Listen to the audio version of the article
There is also Oracle in the race for generative artificial intelligence. Larry Ellisson’s giant was not steady on his legs. The choice was to integrate the large language model (Llm) of the Canadian Cohere with that of Meta, Llama 2 throughout its technological products, both in terms of cloud computing and in on-premises solutions.
«Generative AI, as underlined on various occasions by our founder CTO of Oracle (Larry Ellison ed.) has disruptive potential», observes Giovanni Ravasio, vp and country leader cloud applications, Italy and Iberia, Oracle. «It is capable of revolutionizing many sectors of activity and improving the human condition. Without fear of exaggeration, the “GenAI revolution” is the greatest opportunity since the advent of the cloud. And Oracle is one of the cloud vendors best positioned to help companies seize these opportunities.”
Oracle’s strategy, according to the manager, is based on three elements: providing a solid infrastructure for training and supplying large-scale GenAI models; offer easy-to-use generative AI services for developers and data scientists and integrate these models into their software-as-a-service (SaaS) portfolio. For the first point, they will make superclusters available to customers on Oralce Cloud Infrastructure (Oci), powered by the most recent GPUs in the sector, connected together at very low latency for remote access to data memory (RDMA) on a convergent Ethernet network (RoCE ). This allows Oci to train GenAI models twice as fast as other clouds and at significantly lower cost. As for large languages, the choice, as mentioned, was to rely on Cohere, a company specialized in enterprise-level LLM models. Given the investments and the names on the field, the impression is that AI is a game for the few. Exactly as happened with cloud computing. Indeed, in the case of Ai, the club is even more exclusive.
«Most likely yes – replies Ravasio -. The investments in research and development required by the very rapid evolution in the field of GenAI and LLM, as well as the enormous infrastructure resources for model training, are within the reach of only a few large vendors.”
Gen AI will profoundly change the way businesses deliver value to their customers and increase productivity. «The SaaS architecture of our applications – he adds – allows us to add new use cases to each product release with practically no repercussions for client companies, so that the latter can increase productivity, reduce costs, improve experience of their employees and their customers and continuously optimize business processes. The real news is that just a few months after the word entered our consumer vocabulary, GenAI is already available in our business applications: we have announced more than 50 GenAI use cases in Oracle Cloud Applications as of last June 2023, which are made available through the quarterly updates that customers automatically receive.”