L’generative artificial intelligence (AI). it could revolutionize the software development models known so far. In fact, it has already undermined Moore’s law that the number of transistors in processors tends to double every 18 months. Today, in fact, the computational resources needed for the largest AI calculations double every 6 months. For the world of developers and for that of ISV (Independent Software Vendor) the arrival of ChatGPT it coincided with the possibility of simplifying some typical coding tasks. But what lies ahead now is a totally different way to imagine new products and to change the very way of making them. He is thinking about paving the way for this further phase Google which he presented these days API PaLM e MakerSuite.
What Google’s PalM and MakerSuite API consist of
These are systems with generative AI functions integrated in Google Cloud and in Google Workspace, the well-known corporate collaboration suite of the Mountain View giant. Available for now only for the benefit of a few selected developers, who will be able to test them in preview, the solutions aim to provide additional capabilities to facilitate the development cycle.
Contact Center as-a-Service: discover innovation at the service of customer experience
PaLM API allows you to experiment with large language models from Google such as LaMDA (Language Model for Dialogue Applications). Although less known to the general public than ChatGPT, it uses the same neural network architecture, ie Transformer. As the gateway to language models such as LaMDA, the PaLM API provides developers with the opportunity to use it for content generation, chat, summary and classification. For example, through the so-called embedding (dense numerical representations of real-world objects and relationships expressed in vector form) generated through the PaLM API, developers will be able to create applications with their own data or with external data sources and open-source libraries.
According to Chirag Dekate, an analyst at Gartner, this solution can enable tech companies to “start building their own versions of the generative AI model”, adapting it to the specific industrial context in which they operate. Similarly, MakerSuite presents itself as a tool to simplify the workflow which is often fragmented due to the use of different tools to carry out activities such as creating and iterating a prompt, generating synthetic data and fine-tuning a custom model. All directly in the browser, with the advantage that, in addition to simplifying these processes, it is possible to export the prompts as code in languages and frameworks such as Python and Node.js.
The other generative AI solutions for developers
Google’s offer on the generative AI side for developers also provides for the provision of basic models – initially for the generation of text and images and later for audio and video – on the platform Vertex AI. Google Cloud’s machine learning platform can help data scientists solve use cases, choose from a variety of models, or get recommendations on how to best use corporate data while maintaining control over costs and governance. Simultaneously, the multinational led by Sundar Pichai launched Generative AI App Buildera tool that connects conversational AI flows with research experiences and ready-made building blocks.
“Google Cloud is bringing decades of research, innovation and investment to the world of AI with the launch of generative AI support in Vertex AI and the Generative AI App Builder,” said Ritu Jyoti, Group Vice President, Worldwide Artificial Intelligence. and Automation Research from IDC -. In this way, Google Cloud is ready to enable a whole new generation of builders, innovators and developers to harness the power of artificial intelligence in different ways”. For his part, Mike Gualtieri, an analyst at Forrester Research, declared that the Californian company “had been preparing for this for a long time and I believe that the competition with Microsoft and OpenAI forced them to release these solutions earlier”.
The challenge (also on princes) deartificial intelligence
And that it is a challenge on all fronts of generative AI can also be deduced from the new features added to Google Workspace. Until now, those based on artificial intelligence allowed the use of “smart compose” within Gmail or to automatically create summaries in Google Docs. Now, a small group of testers are testing an upgrade where, in both Gmail and Docs, it is possible type the topic you want to write about and immediately receive a tailored draft. Starting from this initial draft, it will be possible to elaborate or shorten the message, change the tone, contextualize the contents and so on. The decision not to release these features immediately for the benefit of all, as OpenAI did with ChatGPT, derives from the principles that Google places at the center of its approach to artificial intelligence.
Indeed, the first of these principles states that “in considering the potential development and use of artificial intelligence technologies, we will consider a wide range of social and economic factors and proceed where we believe that the probable overall benefits substantially outweigh the foreseeable risks and disadvantages. AI also improves our ability to understand the meaning of content at scale. We will strive to make high quality and accurate information readily available using AI, while continuing to comply with cultural, social and legal norms in the countries in which we operate. And we will continue to carefully evaluate when to make our technologies available on a non-commercial basis.”
Read more here: Source link