Gaining Practical Artificial Intelligence Skills

[The image above is generated by Midjourney. The prompt I used to create the image is listed at the end of this email.]

Understanding AI and applying that understanding to your career is one of the most important skills you can add over the upcoming years. It’s not about replacing what you know but augmenting your skills with productivity gains from artificial intelligence tools.

That’s why I have devoted so much time over the last year to understanding how AI works and applying it to things I already knew how to do, like marketing and IT operations.

I assume you want to learn and remain relevant, so I compiled a list of ways to do that. It runs the gamut from things a knowledge worker can accomplish to more technical advanced activities for those with more software development or systems management capabilities.

Online Learning Platforms

Embrace the wealth of online platforms offering AI courses and projects. Platforms like Coursera, edX, and Udacity provide comprehensive curricula designed by experts, but I suspect some of these courses. I recently took a class on one of these sites, and it was woefully out of date and had been renamed to capture a product name change but didn’t incorporate the new features (to be fair, that’s in AI time, and the product had changed twice in past eight weeks).

Here’s a list of classes I do think are worth looking at. Even though two of the paths are by Google and Microsoft, they have a good amount of content that will be useful beyond a single vendor. The DeepLearning class is also more theoretical than operational but still very useful.

• Google Cloud Skills Boost - Google’s ten-course learning path will teach you the fundamentals of Generative AI from Google Cloud experts.

• DeepLearning.AI: Generative AI with Large Language Models - AI for Everyone”, a non-technical course, will help you understand AI technologies and spot opportunities to apply AI to your organization's problems.

• LinkedIn Learning: Career Essentials in Generative AI by Microsoft and LinkedIn - Discover the skills needed to apply generative AI in your career. Learn the core concepts of artificial intelligence and generative AI functionality.

I’ve created two courses for CreativeLive. If you like the newsletter, I think you’ll enjoy the courses. (I already got paid, so if you take them, I don’t get an additional dime, but I think they are good.)

• 10x Your Marketing Efforts with ChatGPT, Midjourney, and More

• Getting Started with AI Prompt Engineering

Artificial Intelligence Certifications

Certifications from organizations like Google, Microsoft, and IBM may bolster your credibility. But honestly, I am not a big fan of any of these certifications as they aren’t vendor-neutral. I think there’s value in learning, but from what I have seen, the certification is not very valuable. The free badges from Google are nice on a LinkedIn profile, but paying for an AI certification at this stage is not a good investment, in my opinion.

I think there would be a good offering if there were certifications geared toward using artificial intelligence for practical business operations. Still, as it stands today, I haven’t seen an offering I’d endorse.

Hackathons and Competitions

Participating in AI hackathons and competitions exposes you to real-world challenges. These events encourage innovation and creativity, pushing you to apply AI techniques in novel ways. Platforms like Kaggle host various competitions that cater to all skill levels. Others also do that, like the AICrowd, that merits a look.

• Kaggle - Kaggle competitions are contests where individuals or teams work to solve a specific data science problem posed by competition hosts. These hosts can be companies, research institutions, or individuals with a challenging problem they'd like the community to tackle.

• AICrowd - AIcrowd hosts challenges that tackle diverse problems in Artificial Intelligence with real-world impact. AIcrowd Community spearheads state-of-the-art, whether advanced RL innovation or ML applications in scientific research.

Contribute to Open Source

Open-source AI projects are valuable resources for learning and contributing. Explore AI-related GitHub repositories and contribute code, documentation, or bug fixes. This hands-on involvement enhances your coding skills and showcases your expertise.

•  Identify Your Interest and Skill Level: Decide what interests you the most: neural networks, natural language processing, computer vision, reinforcement learning, etc. Are you a beginner, intermediate, or expert? Your contribution can vary based on your expertise.

• Find the Right Project: It's the most popular platform for open-source projects. Search for AI projects or libraries that align with your interests. Libraries like TensorFlow, PyTorch, and Scikit-learn always need contributors.

