The future of AI as seen by Red Hat
Flexibility, Open Source and more!
Hey guys! Today we're going to talk about a topic that's on everyone's lips, and the one who came to bring incredible insights into the future of AI was none other than Red Hat, in a keynote at the Lisbon Data & AI Forum 2024.
But before we delve into this fascinating world, let’s take a step back and remember some past predictions that went wrong. Have you ever heard that the founder of IBM thought there would only be room for five computers in the world? Yes, today we have billions of these little devices in our pockets! And in 1936, they said no rocket would leave the Earth. I guess they didn't count on Elon Musk's stubbornness.
The Hybrid and Open Source Future of AI
But going back to the future (no pun intended, please), Red Hat has made it very clear: AI will be hybrid and open source! Hybrid because companies will have the flexibility to run their models both on-premises and in the cloud, depending on data regulation, privacy and sovereignty needs. And open source because, let's face it, no one wants to be tied to expensive proprietary solutions, right?
The Red Hat speaker highlighted that open source is the key to freedom of choice and innovation. Imagine relying on a single provider for all your AI needs? No thanks! With open source, you have the power to adapt, improve, and share solutions.
AI in business: Is it worth investing?
But then, is it worth investing in AI in business? According to the research mentioned in the keynote, 90% of CEOs think so, even with some risks involved. And get this, 55% of CIOs are already investing in AI, with almost 60% dedicating 5% of their budget to this area. It is no wonder, after all, AI can bring competitive differentiation, improve customer experience and generate efficiencies.
Red Hat has presented some very interesting use cases for AI. A hospital using machine learning to identify suicide risk in early stages, specialized chatbots for customer service and even streamlining bureaucratic processes in the public sector. Who would have thought that AI could reduce a public entity’s response time from two hours to just one second?
Challenges and considerations on the AI journey
But wait, it's not just about implementing AI left, right and center. Red Hat highlighted some key points for a successful journey:
Leadership commitment: Without the support of senior management, forget it, it won't happen!
Focus on business value: AI is nice, but it has to positively impact the bottom line.
Scalability: There’s no point in having a cool model if you can’t scale it across the entire company.
Flexibility: With AI constantly evolving, you need to be open to change and adaptation.
Change management: We cannot forget the human factor! It is crucial to empower and involve people in this transformation.
Additionally, choosing the right AI model is critical. There’s no point in wanting a giant model that answers everything if your use case is specific. Red Hat suggests the "recovery augmented generation" (RAG) approach, which allows for smaller, cheaper models with reasonable quality.
Another point raised was the importance of having the flexibility to implement AI on-premises or in the cloud, depending on regulations and data location. And it's worth remembering that AI is not an island! It needs to integrate with existing applications in the company.
The future of AI as seen by Red Hat
Now that we understand the challenges and considerations, let’s look at how Red Hat is approaching the future of AI. They showcased two main products: one focused on the Red Hat Linux operating system and the other on the OpenShift container platform.
At Red Hat Linux, they offer the Granite family, a partnership with IBM Research that provides robust and reliable open source models. In OpenShift, the idea is to treat AI as just another application, using the same platform and DevOps processes already established in the company. This makes it possible to have a scalable and integrated AI operating environment.
But what really caught our attention was Red Hat's open source approach. As mentioned earlier, open source brings flexibility, transparency, and innovation. Can you imagine being able to see exactly how models were trained, what data was used, and having the freedom to adapt them to your specific needs? Now that's empowerment!
Additionally, Red Hat provides ongoing support and updates for its AI products, ensuring you’re never left behind on this journey. And best of all? Intellectual property is open source, so you don't have to worry about those pesky licensing issues.
But what about the future?
Well, as we saw at the beginning, predicting the future is a thankless task. But one thing is certain: AI is here to stay and will continue to evolve rapidly. Red Hat's approach, with its focus on flexibility, open source, and integration with existing applications, appears to be a solid path for companies to navigate this sea of possibilities.
Of course, there are still many challenges ahead, such as AI regulation, ethics in the use of data and training professionals. But with the collaboration of the open source community and the commitment of companies to use AI responsibly and strategically, we can build an exciting and innovative future.
And what will we do with it?
Red Hat's keynote at the Lisbon Data & AI Forum 2024 showed us that the future of AI is hybrid, open source, and closer than we think. With an approach focused on flexibility, integration and transparency, companies can embark on this journey strategically and reap the rewards of AI in business.
I particularly liked Red Hat's RAG (Retrieval-Augmented Generation) and containerized approach. We here at Yes Marketing have a large project under development that requires differentiated training for image recognition and data inference in those same images. We are strongly tempted to give OpenShift's approach to solving this problem a try. The flexibility and scalability offered by the platform, together with the robustness and reliability of the open source models of the Granite family, seem to be a perfect combination to meet our specific needs.
Furthermore, we cannot forget the fundamental role of the open source community in this process. Together, we can shape a future where AI is a powerful tool for the common good, driving innovation and creating value for all. Collaboration and knowledge sharing are essential to overcoming challenges and making the most of the potential of AI.
So if you’re considering embarking on the AI journey, it’s worth exploring Red Hat’s approach and seeing how it might fit your organization’s needs. With flexibility, transparency and an engaged open source community.
