Generative AI, Quantum And Partnerships: The Amazing Highlights From IBM’s Think 2024

Thank you for reading my latest article Generative AI, Quantum And Partnerships: The Amazing Highlights From IBM’s Think 2024 . Here at LinkedIn and at Forbes I regularly write about management and technology trends.

To read my future articles simply join my network by clicking 'Follow'. Also feel free to connect with me via Twitter, Facebook, Instagram, Podcast or YouTube.

At IBM’s Think 2024 event, IBM CEO and Chairman Arvind Krishna opened his keynote with the message that the convergence of AI, hybrid cloud and quantum computing is set to revolutionize business.

Headlining the event were big announcements in all of these fields, as well as news around IBM’s many partnerships with industry leaders such as Amazon, Meta, Microsoft and Adobe.

But the star of the show, once again, was undoubtedly generative AI. Here, IBM laid out details of its open approach, announced a number of new assistants built around the watsonx platform, and unveiled new tools for training foundation models and generating real-time insights into business performance.

Having spent a few days in Boston learning about IBM’s vision for the future of enterprise AI, here’s a rundown of the highlights of an eventful and insightful Think 2024.

?

Open-Source – The Future Of Generative AI?

One of the biggest announcements was news that 18 of IBM’s Granite large language models (LLMs), including a three billion parameter model, a much larger and more powerful 34-billion parameter model, and a geospatial model developed in collaboration with NASA, will be released under the Apache 2.0 open-source license.

Speaking at the event, Krishna explained that this would contribute to making generative AI more open, cost-effective and accessible to businesses.

The news that this will involve models fine-tuned for coding was particularly well received, as these models benchmark as among the most impressive in their field. All will now be available on Hugging Face and Github.

Krishna told his keynote audience, "We want to use the power of open-source to do with AI what was successfully done with Linux and OpenShift,” a reference to the democratizing impact that open-sourcing had on the development of both operating systems and cloud computing infrastructure. His comments make it clear that he believes IBM will take generative AI on the same journey.

Also newly announced was Instructlab, which enables AI developers to take a new approach to designing and deploying models by incorporating domain-specific or enterprise-specific data to simplify the creation of highly specialized AI applications.

Another newly announced release – Red Hat Enterprise Linux AI (RHEL) combines InstructLab, Granite and the Red Hat Linux distribution to create a fully open-source platform for deploying AI infrastructure onto hybrid cloud environments.

?

IBM Concert

IBM Concert is a watsonx-powered tool for generatively providing real-time insights that acts as a “nervous system” across all of an organization’s technology operations.

During his own keynote, IBM SVP Rob Thomas said “The ‘aha’ for us on this one was when I sit in on our support calls, there was always one problem that came up. When something goes wrong people say ‘what changed? What happened?’ and nobody ever knows the answer to that question.

“So our view was we can use generative AI to have an instant answer to that question and, in most cases, solve the problem before we even get to the point of being on the phone.

“With IBM Concert, you can identify, you can predict, and you can fix problems before they even happen.”

Launching in June, Concert, powered by watsonx, provides holistic insights into the workings of an organization’s entire tech stack, untangling issues around connectivity and dependencies. It aims to enable users to spend less time solving technology-related problems and more time generating value for their businesses.

?

AI Assistants and watsonx Orchestrate Assistant Builder

SVP of Product Management and Growth, Kareem Yusuf, used his keynote to introduce new features for watsonx Orchestrate, specifically the Assistant Builder, which makes it easy for anyone to create AI assistants.

Using simple forms, users can create assistants in minutes, tailoring them to specific business tasks such as sales or customer services. Assistant Builder is designed around a skills system that lets it easily integrate with third-party tools like Microsoft Outlook or Salesforce.

Yusuf also unveiled a number of new assistants which IBM has created using the tool, including? watsonx Code Assistant for Z, watsonx Code Assistant For Enterprise Java Applications, and Watson Code Assistant Enterprise Edition. Yusuf also showed assistant functionality integrated into IBM applications, including Planning Analytics, Apptio, Maximum and Guardium, which is due to happen in the near future.