• Start Small: Before diving in, understand the project's goals, architecture, and contribution process. Many repositories label specific issues as "good first issues" or "beginner-friendly" to help newcomers get started.

• Documentation: Improving documentation is often overlooked but is crucial. It's a great way to start if you're not ready to dive into code.

Hands-on Projects: Learning by Doing

Theory is valuable, but practice makes perfect. Engage in AI projects that align with your interests. Create a sentiment analysis model, build a recommendation system, or develop a chatbot. Practical application deepens your understanding and hone your skills. Here are two ways you can play around with AI for free.

• Google Colaboratory - Colab is a hosted Jupyter Notebook service that requires no setup and provides free access to computing resources, including GPUs and TPUs(Tensor Processing Units). Colab is exceptionally suited to machine learning, data science, and education.

• HuggingFace Spaces - Hugging Face Spaces allows users to deploy machine learning models as interactive web applications without requiring complex infrastructure setup. Once deployed, these applications can be shared with the community, colleagues, or the public. The platform encourages collaboration by allowing users to explore and interact with applications created by others.

Attend AI Conferences and Meetups

AI conferences and meetups facilitate knowledge exchange. Engage with experts, attend talks, and participate in workshops. Meetup.com is an excellent resource for learning from people in your community.

I’ve been running a local meetup.com for about four months, and it’s been beneficial for me to meet and network; we do in-person and online meetups. We have had people from all over the world attend our online meetups. They are free and offer a way for you to network with others. If you’d like to join my meetup group, you are welcome to - RDU Artificial Intelligence Meetup Group. Or find one local to your area on Meetup.com.

Apply AI at Your Workplace

Integrate AI into your job responsibilities, if applicable. Implement AI solutions that streamline processes or enhance decision-making. I have done this personally, from using AI to rewrite emails and escalate situations to creating content.

• Assess the Potential for AI Integration: Before diving in, take a moment to evaluate where AI can be most beneficial in your job role. Are there repetitive tasks that can be automated? Can data-driven decisions be enhanced with predictive analytics? Identifying these areas will give you a clear roadmap for AI implementation.

• Streamline Processes with AI: One of the primary advantages of AI is its ability to automate routine tasks. From sorting emails based on priority to scheduling meetings, AI can handle various administrative duties, allowing you to focus on more strategic initiatives.

  • Calendaring - There are a ton of apps coming on the market right now that offer AI assistance, despite group calendaring and other tools like Calendly. Currently, I am looking at MayDay (link below), which is in public beta.

• Enhance Decision-Making with Data: AI quickly processes vast amounts of data. By leveraging AI-powered analytics tools, you can gain insights that were previously hidden or too time-consuming to extract. This can be particularly beneficial in roles that require data-driven decision-making, such as marketing, finance, or operations.

  • ChatGPT Advanced Business Analytics - In a recent development, OpenAI introduced a business version of ChatGPT named ChatGPT Enterprise. Alongside this announcement, a significant name change was made to a critical feature previously known as "Code Interpreter." This feature, which was launched two months prior, has been renamed "Advanced Data Analytics." Initially, the Code Interpreter allowed ChatGPT to perform tasks such as mathematics, file uploads and downloads, data analysis, and code creation and interpretation. Despite its capabilities, the name "Code Interpreter" was seen as limiting, suggesting its primary use was for coding. The new name, "Advanced Data Analytics," aims to highlight the feature's broader applications, especially for data insights, although some argue that this name might also be somewhat restrictive. The renaming appears to be a strategic move to appeal to business users, emphasizing the feature's data analytics capabilities over its code interpretation functions. Regardless of the name, the quality remains a powerful tool, offering various analytical capabilities. You can use it as part of a ChatGPT Plus subscription.

