Google jostles with AI rivals as it releases a large language model API and other developer tools.
Google Cloud opened some of its AI-powered applications to developers on Tuesday, introducing the Pathways Language Model, or PaLM, API for language models and the Makersuite prototyping tool within it. The developer tools go hand-in-hand with Google’s new AI productivity tools in Google Workspace, which are in testing now. The productivity AI can write emails, job descriptions or other business documents on request. For developers, there are a few new ways to get their hands on these tools.
Google opens gateway to PaLM API
PaLM is a large language model capable of responding to natural sounding prompts and returning sensible prose; the PaLM API lets developers build on top of it. In particular, developers can get an inside look at models optimized for multiturn use cases such as content generation and chat or more general-purpose models intended for summarization, classification and other tasks.
Google said in a blog post about this news that an “efficient model … in terms of size and capabilities” is available, with larger sizes expected “soon.” That efficient model is available only for select developers in the Private Preview at present, with a wait list for other interested parties expected to open before long.
What is Google’s Makersuite?
Makersuite is a low-code tool within the API that helps developers prototype ideas. Google plans to add features for prompt engineering, synthetic data generation and custom-model tuning going forward.
It’s also a way for business owners to experiment with possible use cases or tailor the API to their needs. Once finished, developers can export their prompt in common languages and frameworks including Python and Node.js.
Google wants to make sure developers can do a lot of work without leaving the tool.
Google’s Scott Huffman, vice president, engineering and Josh Woodward, senior director, product management, wrote in a blog post: “We learned a lot about generative AI development workflows and how fragmented they can be. Developers have to use different tools to accomplish tasks like crafting and iterating on a prompt, generating synthetic data, and tuning a custom model.”
In addition, Google emphasized that the API and tool are built in accordance with its AI safety standards. Training data accrued during non-Google developers’ use of the API won’t be fed back into the large language model and will be kept private to the company that uses it.
Other AI tools for Google Cloud enterprise customers
Google continues to try to claw into the natural language AI chat space currently dominated by Microsoft and OpenAI. Several new AI features for enterprise cloud customers are on the way, Google announced on Tuesday.
Generative AI App Builder is a tool designed to create apps that use conversational AI for whatever the user needs, connecting directly to Google’s out-of-the-box search capabilities and foundation models.
Google Cloud’s Vertex AI platform, which enterprises can use to build and deploy machine learning models and AI applications at scale, now has access to foundation models. This means enterprise customers can discover models, create and modify prompts and fine-tune those prompts with data from their own companies. For now, Vertex AI can create text and images; Google expects video and audio to follow.