directed information technology and operations staff to investigate employees without management's knowledge or consent.
Sam Altman, who has carefully used OpenAI's nonprofit resources to advance his own personal goals, holds a grudge against Musk, especially after their falling out.
The operations team acquiesced to special rules that applied to Greg Brockman and went through complex requirements to avoid being blacklisted.
BradLightcap failed to deliver on its promise to release documents detailing OpenAl’s profit cap structure and the profit cap for each investor.
Sam's calculation quotas for research projects were inconsistent, causing internal distrust and infighting.
The authors of the letter emphasized that OpenAl's work is "too important to be compromised for the personal purposes of a few people," and they asked the board to launch an independent investigation into Sam Altman and Greg Brockman. “We implore you, the Board, to remain steadfast in your commitment to OpenAI’s original mission and not to succumb to profit-driven pressure. The future of AI and the well-being of humanity depends on your unwavering commitment to ethical leadership and transparency,” the employees urged.
However, there is a question mark over whether the board of directors can take on this important task, because there are also internal divisions and discord within them.
Today, OpenAI is in danger because of this chaos, and the service has become unstable. But for opponents, this is an excellent pursuit opportunity.
ChatGPT suffered an outage this morning
Competitor Claude upgrades
Before being fired by the board of directors, Sam Altman said that in order to ensure a good user experience for each user, the registration of ChatGPTPlus accounts needs to be suspended for a period of time. The reason for this is that since the OpenAI Developer Days, the traffic surge has far exceeded its capacity.
Recently, OpenAI's board of directors dismissed Sam Altman and a series of subsequent incidents, ChatGPTPlus users have become "fish in the pond".
This morning, ChatGPT was down for a while, and the prompt that popped up was "We are experiencing unusually high demand. Please be patient, we are working hard to expand our system."
Currently, ChatGPT has returned to normal use.
In addition, after several days of silence, OpenAI finally made its voice on social platform X. This time, the voice version of ChatGPT is open to all free users, and the App can be downloaded on mobile phones. When in use, click the headset icon to start a conversation. This seems to be a move to stabilize users and show the outside world that OpenAI is still functioning normally.
While OpenAI is busy "fighting", its main competitor Anthropic upgraded Claude to version 2.1.
The biggest update of Claude2.1 this time is to support 200K context windows, which is equivalent to 150,000 words or more than 500 pages of text material. Now users can upload complete code bases, technical documents such as financial statements (such as S-1s), and full-length literary works such as The Iliad and The Odyssey.
Claude2.1 has also made great progress in hallucination, and the hallucination rate is only half of that of Claude2.0. Improvements have also been made in understanding and summarizing, especially for complex documents that require high accuracy, such as legal documents, financial reports, and technical specifications. Evaluation shows that Claude2.1 reduces the probability of giving wrong answers by 30%.
OpenAI's chaos seems to be continuing, and competitors can take advantage of the opportunity to develop.
Related reports:
"Sudden!" OpenAI CEO Sam Altman was fired and resigned from the board of directors》
"Behind the ouster of Sam Altman: Chief scientist Ilya pushed, senior employees resigned, and competition began to grab people"
"48 hours after being fired, "visitor" Sam Altman returns to OpenAI headquarters, Musk calls for disclosure of the truth"
"Sam Altman wants to join Microsoft, but the promoter Ilya regrets it. More than 500 employees resigned and forced the board of directors to go to the board of directors"
"OpenAI's board of directors was kicked out and had discussed a merger with competitor Anthropic, and there are still variables whether Altman will stay or not"
Reference links:
https://www.timesnownews.com/world/elon-musk-shares-openais-letter-to-board-of-directors-concerns-worth-investigating-article-105397446