5 AI trends we spotted at Google Cloud Next

The Tabnine team recently participated in Google Cloud Next in Las Vegas, where they showcased their latest features amidst a focus on the transformative power of AI. Attendees were introduced to Tabnine's efficient AI models, designed specifically for code generation, which can be deployed at a lower cost compared to other larger models, and offer options for self-hosting to enhance security. The event highlighted key trends in AI, such as the challenge of running Generative AI models due to their computing demands, and the importance of data protection, as exemplified by Tabnine's use of permissively licensed open-source code for training. AI agents, which perform tasks and make decisions akin to human interaction, were a focal point, with Tabnine introducing Chat agents to aid software development. The proliferation of specialized AI models for software development was noted, with Tabnine offering switchable models to optimize performance for specific tasks. Additionally, the company addressed AI hallucinations by employing techniques like retrieval-augmented generation to improve the accuracy of AI recommendations, alongside offering model customization to enhance performance based on user-provided training data.