Google takes on OpenAI’s GPT-4, unveils PaLM API and MakerSuite

In April last year, Google introduced its own AI model named PaLM, an acronym for Pathways Language Model. With 540 billion parameters, Google said PaLM showed ‘breakthrough capabilities on numerous very difficult tasks’.

Similar to GPT or LaMDA, PaLM is able to generate and edit text and is being trained to be a conversational bot. It can also be used to summarise tasks or write code just like the new generative AI features for Workspace Apps like Gmail and Docs.

Now, the tech giant has launched an API for PaLM to take on the recently announced GPT 4 from OpenAI and new AI tools for enterprise users. Google says PaLM will help businesses generate text, photos, code, videos, audio and more using natural language prompts. To help developers with PaLM, Google has launched a new tool called MakerSuite that lets them quickly prototype ideas with plans to bring features for ‘prompt engineering, synthetic data generation and custom-model tuning.’

Google is also bringing generative AI capabilities to Google Cloud. This is will be useful for developers who want to customise or build their own apps or models using Google’s AI models. The company is providing foundation models for image and text generation for its Vertex AI platform, with plans to include features like the ability to ‘discover models, create and modify prompts, fine-tune them with their own data, and deploy applications’ in the near future.

If you own a business or work for the government and want to develop an AI-powered chat interface or digital assistant, the new Generative AI app builder might come in handy. The new tool can ‘connects conversational AI flows with out-of-the-box search experiences and foundation models’, thus enabling companies to build a generative AI application in a few hours or minutes.


Read more here: Source link