Yusuf told the audience during his keynote “This is what I mean by building high valued assistants that meet tailored, specific work … this is assistance, in context, for what needs to be done.”

?

IBM BI Assistant

Also hot off the press was the announcement of IBM watsonx BI Assistant. This aims to solve the complexity that has traditionally existed around BI tools, making them difficult to use for those without a deep understanding of data and analytics.

BI Assistant lets anyone analyze data, whatever their level of expertise, using simple natural language queries.

IBM hopes that removing this complexity will help bridge the gap between the huge investments being made in enterprise BI systems – predicted to surge to $27 billion by 2027 – and the fact that only 30 percent of employees are reported to be regularly using them to make decisions.

?

The Role Of Synthetic Data In Enterprise AI

I found the keynote given by Dario Gil, IBM SVP and Director of Research, particularly enlightening. Gil spoke about the fac that while a huge amount of public data has been used to train foundation models, only a comparatively tiny amount of enterprise data has been put to the same use. This creates challenges when it comes to using generative AI platforms to answer enterprise-specific questions.

Introducing InstructLab, Gil described a methodology for enhancing the ability of a foundation model to learn incrementally, by supplementing training with large datasets of synthetically-generated questions and answers. This enables it to become better at specific domain knowledge without losing generality.

One use case involved the development of the Code Assistant For Z tool, which was built to assist with application modernization. Specifically, it’s aimed at modernizing applications created in COBOL, the legacy programming language that still underpins many enterprise applications, particularly in the finance and government sectors. It’s well-documented that it is becoming increasingly difficult to find coders who can program in COBOL, creating challenges for businesses that need to update these systems.

By applying the InstructLab methodology, Gil said, “In one week, they were able to build a model that was significantly better than the model we had worked so hard on over months to fine-tune and get to that level. That's really impressive."

?

Quantum Computing

IBM is a pioneer in the emerging field of quantum computing. Krishna used his keynote to announce that the company has built more than 70 quantum computers, which are now being put to use by over 250 of their customers.

He said “People have run over three trillion individual experiments on our quantum systems over the last years … when you run three trillion of something it’s no longer a science experiment, it’s something that becoming a real computer system.”

He explained that the systems – which harness the strange properties of quantum systems such as superposition and entanglement to carry out calculations millions of times quicker than conventional computers - are being used to solve challenges around developing new fertilizers to tackle food supply issues and financial risk.

?

Partnerships

IBM also unveiled a series of announcements around new developments within its partnership ecosystem.

These include bringing AI governance capabilities to Amazon Sagemaker through its integration with watsonx.governance. It will also integrate watsonx into Adobe’s Experience Platform and Adobe Acrobat AI Assistant, and launch a new consulting service aimed at encouraging its clients to adopt Adobe Express.

Salesforce plans to integrate IBM’s Granite models into Einstein, increasing the number of models available for natural language and decision-making on the popular CRM platform.

Meanwhile, Microsoft will let customers run Watsonx tools and applications on Azure, allowing them to leverage Watsonx AI capabilities on the flexible and scalable cloud infrastructure.

And Meta will make its own LLM, Llama 3 available through watsonx, building on the commitment both companies have made towards advancing the use of open-source AI.

Overall, Think 2024 in Boston was a great event, packed with announcements and innovations that will enable IBM’s customers and partners to drive internal efficiencies and delight customers.

For more on the key announcements and on-demand keynotes, click here: https://ibm.biz/BdmnXL

#IBMPartner

About Bernard Marr

Bernard Marr is a world-renowned futurist, influencer and thought leader in the fields of business and technology, with a passion for using technology for the good of humanity. He is a best-selling author of over 20 books, writes a regular column for Forbes and advises and coaches many of the world’s best-known organisations.

He has a combined following of 4 million people across his social media channels and newsletters and was ranked by LinkedIn as one of the top 5 business influencers in the world. Bernard’s latest book is ‘Generative AI in Practice’.