On Tuesday, Google introduced a wide range of new artificial intelligence products and collaborations that are intended to increase the use of the developing technology by big organisations.
New clients for its cloud software, such General Motors and Estee Lauder Companies, were among the announcements made at Google Next in San Francisco.
The Alphabet subsidiary showcased a new version of its custom-built AI processors, as well as solutions for security and its office suite. It also unveiled an enterprise-scale application to watermark and identify photos created using AI.
The sudden spate of announcements is a result of Google's recent push to highlight its AI initiatives after Microsoft surprised it with an ambitious AI strategy that it has been implementing since last year.
Thomas Kurian, the head of Google Cloud, told Reuters that its major business clients must be thoughtful and operate at a different speed.
"We've generally told enterprise customers, 'Go slowly and methodically because it's important that you treat this as a strategic software development,'" he said. "There's been this sort of FOMO of, 'I need to be in generative AI for generative AI's sake.'"
Fear of missing out, or FOMO, has become a recurrent theme in AI in recent months.
Google increased the number of its AI models in its collection from 20 to 100 in order to strengthen its enterprise cloud service. The AI infrastructure comprises agreements that provide Google Cloud users access to the AI model LLaMa 2 from Meta Platforms and Claude 2 from the firm Anthropic.
New versions of Google's own foundational AI infrastructure have been released, and they perform better and have more functionality. For instance, the PaLM text model's new iteration boosted the amount of text users could enter, making it simpler to handle larger materials like books and court papers.
A technique that makes it possible to watermark AI-generated photos was discussed by Google. The technique, known as SynthID, modifies a digital image file in a way that is imperceptible to the human eye. It is intended to hold up even after a picture has been changed or tampered with.
Google updated its office software and security tools with AI as well. The business presented a tool that uses artificial intelligence (AI) to migrate databases from Oracle (ORCL.N) to an open-source version, a famously challenging job.
Google has made a version of its fifth-generation tensor processing unit (TPU) that is suitable for genAI and big language models accessible in advance of the official unveiling of the full-featured model.
The brand-new TPU v5e chip is made to efficiently feed content from huge models while simultaneously training big models. It is not as potent as the fifth AI generation flagship processor, which has not yet been released.
Google has put the TPU v5e chips together in groups of 256 to create what it calls a "supercomputer." Customers of the cloud can join multiple pods together to address increasingly challenging computing issues.
(Source:www.usnews.com)
New clients for its cloud software, such General Motors and Estee Lauder Companies, were among the announcements made at Google Next in San Francisco.
The Alphabet subsidiary showcased a new version of its custom-built AI processors, as well as solutions for security and its office suite. It also unveiled an enterprise-scale application to watermark and identify photos created using AI.
The sudden spate of announcements is a result of Google's recent push to highlight its AI initiatives after Microsoft surprised it with an ambitious AI strategy that it has been implementing since last year.
Thomas Kurian, the head of Google Cloud, told Reuters that its major business clients must be thoughtful and operate at a different speed.
"We've generally told enterprise customers, 'Go slowly and methodically because it's important that you treat this as a strategic software development,'" he said. "There's been this sort of FOMO of, 'I need to be in generative AI for generative AI's sake.'"
Fear of missing out, or FOMO, has become a recurrent theme in AI in recent months.
Google increased the number of its AI models in its collection from 20 to 100 in order to strengthen its enterprise cloud service. The AI infrastructure comprises agreements that provide Google Cloud users access to the AI model LLaMa 2 from Meta Platforms and Claude 2 from the firm Anthropic.
New versions of Google's own foundational AI infrastructure have been released, and they perform better and have more functionality. For instance, the PaLM text model's new iteration boosted the amount of text users could enter, making it simpler to handle larger materials like books and court papers.
A technique that makes it possible to watermark AI-generated photos was discussed by Google. The technique, known as SynthID, modifies a digital image file in a way that is imperceptible to the human eye. It is intended to hold up even after a picture has been changed or tampered with.
Google updated its office software and security tools with AI as well. The business presented a tool that uses artificial intelligence (AI) to migrate databases from Oracle (ORCL.N) to an open-source version, a famously challenging job.
Google has made a version of its fifth-generation tensor processing unit (TPU) that is suitable for genAI and big language models accessible in advance of the official unveiling of the full-featured model.
The brand-new TPU v5e chip is made to efficiently feed content from huge models while simultaneously training big models. It is not as potent as the fifth AI generation flagship processor, which has not yet been released.
Google has put the TPU v5e chips together in groups of 256 to create what it calls a "supercomputer." Customers of the cloud can join multiple pods together to address increasingly challenging computing issues.
(Source:www.usnews.com)
Homepage
Send to a friend
Printable version
Share
