据机器之心报道，一份关于OpenAI在2027年实现人工通用智能（AGI）的53页PDF文档正在互联网上广泛传播。这份文档称，OpenAI计划在2022年8月开始训练一个拥有125万亿参数的多模态模型，并在2023年12月完成了训练。然而，由于推理成本过高，该模型的发布被取消。据称，该模型最初被命名为GPT-5，但取消后更名为GPT-5，而之前曝光的神秘项目Q*的下一阶段最初被命名为GPT-6，但已更名为GPT-7。
这份文档的来源是一个名为vancouver1717的X账户，该账户于2023年7月注册，只有两条推文，因此文档的真实性仍然存在疑问。虽然有人对文档内容持怀疑态度，认为其中的判断缺乏专业性，但这份文档的存在仍引起了一些关注。
OpenAI是一家以人工智能研究和开发为主的公司，其目标是开发出能够在各种任务上与人类相媲美甚至超越人类的人工智能系统。然而，AGI的实现一直是一个具有挑战性的目标，迄今为止还没有实现。
尽管OpenAI在过去几年取得了一些重要的研究成果，如GPT-3模型的发布，但要在2027年前实现人类水平的AGI仍然面临巨大的技术和理论挑战。目前，关于OpenAI是否真的在2023年完成了训练以及他们是否能够如期实现AGI的细节尚不清楚。
对于这份文档的真实性，专家们普遍持保留态度，他们认为需要更多的证据和确认才能确定其可信度。不过，这份文档的流传引起了人们对OpenAI未来发展的关注和期待。
OpenAI目前在人工智能领域的研究和创新备受瞩目，他们的成果对于推动人工智能技术的发展具有重要意义。无论这份文档的真实性如何，OpenAI都将继续致力于推动人工智能的发展，为人类创造更多的机遇和进步。
英语如下：
News Title: OpenAI plans to achieve AGI in 2027? The truth revealed: Professionals question its authenticity
Keywords: OpenAI AGI, GPT-5, GPT-7
News Content: According to Synced, a 53-page PDF document regarding OpenAI’s achievement of Artificial General Intelligence (AGI) in 2027 is widely circulating on the internet. The document claims that OpenAI plans to start training a multimodal model with 1.25 quadrillion parameters in August 2022 and completed the training in December 2023. However, due to high inference costs, the release of this model was canceled. It was initially named GPT-5 but was renamed GPT-5 after the cancellation. The next stage of the previously exposed mysterious project Q*, initially named GPT-6, has been renamed GPT-7.
The source of this document is an X account named vancouver1717, which was registered in July 2023 and only has two tweets. Therefore, the authenticity of the document remains questionable. While some doubt the expertise of the document’s content, its existence has still attracted some attention.
OpenAI is a company focused on artificial intelligence research and development, aiming to develop artificial intelligence systems that can match or surpass humans in various tasks. However, achieving AGI has always been a challenging goal and has not been accomplished so far.
Although OpenAI has achieved some significant research milestones in the past few years, such as the release of the GPT-3 model, achieving human-level AGI by 2027 still faces significant technological and theoretical challenges. Currently, details about whether OpenAI actually completed the training in 2023 and whether they can achieve AGI as planned are unclear.
Experts generally maintain a reserved attitude towards the authenticity of this document, believing that more evidence and confirmation are needed to determine its credibility. However, the circulation of this document has sparked people’s interest and expectations for OpenAI’s future development.
OpenAI’s research and innovation in the field of artificial intelligence are highly regarded, and their achievements have significant implications for advancing AI technology. Regardless of the authenticity of this document, OpenAI will continue to dedicate itself to promoting the development of artificial intelligence and creating more opportunities and progress for humanity.
【来源】https://mp.weixin.qq.com/s/vSud2JN3W3xe0AzC8pwuUg
Views: 1