Generative AI training programs are popping up all over the place today. One of my acquaintances said that he was “joking” about enrolling in such a program. Many are expensive, time consuming and of low value. Participants proudly display their certificates and even praise their instructors. They don’t know that it won’t lead to a job offer after all. Many programs simply take advantage of the ignorance of the general public. In some cases, operators and instructors truly believe themselves to be experts in generative AI.
But if you can get in without any real-world coding or machine learning experience, how can you even learn the basics of generative AI? This article is about serious technical training. Excludes general courses equivalent to cooking classes for amateurs. Therefore, I focus on training that is advertised as technical and sometimes described as advanced.
academic model
Academia usually offers some of the most expensive programs. There are two reasons for this. First, higher education takes advantage of the large loans students get to pay for tuition. Second, the fact that many universities are still respected by employers. Out-of-state students usually have to pay more. But much of the money raised, at least in the United States, goes to management and sports-related investments. Teachers get very little money. In fact, most of them are part-time, non-regular employees, very poorly paid, who either don’t have a PhD or are inexperienced PhD students.
Also, due to the tenure system, many subjects are outdated. Some feature older techniques such as linear regression and rename them statistical machine learning. Or focus on poor concepts and outdated jargon such as: p-worth. But the content hasn’t changed in decades. Reading the textbook makes me feel like I’ve gone back in time. But in many curricula it is a large part of the training. The program may include one or two sessions with him to update the presentation. For example, students may run deep neural networks (much simpler than GANs) or play with bags of words. The latter, advertising materials even describe him as an NLP or LLM. In practice, even projects involving web crawling are limited in scope. You’re unlikely to learn clever techniques for crawling and organizing unstructured tex.
Your best bet is to look for programs where external practitioners from top companies actually teach most of their classes. Universities that have produced many successful entrepreneurs or have attracted private funding usually do so.
Weird selection process
As if this wasn’t enough, you may have to deal with a ton of paperwork. It starts with filling out a lengthy questionnaire and paying a non-refundable fee. You may be required to take a test unrelated to your certification. They will typically provide a GPA score, answer questions about race, gender, disability, as well as sexual orientation and immunizations, and sometimes send a letter of recommendation. I’m not sure how this relates to the qualifications to learn generative AI. I don’t have a GPA, and if I had a score, it would have been 30 years ago. Which state you live in is an important issue that may require official documentation. But surprisingly, you don’t have to present your GitHub portfolio as if it were irrelevant.
Datacamp and MOOC model
MOOCs (massive open online courses such as Coursera) and intensive data camps range from cheap to very expensive. Quality does not necessarily correlate with value obtained. You can also exclude instructors who don’t even reveal the instructor’s name. Avoid. Yet the common drawbacks are similar. Participants are often treated as if they know very little. Probably because they are trying to attract as many people as possible in order to increase their income. Entry standards may be lower than stated in marketing materials.
Here is the final result: Spend a lot of time learning the basics you already know. The progress of the video is painfully slow. Often the instructor spends several minutes typing Python code into a Jupyter notebook, one character at a time for her. This will put you to sleep. The exam will then consist of multiple-choice questions. It is not an assessment of your ability to work in a professional environment. Sometimes it involves judgmental questions and the “correct” answer is questionable. However, you can preview the material, which is usually free, and decide for yourself if it’s worth your time and money. Beware of programs that guarantee employment or money back. You may end up not getting the job you wanted. Unfortunately, beginners may not be able to judge the value of the program.
One way to assess value is to reach out to someone who has the certification in question on their LinkedIn profile page and ask them what they think about it.
how people learn
People learn very differently. There is no “one size fits all”. If, after due diligence, you find a program that meets your needs and budget, go for it. Be mindful of time constraints. Is spending weeks on a program (even if it’s free) really worth it compared to spending time on something else?
For me, I learn by practicing on my own, finding answers to my questions on Stack Exchange and other discussion forums, and asking my own questions. Find datasets on Kaggle or similar. I learned TensorFlow and GANs (Generative Adversarial Networks) by testing and improving Python code I found on Machine Learning Mastery and Medium. Now I can find a much better version on my own platform. In some cases, you pay someone to develop and deploy your application. Recently, I designed a web API with a Python backend for one of my GenAI apps. The next project on my list is about designing the SDK. Little by little, I’m starting to look like an engineer. The interaction with the people I hired was very informative, less expensive than most courses and much quicker. And, of course, my “work” focus. But it’s just me.
alternative model
I couldn’t find a workout that suited me, so I finally decided to train myself. I realized I’m not the only one. The same is true for many professionals who have limited time and already know more than the basics. Like me, they can learn a lot on their own and essentially only need limited instruction to get started right away. I made a program for them. Participants typically have a minimal technical background and know Python. For example, engineers and product people from various industries.
If you’re interested, check out my AI/ML lab certifications here. The one on generative AI is based on my book currently accepted by Elsevier. Fees are lower than the entrance fees of any academic program. The value is excellent, especially due to the depth and novelty of the material on offer, as well as the customization. You’ll also spend orders of magnitude less time as you focus on the aspects that are most valuable and go straight to your goal without creating boring videos.
About the author
Vincent Granville is a pioneering data scientist and machine learning expert, founder of MLTechniques.com and co-founder of Data Science Central (acquired by TechTarget in 2020), VC-funded Former executive, author and patent owner. Vincent’s past corporate experience includes Visa, Wells Fargo, eBay, NBC, Microsoft, CNET and InfoSpace. Vincent is also a former Postdoctoral Fellow at the University of Cambridge and the National Institute of Statistical Sciences (NISS).
Vincent published number theory journal, Journal of the Royal Statistical Society (Series B), and IEEE Transactions on Pattern Analysis and Machine Intelligence. He is also the author of Intuitive Machine Learning and Explainable AI. You can see from here. He lives in Washington State and enjoys studying stochastic processes, dynamical systems, experimental mathematics, and probabilistic number theory.
