终极三国

PLA Southern Theater Command tracks, monitors and warns 3 Philippine aircraft after their illegal entry into Huangyan Dao airspace_我的网站

汽车之家

A | 　　OpenAI首席执行官萨姆·奥尔特曼（Sam Altman）周日表示，他担心人工智能（AI）有朝一日会被少数公司、模型或个人控制，使绝大多数消费者对这项技术如何发展失去话语权。　　当天，他在一档播客节目中说道：“正确的做法是，我们希望人们能够深度掌控未来。社会和各种模型将共同演化。

An aerial drone photo taken on May 2, 2026 shows reefs in the territorial waters of China's Huangyan Dao in the South China Sea. (Xinhua/Chen Bin)The PLA Southern Theater Command said on Friday that three Philippine aircraft illegally entered China's territorial airspace over Huangyan Dao without approval, prompting Chinese naval and air forces to track, monitor and issue verbal warnings to the aircraft.
On Friday, three small Philippine aircraft - an Islander, a C-208, and a PA-31T - illegally intruded into China's territorial airspace over Huangyan Dao without approval from the Chinese government. The naval and air forces dispatched by the PLA Southern Theater Command (STC) tracked and monitored the Philippine aircraft, and issued verbal warnings to them in accordance with laws and regulations, senior Colonel Li Jianjian, spokesperson for the PLA STC Air Force, said in a statement on Friday.
The acts of the Philippine side seriously violate China's sovereignty, and are highly likely to cause misunderstanding and miscalculation. "We sternly warn the Philippine side to stop rights-violation and provocative acts immediately. Forces of the STC remain on high alert at all times to firmly safeguard national sovereignty and security as well as regional peace and stability," said the spokesperson.
Ding Duo, director of the Research Center for International and Regional Issues at the National Institute for South China Sea Studies, called the Philippines' latest move another "stunt."
The Philippines previously submitted its so-called "territorial sea baselines" around Huangyan Dao, so it is not unexpected that Manila would follow up with actions at sea or in the air to stir up tensions, keep the issue in the spotlight and stage further provocations, said Ding.
At present, the Philippines can barely gain access to the relevant waters and has secured no tangible advantage there. Aircraft, however, can enter and leave the area quickly. Such operations not only provide material for media hype but also generate footage of China's response, allowing Manila to project resolve and reassure its military and law-enforcement forces at home, Ding told the Global Times.
In late July, the Philippines notified the United Nations of the so-called "baselines and maritime zones" surrounding Huangyan Dao. On July 31, the Chinese Foreign Ministry issued a statement on the Philippines' delimitation of the so-called "baselines of the territorial sea" around Huangyan Dao, calling the move "a grave infringement upon China's territorial sovereignty and a serious violation of international law."
"It is illegal, null and void. China firmly rejects and will never accept it," the statement said.
Meanwhile, Manila appears emboldened by growing security and defense cooperation with the US and Japan. Backed by these external partners, the Philippines has shown greater willingness to stage provocations, while its military and law-enforcement forces have displayed increasingly risk-prone and opportunistic behavior, said Ding.
China has exercised considerable restraint over the Huangyan Dao, but the Philippines should not mistake that restraint for weakness. If China were to employ the full range of available measures, Philippine aircraft would likewise be denied access to the area, said Ding.
By dispatching small aircraft for deliberate provocations, the Philippines is further increasing the risk of miscalculation. Such aerial incursions are highly unpredictable and carry serious escalation risks. If China responds with resolute or more forceful measures to defend its rights and an unintended incident occurs as a result, the Philippines should bear full responsibility for the consequences, the expert warned.
。”　　奥尔特曼表示，他担心的另一种可能性是人工智能摆脱人类的控制。矛盾的是，他还认为，对后一种可能性（AI摆脱人类控制）的恐惧，可能会导致前一种可能性（即AI被少数人控制）成为现实。

B | 　　“有很多人对于这些风险的严重程度感到非常紧张，并被这种担忧深深影响，并觉得自己有必要保护世界。比如，他们会认为：‘我们应该为了安全而牺牲很多自由。’”他说。　　奥尔特曼这番话很可能是在暗指Anthropic首席执行官达里奥·阿莫迪（Dario Amodei）。　　阿莫迪曾在播客、文章和公司博客中多次警告AI快速发展所带来的危险。

C | 他在6月的一篇文章中写道，该公司的新Mythos模型对网络安全、金融行业、关键基础设施和国家安全构成了“非常现实的风险”。

D | 　　阿莫迪利用这些警告，呼吁政府在人工智能监管方面发挥更大作用。在奥尔特曼看来，这可能是以自由换取安全。　　上个月，Anthropic也是唯一一家没有签署支持开放权重模型的信函的主要人工智能公司。

E | 开放权重模型允许开发者在自己的基础设施上共享、下载、修改和部署AI模型。　　中国已积极拥抱开放权重模型，这使得其能够以极低的成本跟上硅谷的发展步伐。美国国内的一些力量，包括英伟达、 Meta 和微软，则认为美国也需要采取同样的做法，否则就有可能在AI竞赛中将阵地拱手让给竞争对手。　　虽然OpenAI尚未发布其前沿模型的源代码，但该公司去年确实发布了两款开放权重模型：gpt-oss-120b和gpt-oss-20b。