What Happened

In an era of influencer economy, famous Youtubers like Siraj Raval is considered very successful in spreading words of and enlighting newcomers to the somewhat daunting AI world. As of now, his YouTube channel has 692K subscribers which is something. His content has a very broad reach. His ‘Doing X in 5 minutes’ tutorial videos are so popular, there is a meme for that:

But recently, he ran into some problems. It’s with his new course ‘Making Money with Machine Learning’ that he charges $200 for it. Charging $200 for a course is totally fine, the controversial part of it is that people were complaining that he used some code from other people’s GitHub repo and also he accepted way more students than he promised and can properly handle without downgrading the experience. When some people seeking for refund, they also ran into some hiccups.

‘Growth Hacks’ vs ‘Growing Pain’

Some people accused him of fraud, I’m not so sure about that. This to me looks more like an honest mistake due to lack of experience in scaling up and entering into a field he is not familiar with.

See, most of his more popular videos are entry-level tutorials with a bit of ‘entertainment’ (rap, parody, etc.) elements to it. He is very natural at that and that makes AI or Machine Learning not that boring or scary. And I believe that’s why people like his content the most. However, when it comes to more serious education, especially when you want to charge $200 for it, people expect certain standards and he might not be fully aware of what those standards are. For example, you need to own your content, if you want to use other’s contents like code snippet or diagrams, give credit, and very proper credit at that.

I remember when fast.ai’s Jeremy Howard was teaching the ‘Introduction to Machine Learning for Coders’ course. He used an ‘overfitting vs underfitting’ diagrams, and he credited it as from Quora.

But turns out, the above image originally came from Andrew Ng’s Stanford Machine Learning course. So Jeremy Howard actually specifically corrected this error on his next lecture and went out of his way and praised Andrew Ng teaching style. I think this is the respective way of handling this. I don’t think Siraj Raval is intentionally not giving out credit, he might not be so used to the academia convention of giving references, etc. People won’t complain about it if you are doing free videos on YouTube and used their code, but if you try to charge for the content on a formal educational program, that’s a totally different animal.

On the accuse of taking too many students than he promised, I think it is mostly a scaling challenge. When the success hit you too fast and you are not prepared, you most definitely will stumble. It is understandable. However, how you handle it afterward is what will make all the difference. You can either get more help or properly refund people for the level of service you fail to live up to, just don’t run away from it or trying to solve a mistake with more mistakes. This will at least save the reputation and get a second chance to make things right.

Final Thoughts

The situation is still developing and we’ll see how he handles this. I’d like to say we give him the benefit of doubt. This guy did a lot for the AI community, helped people get started with AI/ML and made a lot of people laugh after all. My suggestion?

Directly face the issue, acknowledge that he made a mistake and compensate people properly, then make a video on YouTube on how he handles scaling and did damage control! Turning crisis into opportunity. Take a page from The Tech Lead’s playbook. Gotta monetize whenever you can, right?