news

openai is no longer open? company reorganization, top management shake-up, but it's not that bad

2024-09-26

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

openai is no longer open?

based on previous reports from many reliable foreign media and openai's actions in recent days, the above question may be replaced by a declarative sentence.

early this morning, reuters reported that openai plans to reorganize into a for-profit benefit corporation and will no longer be controlled by a non-profit board of directors. however, the plan is still under discussion with lawyers and shareholders, and the timetable for completing the reorganization remains uncertain.

openai's reorganized corporate structure will be similar to that of its old rivals anthropic and elon musk's xai. this is a special form of enterprise that aims to pursue economic interests while also promoting social responsibility and sustainable development.

the report said that openai's nonprofit organization will continue to exist and own a minority stake in the reorganized company.

sources said the reorganized openai could be worth $150 billion, depending on whether it can overturn its corporate structure and remove caps on investor returns.

reuters previously reported that removing the reward cap would require approval from openai’s nonprofit board, which includes altman, entrepreneur bret taylor and seven other members.

the massive round is seeing strong investor demand given openai’s rapid revenue growth and could be finalized within the next two weeks, the sources added.

existing investors thrive capital, khosla ventures, and microsoft are expected to participate, as are new investors including nvidia and apple. sequoia capital is also in talks to return.

reuters reported that openai's new round of financing is expected to come in the form of convertible notes. if the restructuring is unsuccessful, openai will need to renegotiate the valuation with investors and may convert to a lower number.

it is worth noting that openai’s helmsman sam altman will also acquire equity in the company for the first time.

it is not clear how much equity altman will get. as a billionaire, he once said, "i don't have any equity in openai. i do this work because i love it."

however, altman is not telling the whole truth; he does not own any equity in openai, but he does own equity in y combinator, which in turn owns equity in openai.

an openai spokesperson said:

we will continue to focus on building ai that benefits everyone, and we are working with our board to ensure we are in the best position to succeed in our mission. the nonprofit is core to our mission and will continue to exist.

in addition to the seismic restructuring of its corporate structure, openai is currently also facing a wave of executive departures.

since sam altman was dismissed last year, openai's internal turmoil seems to have never ended, and has frequently caused high-level earthquakes in recent times.

early this morning, openai's chief technology officer mira murati suddenly announced her resignation.

at openai's most important press conference this year, mira released the new model gpt-4o.

in early august, john schulman, co-founder and important veteran of openai, announced his resignation from openai. in addition to schulman's departure, openai president greg brockman is also taking a long vacation, and peter deng, the product director who joined last year, has also left.

a few hours after mira announced his resignation, openai chief research officer bob mcgrew and vice presidents of research barret zoph and mark chen also announced their resignation plans.

in response, altman hinted on the x platform that mira, bob, and barret made these decisions independently and amicably.note, there is an important point here, "independent and friendly".

he further said:

leadership changes are a normal part of any company, especially one that is growing so quickly and is so demanding.

but we’re not a normal company, and i think the reasons mira explained to me (that there’s never a good time, anything that’s not sudden will leak, and that she wanted to do this while openai was on the rise) make sense.

while the above two pieces of news are occupying the mainstream public opinion, another piece of news about openai's video generation model sora is also attracting widespread attention.

remember sora released in february this year?

according to foreign media the information, openai is training a new version of sora, hoping to generate higher quality and longer video clips, and the training requires collecting millions of hours of high-resolution video data in a variety of styles and themes.

previously, sora was revealed to have actual results that were below expectations, and it was still a long way from being deified by the fact that “reality no longer exists”.

the video generation speed is slow, initially it takes more than 10 minutes to generate a short video of about 1 minute;

difficult to use, filmmakers need to generate hundreds of clips to find something usable;

it’s hard to keep the style consistent, and objects and characters are hard to keep consistent across different edits;

there are errors in physics and anatomy, etc.;

the improved sora model will hopefully solve the above problems, which is also good news for creators and filmmakers.

with the wolves watching, openai seems to have always been eager to release demos in the past, but the products are not really ready for actual commercial implementation. this is also openai's old problem.

similar situations exist for chatgpt advanced voice assistant and searchgpt.

two weeks ago, zuckerberg said that meta prefers to release new products to get feedback rather than wait until the product is perfect before releasing it. putting aside the grudges of corporate competition, zuckerberg and altman can chat on this point.

at a time when the domestic ai video models are doing very well and are even overtaking the market, sora's reappearance will inevitably require more careful polishing, with both loud thunder and heavy rain. we also hope that we can see some eye-catching new changes by then.

in conclusion, if you think about it from another perspective, the frequent resignations at openai are not necessarily a bad thing that deserves criticism.

talent is the first resource for the development of the ai ​​industry. the recent popularity of flux vividly illustrates the classic truth that when you gather together, you are a ball of fire, and when you disperse, you are a sky full of stars.

batches of executives with mature ai r&d and management experience have left openai one after another, spreading like seeds throughout the ai ​​industry, better nourishing the industry's thriving growth.

the most typical benchmark is anthropic.

the founder of anthropic, who left openai, has paved a new path for later comers, and the company has recently been revealed to be negotiating a new round of financing, with a valuation expected to reach us$30-40 billion.

for users, if the ai ​​industry follows the cruel law of no winners, then users are undoubtedly standing in the position of winners, enjoying the benefits brought by technological involution.

for altman, although he has been controversial for deviating from the original intention of establishing openai, the major changes in the top management may also make this ai unicorn, which is on a strong upward momentum, more united and become a chariot that only obeys altman's will.

in other words, regardless of whether the path chosen is correct or not, isn't openai ushering in a new life?