aifollow.news
返回 量子位 AI 报道
量子位 AI 报道· · 原发布时间 AI 评分20

奥特曼谈黄仁勋离职:称无法判断其是否应重返英伟达

AI 辅助摘要

在11月16日的Semafor硅谷峰会上,OpenAI CEO山姆·奥特曼谈及黄仁勋离职,表示对其处境有“强烈的个人共情”,但因掌握的信息不足,无法判断他是否应重返英伟达。奥特曼还透露,OpenAI已用AI加速模型研发,但尚未启动无人监督的递归自我改进。

正文 · AI 翻译

Altman Speaks for the First Time About Jensen Huang's Departure

“Heroes have flaws, too.”

“Heroes have flaws, too.”

As the controversy over Jensen Huang's departure continues to grow, Altman has finally spoken publicly.

On November 16 local time, OpenAI CEO Sam Altman spoke publicly for the first time about the controversy over Jensen Huang's departure at a Silicon Valley summit hosted by Semafor.

Altman also made no attempt to hide his admiration for Huang at the event: “Jensen Huang has always been a hero to me.”

Altman also shared his fuller views on the controversy surrounding Huang's departure and the future of the AI industry as a whole:

Altman acknowledged that heroes are not perfect.

Huang has always been a hero to Altman, and it is hard to imagine the AI revolution without him.

Huang's departure will not stop the AI industry as a whole from moving forward.

OpenAI and Nvidia established a mutually supportive relationship early on, and each placed a bet on the other at an early stage.

As for whether Huang should return to Nvidia, Altman said he did not have enough information to make a judgment.

3 years ago, Altman himself went through a board ouster that rocked Silicon Valley.

Now that it is Huang's turn, Altman can clearly understand what that feels like better than most people.

Rewind a few days.

On November 12, Nvidia founder Huang, who had led the company for 33 years, suddenly announced his resignation as CEO and from the board.

Following the news, Nvidia's share price fell as much as approximately 15% within a week, wiping more than $570 billion off its market value.

Subsequently, several Silicon Valley executives, including Elon Musk, signed an open letter, while 180,000 people signed a petition urging Nvidia's board to reinstate Huang.

Many observers, however, noticed the wording of the open letter supporting Huang—

it was virtually identical to the open letter in which OpenAI employees pressured the board to bring Altman back in 2023.

At the Semafor summit, moderator Reed Albergotti put the question directly to Altman.

After all, in November 3 years ago, Altman was also abruptly fired by OpenAI's board, only to make a dramatic comeback 4 days later.

In response, Altman admitted that he felt “strong personal empathy” for Huang's current situation.

He recalled that Nvidia had bet on OpenAI early on, and OpenAI had likewise bet on Huang.

In Altman's view, it is hard to imagine what today's AI revolution would look like without Huang and the Nvidia he built.

But personal feelings are one thing; business is another.

Asked whether Nvidia's business commitments still carried the same weight after Huang's departure, Altman gave a highly pragmatic answer.

Over the past few years, Nvidia has been not only OpenAI's largest chip supplier but also one of its most important financial backers.

Just recently, Nvidia agreed to provide up to $105 billion in financial guarantees for OpenAI's data center project in Ohio.

Statistics show that Nvidia's investments and commitments across the AI ecosystem now total more than $100 billion, covering major players including CoreWeave, Nebius, and Anthropic.

There is widespread concern that if Chief Financial Officer Colette Kress becomes interim CEO, Nvidia may scale back its activities, affecting liquidity across the AI industry.

Altman does not share that pessimistic view.

He pointed out that Nvidia is now a giant company of considerable size, and people still have great confidence in its ability to honor its guarantees.

As for whether the AI industry as a whole will lose momentum without Huang, Altman is quite optimistic.

He believes demand in the AI market is now strong, AI's value has been proven, more and more people are actually using AI, and industry revenue and market demand are growing strongly.

Therefore, even if Huang leaves, the AI industry as a whole can continue to move forward.

Altman also stressed that Huang's extraordinary ambition has played an exceptionally distinctive and valuable role in advancing the AI industry.

Turning to where the next wave of AI breakthroughs might occur, Altman pointed to scientific research.

He revealed that over the past few months, AI has begun to show its potential to drive scientific discoveries in fields such as mathematics and physics, with some exciting advances emerging.

In his view, this is only the beginning.

As models become more capable, AI may take part in complex engineering tasks such as chip design, helping people solve technical problems that previously required extensive work by experts.

At the same time, the way companies work internally will also change:

A great deal of repetitive, process driven work may gradually be handed over to AI.

From scientific discovery and engineering design to everyday business operations, Altman expects AI to take on increasingly complex practical work and thereby create new economic value.

Besides discussing his longtime partner Huang, Altman also addressed several recent points of public criticism head on during the interview.

First were his remarks about RAM sticks, which recently angered PC gamers.

At the beginning of this month, global DRAM prices surged under pressure from demand for AI computing power, at one point rising by 5 to 7 times.

Altman had said in an interview that it was rather bad to see scarce RAM sitting in ordinary people's personal computers when it could be more useful in AI data centers.

Those remarks quickly drew widespread mockery in online communities overseas, and many PC builders even nicknamed him “Ram Altman.”

At the summit, Altman took the opportunity to clarify that his earlier remarks had been misunderstood.

He changed his tune, saying—

that he actually hopes everyone can afford, and should install, plenty of RAM in their own computers, and that he strongly supports people tinkering with hardware at home.

Next were the practical issues of double counting demand for AI chips and the difficulty of getting data centers built.

Earlier reports said that Altman had told Samsung executives there was a bubble caused by double counting chip demand in the current market.

Altman explained the reasoning behind this at the event.

He said many companies are now developing AI chips, and each supplier hopes to sell most of its capacity to a handful of leading AI model companies, such as OpenAI and Anthropic.

From the buyers' perspective, however, OpenAI does not need to buy up every supplier's chips, resulting in overlapping expectations along the upstream supply chain.

He stressed, though, that the actual need to build data centers remains very real.

Altman acknowledged that the old approach of building data centers near residential areas no longer works, and that developers will have to build them far from population centers in the future.

Altman also discussed the latest developments in recursive self improvement (RSI) within OpenAI.

Altman revealed that OpenAI has begun using AI technology to accelerate development of its own models.

However, he made clear that OpenAI has not yet begun fully autonomous RSI, in which AI continuously improves its own capabilities without human oversight.

The reason is that it is not yet possible to ensure the process is sufficiently safe.

As AI capabilities continue to improve, ensuring that humans always retain control will also become an important problem for OpenAI to solve.

Finally, Altman also gave his latest comments on capital market developments drawing public attention.

This week, competitor Anthropic just began its closely watched IPO process.

Some analysts worry that Anthropic's listing will draw substantial Wall Street funding away early, potentially leaving less room for OpenAI to raise capital.

Altman appeared quite unruffled by this—

he believes the reality in private markets over the past year has been that investor demand far exceeded the supply of quality assets, and he is not at all worried about a lack of financial support in the future.

Moreover, giving more ordinary people the opportunity to invest in OpenAI and share in the economic benefits of AI development is one of the important reasons for taking the company public.

At the same time, he acknowledged that becoming a public company does have a number of practical drawbacks at a time when the industry is highly volatile.

Thus, although OpenAI will certainly enter the public markets at some point in the future, it is in no hurry to do so now.

Reference link:

[1]https://www.youtube.com/watch?v=oi-E6xB4wSU&t=39s

发现内容有误?提交纠错