IBM THINK 2024: Key announcements 

IBM THINK banner

United Kingdom, Jun 28, 2024

 At IBM THINK 2024, several key announcements were made that highlight IBM's ongoing innovations and partnerships across AI, hybrid cloud, and quantum computing, but above all else, the overarching message was that the Future of AI is Open.

IBM’s key announcements at THINK include:

  • The Future of AI is Open: The Open Sourcing of IBM’s 18 Granite LLMs (Large Language Model)
  • IBM Concert, the ‘nerve-centre’ of an organisation’s IT operations.  
  • InstructLab (in partnership with RedHat) a new AI training method for large language and code models.
  • AI Ecosystem partnerships with AWS, Meta, Nvidia, Azure, Mistral, SAP, Salesforce, and Adobe.
  • Quantum: no longer a science experiment.
  • Acquisitions: HashiCorp is the latest acquisition to strengthen IBM’s Automation portfolio.

The Future of AI is Open

IBM announced a significant milestone at their annual showcase event, THINK 2024 that emphasizes their commitment to open AI, with the open sourcing of their Granite LLM, available under an Apache 2.0 license. IBM capabilities and innovations are rooted in decades of pioneering work by IBM Research, and Forrester’s recognition of Granite models just 8 months after their release exemplifies IBM’s commitment to delivering high-performing and trusted LLM options to its customers. This approach contrasts with proprietary systems, aiming to foster innovation through shared research, open standards, and accessible technologies. The concept of Open AI promotes diverse contributions and accelerates advancements across industries.

The Granite models cover a range of applications, including language processing and geospatial analysis. IBM Research Director Darío Gil highlighted the importance of enterprises moving their data into foundation models that they own and control, stressing transparency and trust in AI model development.

The Granite LLM is a standout due to its advanced architecture and capabilities. It excels in natural language understanding and generation, making it highly effective for a wide range of applications. Granite's strengths include:

  1. Enhanced Accuracy: Fine-tuned on diverse datasets, Granite provides precise and contextually aware responses.
  2. Scalability: Efficiently handles large volumes of data and complex queries, suitable for enterprise-level applications.
  3. Adaptability: Easily customizable for specific industry needs, enhancing its utility across different sectors.
  4. Ethical AI: Incorporates robust ethical guidelines and bias mitigation techniques, ensuring responsible AI use.

This combination of accuracy, scalability, adaptability, and ethical considerations makes it a powerful tool in IBM's open AI ecosystem, driving innovation and practical solutions.

CEO Arvind Krishna also went on to say that there are three key elements needed to deliver the productivity that Gen AI promises:

  1. Trust the underlying artificial intelligence, which is why IBM teamed together with Meta to help form the Open AI Alliance.
  2. Flexibility  in combining models based on needs, from parties like IBM, from other commercial entities, and from open source.
  3. Safety is critical, which is one of the reasons why IBM have pushed to bring a lot of capabilities around AI inferencing and deployment into Red Hat Linux.

IBM Concert: Enhance AI Innovations with the power of Automation

A major focus of THINK was the launch of IBM Concert, a generative AI-powered tool designed to help enterprises manage and optimize their IT operations through automation. Concert leverages AI from IBM's watsonx platform to provide insights and predictive automation across applications, helping organizations streamline compliance processes and enhance operational efficiency. This tool aims to reduce the complexity of managing IT environments and is set to be generally available by June 2024. Simply put, this is next-generation Observability, the ‘nerve-centre’ of IT operations, powered by AI.

InstructLab with Red Hat

IBM and Red Hat co-launched InstructLab—a collaborative approach to accelerating  open-source innovation around LLMs.

"With InstructLab, developers can build models specific to their business domains or industries with their own data, so that they can see the direct value of AI rather than just the model providers seeing the value." (“IBM Unveils Next Chapter of watsonx with Open Source, Product ...”)

IBM Research Director, Dario Gil, urged the audience to tap into the benefits of dark data in the Enterprise. He said that nearly 100% of publicly available data from the internet has been used to train LLMs, whereas less than 1% of enterprise data is used in LLM training, which creates a huge opportunity for value creation from AI. Using IBM InstructLab, organisations can rapidly update LLMs with new or untapped organisational ‘dark data’.

RHEL AI combines  an enterprise-ready version of InstructLab, IBM’s Granite LLMs, and Red Hat Enterprise Linux to simplify AI deployment across hybrid infrastructure environments.  

InstructLab will enable foundation models to learn incrementally, similarly to humans, and provides toolset to feed new data into your model without needing to retrain it from scratch. The inclusion of IBM’s Granite-7B language model into InstructLab allows anyone to add new skills and knowledge.  

AI Ecosystem Partnerships

IBM’s ethos to partnerships in AI centres on trust, transparency, and collaboration. In its partnerships, such as with Meta and within the AI ecosystem, IBM emphasizes ethical AI development, responsible data usage, and open innovation. The AI Alliance fosters a collaborative environment to advance AI technologies while ensuring they are fair, accountable, and secure. IBM prioritizes building solutions that augment human capabilities and align with ethical guidelines, reinforcing a commitment to creating AI that benefits society.

IBM announced many new strategic partnerships with other vendors including Mistral AI, Adobe, Microsoft and Salesforce. IBM also announced it is expanding ecosystem access to watsonx for AWS, Meta, SAP, Nvidia and others.

Lastly, IBM also announced the integrations of third-party models into watsonx. Notable partnerships include collaborations with AWS to bring advanced AI governance to Amazon SageMaker, and with Adobe to integrate watsonx and Red Hat OpenShift into the Adobe Experience Platform. Additionally, IBM is working with Meta to include the latest Llama 3 model on watsonx and has established strategic partnerships with companies like Microsoft, Salesforce, and Palo Alto Networks to further enhance AI capabilities across various platforms.

Quantum Computing

IBM's advancements in quantum computing were also a highlight. CEO Arvind Krishna noted the significant progress in this field, with IBM's quantum systems having run over 3 trillion experiments. He said,

“When you run 3 trillion of something, it’s no longer a science experiment. "It is something that is becoming a real computer system." (“IBM CEO’s 5 Boldest Remarks On New Products, AI And Quantum ... - CRN”)

CEO Arvind Krishna also left the audience with the teaser that although THINK announcements this year focused on IBM’s advancements in AI, next year we can look forward to hearing more about Quantum. Watch this space…

Strategic Acquisitions

IBM announced its intent to acquire HashiCorp to bolster its capabilities in automating multi-cloud and hybrid environments, it’s expected that this will enhance IBM's infrastructure lifecycle management and security offerings and empower customers with an even more robust approach to their sustainability and FinOps agenda.  

 

Topic

Related Insights