By M.Joel Dubinko
I’ve been interviewed by AI about my LinkedIn Learning course on AI. Scroll to the end to interact with a live instance of this AI interview host, called “Compelling Conversations,” for yourself. And to find out about a FREE webinar.
On with the interview…
Compelling Conversations: Welcome to today’s episode! We’re thrilled to have with us M.Joel Dubinko, but you can call him Joel. With an extensive background in software engineering, Joel has worked on a wide array of projects over the last 30 years, ranging from embedded software and search technologies to Enterprise DevOps. Joel, it’s an honor to have you here. Could you start by giving us a brief introduction about yourself and what brings you to our show today?
MJD: That was a pretty good introduction you already gave. Thanks for that! All I’d add is that I delight in solving difficult problems, which is possibly the only common thread weaving throughout my career. But what I’m really excited about this week is my new course on LinkedIn learning, all about how to build AI apps with orchestration frameworks like LlamaIndex and LangChain.
Compelling Conversations: Joel, the pleasure is all ours, and your passion for tackling tough challenges shines through! Diving right into the heart of today’s topic, your course on building AI applications is bound to pique the interest of many, especially with the unique focus on orchestration frameworks like LlamaIndex and LangChain. For our listeners who might not be as familiar with these terms, could you explain what LlamaIndex and LangChain are and why they’re important for developing AI applications?
MJD: I see lots of emerging projects that provide chatbots, or AI assistants, or ways to “chat with your own documents”. Personally, I’ve shied away from these services due to privacy concerns. As I say in the course, Large Language Models are trained on billions of documents, but none of them from your private files–and this is a good thing! Right now, the best way to play with these technologies is to roll your own, and frameworks like LangChain and LlamaIndex make this a whole lot easier.
Compelling Conversations: That’s a fascinating approach, Joel, especially in today’s digital age where privacy concerns are paramount. By focusing on locally-running Large Language Models (LLMs) and leveraging frameworks like LangChain and LlamaIndex, it seems you’re empowering individuals to harness the power of AI while maintaining control over their data.
This brings us to a curious point about your course: the idea of turning ordinary English prose into a programming language in its own right. Could you elaborate on this concept and its implications for AI programming, especially for our audience who might be more accustomed to traditional coding paradigms?
MJD: Well, programming used to be for programmers. But making English into a programming language rewrites all expectations about what even is a ‘programmer’. Take OpenAI’s release of GPTs for example. When you walk through their setup workflow, you don’t write a line of code. Just describing things in plain language is enough to build a workable app. But let’s be honest, this has limitations. The whole field is changing so rapidly, that it’s hard to keep up. And in conditions like these, I encourage folks to dive in to the fundamentals. Get a good handle on the basics, and you’ll be able to quickly adapt to new advances. Which, not so coincidentally, is the approach I took in the LinkedIn course. 🙂
Compelling Conversations: Joel, that’s an insightful perspective on the evolving role of programming and the democratization of technology through AI. It’s thrilling to consider how these developments are opening up new avenues for creativity and problem-solving across various fields.
Given the fast pace of change in AI technologies, what challenges did you face in creating a course for such a rapidly advancing subject? How did you ensure that the content would remain relevant and valuable to learners, especially considering the technical expertise of our audience?
MJD: While I was writing the script for the course, and even as I was onsite recording the videos, I kept having to rewrite code samples to track new releases of the frameworks and APIs. There’s a GitHub repo for the course where I checked in all these samples–and I anticipate revisiting that as things continue to change.
Compelling Conversations: Joel, that sounds like a meticulous and adaptive approach to a field known for its dynamism. Keeping the course materials up-to-date through a GitHub repository is a fantastic strategy, ensuring that learners have access to the latest advancements and can see real-time updates as the technology evolves.
Your work, particularly in the realm of AI-amplified problem-solving, seems to not only reflect your expertise but also your passion for innovation. Could you share more about how AI amplifies problem-solving in your day job at Dubinko Consulting, and how this theme is woven into your course?
MJD: I’m keeping a close eye on code completion models. There’s some early research that shows that code quality suffers when using these tools. On the other hand, I know some fantastic architects who are multiplying their productivity with coding assistants. It boils down, again, to fundamentals. I help people to build core skills, not only in computer science, but also in Systems Thinking and problem-solving, topics that aren’t routinely taught in schools. The stronger your base, the more of a multiplier AI tools will be for you. And these are the sort of people that are going to thrive in the AI economy.
Compelling Conversations: Joel, your emphasis on the importance of a strong foundational skill set in both computer science and Systems Thinking is incredibly valuable. It highlights a critical perspective in today’s rapidly evolving tech landscape: AI tools and technologies, like coding assistants, are not just about automating tasks but also about enhancing human capability and creativity when used wisely.
As we’re nearing the end of our conversation, could you summarize the main points or takeaways you’d like our audience, especially those interested in diving into AI and its applications, to remember from your LinkedIn Learning course and your insights today?
MJD: It’s kind of scary out there. Change always is. But there’s opportunity like never before. I also want to give a shout-out to the LM Studio project. This is a desktop app for Mac, Linux, or Windows, that makes it easy to locally host LLMs, even if you don’t have pricy GPU hardware. It serves an OpenAI-compatible AI and works great with LangChain and LlamaIndex. So check out the course, and enjoy yourself while working through it. I’d love to hear how you fare.
Compelling Conversations: Joel, your enthusiasm and forward-looking attitude are infectious. Emphasizing the blend of challenge and opportunity in the AI landscape, alongside practical tools like the LM Studio project for locally hosting LLMs, is invaluable advice for our audience. Your course sounds like a fantastic resource for anyone looking to navigate the complexities of AI with confidence and creativity.
Before we wrap up, is there anything else I should have asked? And could you share how our listeners can find you on the internet or social media to follow your work or reach out with their experiences from your course?
MJD: One question I like to ask is “how should I be thinking about this?” In the case of AI, it’s like the early days of the Internet. In a few short years, knowledge-worker jobs that don’t involve AI in some capacity will be as rare as jobs that don’t use the Internet or email today. You can keep up with me on my blog at dubinko.consulting or on mastodon.social as @dubinko. On the blog, you’ll get prompted to sign up for a weekly email, which is the best way to keep in touch.
Compelling Conversations: That’s a powerful analogy, Joel, positioning AI in the same transformative light as the early days of the Internet. It certainly frames the importance of embracing AI technology for the future of work in a compelling way.
Joel, it’s been an absolute pleasure having you on the show. Your expertise and perspectives have not only illuminated the exciting possibilities of AI but also the practical steps our listeners can take to be a part of this evolution. Thank you for your time and for sharing your knowledge with us today.
Listeners, be sure to check out Joel’s LinkedIn Learning course and connect with him to continue the conversation. Until next time, keep exploring, learning, and innovating. Goodbye for now!
Compelling Conversations is a GPT programmed by Dubinko Consulting, and you can have it interview you right now. Link
For a behind-the-scenes look at putting together this AI interview host, including secrets of writing effective prompts, you can find this exclusive information solely in The Problem Solvers Digest.