IBM unveils major Watson, AI & open source innovations

At its annual THINK conference, IBM announced significant enhancements to its watsonx platform, unveiling updates and new data and automation capabilities designed to make artificial intelligence (AI) more open, cost-effective, and flexible for businesses.

Among the key updates, IBM has open-sourced a family of its most advanced and high-performance language and code Granite models. The Granite models, available under Apache 2.0 licenses on platforms like Hugging Face and GitHub, range from 3 billion to 34 billion parameters. These models, trained across 116 programming languages, excel in various tasks such as code generation, bug fixing, and application modernisation.

IBM has also launched InstructLab in collaboration with Red Hat. This innovative capability allows for continuous development of base models through incremental contributions, similar to open-source software development. InstructLab enables developers to build models tailored to their business needs, adding direct value from AI innovations.

In his keynote, IBM CEO Arvind Krishna highlighted the company's commitment to open-source AI innovation, stating, “We want to use the power of open source to do with AI what was successfully done with Linux and OpenShift.” Krishna emphasised that open innovation brings choice, transparency, and collaboration to the AI landscape.

IBM also introduced a new AI-powered tool, IBM Concert, which will be available from June 2024. IBM Concert provides generative AI-driven insights across clients' application portfolios, helping to identify, predict, and address problems before they escalate. By integrating into existing systems and using AI to connect with data from various sources, it offers a comprehensive view of connected applications, making teams more effective and proactive.

The THINK conference served as a platform for IBM to reveal updates to its watsonx assistants. These include the upcoming watsonx Code Assistant for Enterprise Java Applications and watsonx Assistant for Z, aimed at transforming user interaction and knowledge transfer within systems. Additionally, enhancements to watsonx Orchestrate will aid clients in building their own AI assistants.

IBM’s recent annual Global AI Adoption Index indicated that while 42% of enterprise-scale companies have implemented AI, 40% of those experimenting with AI have yet to deploy their models. To overcome barriers such as the skills gap and data complexity, IBM's new AI innovations aim to simplify and accelerate AI adoption.

Further bolstering its AI capabilities, IBM announced new and upcoming generative AI-powered data products. These include IBM Data Product Hub and Data Gate for watsonx, both set for release in June 2024. These products will help organisations manage and optimise complex data for AI workloads.

In addition, IBM is expanding its NVIDIA GPU offerings to support AI and other mission-critical workloads. New GPU options include NVIDIA L40S and L4 Tensor Core GPUs, alongside support for Red Hat Enterprise Linux AI and OpenShift AI.

IBM's THINK conference also underscored the company's collaborations with major industry players such as AWS, Adobe, Meta, Microsoft, Mistral AI, Palo Alto Networks, SAP, Salesforce, and the Saudi Data and Artificial Intelligence Authority. These partnerships aim to expand AI capabilities, offer model choice, and enhance governance through the watsonx platform.

Krishna concluded by reiterating IBM’s dedication to advancing AI through open-source contributions and strategic partnerships, aiming to drive enterprise AI at scale and unlock significant economic benefits across industries.