• Legal Considerations: While AI offers numerous benefits, it's essential to approach its implementation with a legal lens. Are you obligated to keep certain information confidential? You probably want to check with your legal counsel on their thoughts before copying or sharing information in ChatGPT or any other chatbot, for that matter. This is one of the reasons why I often write about training and fine-tuning LLMs. Because of privacy and regulatory concerns, we’ll have private chatbots for many businesses.

Continuous Learning and Adaptation

AI is dynamic, requiring continuous learning and adaptation. Embrace a growth mindset, and be prepared to evolve alongside technological advancements. I also think this shouldn’t be considered extra work. In the long term, your AI skills will help you accomplish your job tasks more quickly and effectively.

Tip of the Week: Chatbot Workflows

If you are a regular user of ChatGPT, you probably notice a bit of a decline in its dependability and the results it’s generating. This article on LinkedIn has a nice commentary on the what and the why. However, if you are depending on AI to help you get your work done, you might want to add a couple of backups to ChatGPT.

I generally get better results when reducing the scope of my tasks to ChatGPT. Rather than generate a long-form piece of content, I create an outline and then have ChatGPT work on the outline in conjunction with me. So I will stick to one point at a time and then cut and paste the output in Google Docs, where I edit it, then have GrammarlyGO help me improve the writing. It’s not ideal, but much faster than doing it from scratch.

That process involves cutting and pasting from Google Docs to ChatGPT. I have had good results by incorporating Wordtune, a browser extension and service combo that integrates an AI writer into my web pages, including those on Google Docs.

I also have been using Claude quite a bit lately. I like it as an alternative to ChatGPT, and if you read the article above, you can probably guess why. It’s not suffering from some of the overuse that ChatGPT is. I often create content in ChatGPT and then paste it to Claude to improve the content.

Finally, my biggest tip is to lengthen your prompts but shorten your scope. What I mean by that is to guide ChatGPT’s results, you should provide a particular prompt:

• Specify the output - I often tell ChatGPT to provide the output in prose or a bulleted list because it often defaults to a bulleted list and I need prose for blog posts.

• Provide constraints - I often tell ChatGPT to stick to a specific word count or to avoid certain topics. For instance, if I'm writing about a particular product, I might instruct it to avoid mentioning competitors or to stay within 300-500 words. This ensures the content remains focused and doesn't stray into unrelated areas.

• Set the tone - It's essential to let ChatGPT know the style you're aiming for. Whether it's formal, casual, humorous, or professional, specifying this can help get the desired output. For example, "Write a humorous take on the latest iPhone features" will yield a different result than just asking for information on the iPhone features.

• Use examples - Sometimes, providing an example of what you're looking for is beneficial. If you want a product description, you might include a sample product description to give ChatGPT a clearer idea of your expectations.

• Iterate and refine - Don't expect perfection on the first go. Reviewing the content, making edits, and then asking ChatGPT or Claude to refine or expand on specific points is a good practice. This iterative process can lead to more polished and tailored content.

In conclusion, AI tools like ChatGPT are powerful but not infallible. By narrowing down your scope, being specific with your prompts, and using a combination of chatbots, you can harness the power of AI to produce high-quality content efficiently. Remember, it's all about finding the right workflow that suits your needs and refining it over time.

What I Read this Week

• What OpenAI Really Wants - Wired

• Why Meta’s Yann LeCun isn’t buying the AI doomer narrative - Fast Company

• Time 100/AI - Time Magazine

• Nvidia Muscles Into Cloud Services, Rankling AWS - The Information

• Sundar Pichai on Google’s AI, Microsoft’s AI, OpenAI, and … Did We Mention AI? - Wired

What I Listened to this Week

AI Tools I am Evaluating

• Pika - A powerful Text-to-Video platform

• Mayday - AI assistant + calendar that helps you organize, schedule, and protect your day. It’s in public beta right now.

Midjourney Prompt for Header Image

For every issue of the Artificially Intelligent Enterprise, I include the MIdjourney prompt I used to create this edition.