Google made waves this week with the announcement of major enhancements to its artificial intelligence portfolio. Not only did the tech giant unveil new AI models, tools, and partnerships for enterprises, but it also lifted the lid on a powerful custom chip purpose-built for advanced AI workloads.
The announcements came at Google’s annual Next conference, where executives pulled back the curtain on the company’s latest efforts to lead the AI revolution. Thomas Kurian, Chief Executive of Google Cloud, emphasized Google’s strategic approach, encouraging businesses to take their time implementing new technologies rather than rushing in purely for novelty’s sake.
Some key highlights include over 20 new AI models for Google Cloud, bringing the total offering to 100. Companies will gain access to cutting-edge language models from Meta and Anthropic via new partnerships as well. Google also strengthened its own foundational AI toolkit with improved performance across the board. Its iconic PaLM text model now handles way more words for vast document processing.
Additionally, SynthID was revealed – an invisible watermarking technique for images. The technology leaves a hidden digital imprint intact even after modifications.
Also, Google Office Suite is getting a new set of AI updates, such as the introduction of an AI-powered tool capable of porting a database from Oracle to an Open source Version.
New AI CHIPS
To power it all, Google unveiled the fifth generation of its Tensor Processing Unit chips. The new TPUv5e supercharges generative AI and giant language models on a scale previously unattainable. Multiple pods can link up like building blocks to solve the largest, most demanding AI problems.
With these innovations, Google aims to simplify and expand the use of transformative technologies for businesses worldwide. Executives claim it’s only the beginning – more is yet to come.
Source: Reuters
