IBM is making significant strides in the field of generative AI with new technologies and partnerships announced at the company’s Think 2024 conference. IBM has been involved in AI research for many years, but it was only recently, at Think 2023, that the company introduced its Watsonx genAI product platform. This platform has become the foundation of IBM’s efforts in generative AI, providing organizations with enterprise-grade models, governance, and tools.
At Think 2024, IBM is unveiling a series of its Granite models as open-source code. These models range from 3 to 34 billion parameters and include models for both code and language tasks. Additionally, IBM will be integrating Mistral AI models into its platform. The use cases for generative AI are varied, but one of the most common applications is building assistants. To support this, IBM is introducing new Watsonx assistants.
During a roundtable discussion with the press, IBM CEO Arvind Krishna expressed his belief that AI will have a transformative impact on society, comparable to that of the steam engine or the internet. This highlights the significant opportunities that AI presents for various industries and organizations.
One aspect that sets IBM apart in its AI efforts is its commitment to open source. IBM is making a group of its most advanced Granite models available under the open-source Apache license. Unlike other vendors who claim to have open models, IBM emphasizes the importance of using a bona fide open-source license. IBM wants to enable contributions and growth in the AI community, and real open source is essential for achieving this.
IBM is also expanding its Watsonx assistants to further advance enterprise AI. AI assistants provide a more consumable approach to implementing AI in different organizations. Watsonx code assistant for Java is aimed at helping developers write Java application code. Watsonx assistant for Z helps organizations manage IBM Z environments, and Watsonx Orchestrate enables enterprises to build their own assistants.
Retrieval Augmented Generation (RAG) is a common enterprise deployment pattern for generative AI. RAG empowers assistants and chatbots with real enterprise information that the models were not originally trained on. IBM recognizes the importance of RAG but does not see the need to build its own vector database. They have integrated with various options available in the market.
IBM also sees potential in the InstructLab technology recently announced by its Red Hat business unit. InstructLab enables continuous improvement of models in an optimized approach. This technology allows for better performance and accuracy in AI models.
One of the concerns surrounding AI is its impact on employment. However, IBM believes that gen AI will create more jobs than it will eliminate. Increased productivity leads to more business opportunities and, in turn, more employment. With the aging population in many countries, the need for AI capabilities becomes even more critical for economic growth and improving the overall quality of life.
In conclusion, IBM is continuing to push the boundaries of generative AI with its Watsonx genAI product platform and the introduction of open-source Granite models. By expanding its Watsonx assistants and exploring technologies like RAG and InstructLab, IBM is taking a comprehensive approach to enabling organizations to leverage the power of AI. With a focus on creating opportunities for collaboration and growth, IBM is positioning itself as a leader in the AI space.
Source link
