Transcript:
All you need is love and data and computation. Hello, world it’s! Suraj and I’m gonna list seven ways to make money with machine learning. If you’re new to my channel, welcome hit, subscribe to get notified when I release New videos. Billions of dollars are being invested into the AI ecosystem. So an important question to ask is who is currently making money in Ai well at the heart of the AI. Ecosystem is the data itself without data. There is no Ai. It needs data to learn from and currently, the vast majority of the world’s data is owned by just a few corporations on top of that comes the base layer of AI chips. These chips are what enable AI systems to actually perform the processing necessary so far. NVIDIA has been the biggest winner in this space, just a quick glance At their stock price over the past few years shows how influential their GPU chips have been in the industry, even though they were designed for rendering graphics, It turns out that they’re perfect for the matrix operations that a lot of machine learning applications require, but it’s not just NVIDIA, IBM, Intel, Google and now a few Chinese companies like cambric on technology are designing their own. Ai chips, as well built on top of these chips, are the platform and infrastructure of Ai that is where does the AI processing and storage take place? There’s fierce competition amongst companies like Amazon and Microsoft to provide a complete cloud solution for developers to run Ai. Applications and the cost of maintaining these massive global data centers is something very few can afford on top of that are the actual models and algorithms of Ai Google’s. Tensorflow framework is still the most popular in the field and it’s free the reason being once a company starts using it. They’ll likely need a lot of computing power and who better suited to offer that than Google since Google Cloud is optimized for tensorflow. Once a company becomes reliant on their software, they will likely beat a returning customer for years very smart, but that’s not stopping the other Giants from releasing their own frameworks as well. Then we have The Enterprise Solutions built on all of this technology again dominated by Giants like Salesforce Oracle, which is somehow still alive and S AP consumer facing solutions are being offered by corporations that have massive budgets and teams to execute on innovative ideas and at the very top are entire countries like China, the UK, Germany, France and Japan, which are leading in terms of investment into their respective AI ecosystems to foster economic growth. So where do you fit into all of this? At the very top above countries? No, at a glance, it might not seem like there’s room to make money in this space with so many large entities with massive budgets, But there most definitely is lets. Start with the first way to make money with machine learning. Start a start-up. There is a massive space for startups who can provide the best vertical solutions across industries. Big corporations don’t have time to tackle small niche problems. But you do. We’ve seen that startups can succeed if they have access to four components, large data sets domain knowledge that gives them deep insights into the opportunities within a sector, a deep pool of talent around applied. Ai and capital to fuel growth. One leading startup example is a firm they offer loans to consumers at the point of sale, a very competitive space and have still managed to build a successful company that’s raised over 700 million. US dollars! They quickly created a defensible moat to ensure that they remained competitive. This is called the flywheel effect and it applies to all. Ai companies, more baguettes, better, Ai based services and products that gets more revenue and customers that gets more data and it continues its. The circle of data startups are actually disrupting every part of the AI value chain, even at the lowest level. Habana, for example, is a small team developing a chip that supposedly out runs GPUs for inference algorithm. IYA lets developers deploy and manage ML models in the cloud in just a few minutes with a much more user-friendly interface than AWS clarify provides custom models and algorithms that perform different tasks via their easy-to-use API and peek is providing data analytics as a service directly to enterprises with so many tools freely available on the internet. Anyone with the right amount of motivation and work ethic can succeed in this space if I were to build an AI startup from scratch today. I would write down a list of problems that I personally cared about solving. That’s the only way to stay motivated when issues arise in your journey, and they 100% will. Then I divide these problems into whether they fit into the enterprise or consumer markets. The enterprise is where the most high value use cases for individual startups are. So I probably pick an enterprise service. Then I create a well-designed landing page with a sign up form that details what my imaginary product would do with a price point after getting feedback from friends on the pricing and features and tweaking it a bit, I would share it on social media platforms to see if anybody signs up if people actually sign up knowing the price, that’s extremely exciting and motivating, isn’t it? I get right to work collecting a relevant data set and training a model on it once. I have a working model. I take on clients one by one lowly building, the reputation of my new brand and used the money I earn to hire and scale. Remember where there’s a will. There’s a way nothing lasts forever. Amazon’s cloud product AWS, which earns more revenue by itself than many big corporations, do seems unstoppable, but a decentralized computing network that allows anyone to use their local machine to store and process Encrypted data could provide a lower-cost option to consumers in the near future. The second way is contract work either as an individual or as a team. I used to do contract work when I was a college student. As a way to make money on the side and can definitely recommend it to you as well. It’s a hustle clients won’t come to you. You have to go out and find them, but before you do that, you need to make sure you’ve established a solid personal brand online. That includes a resume, a portfolio website, a Linkedin profile and most importantly, a Github account with a few well documented projects that showcase your skills to prospective clients. I’ve got a great video on this called resume for machine learning and how to do freelance. Ai programming definitely check out both of those before browsing one of the many freelance job board websites available on the Web. If you want to start a consulting firm having built an online brand is helpful if you want access to a larger pool of talent to join you. Many teams these days are fully distributed, including school of Ai. You want to find people that you can trust that you enjoy working with, and that have equal or greater talent than you. Then you can use a team management board like Trello to assign tasks and split the work accordingly, decide what your target market is, then find clients in that market by browsing social media websites like Linkedin and Reddit, as well as local event in your area. For example, consulting firms like Sigmoid, I’ll focus on working with big enterprises and medium-sized companies from just four countries. The third way is to find a job or internship, And I’ve got a great video on this called how to find an AI internship. Check it out! The fourth way is to write a technical book. You have two options here, self publishing or using an existing publisher. I used to write blog posts on Cryptocurrency and one of the editors at Oreilly reached out asking if I’d be interested in writing a book for them because they liked one of my blog posts so much. I ended up. Writing decentralized applications using their brand and Oreilly took care of marketing and distribution, including printed copies across many bookstores. This is a great way to earn a recurring revenue, increase. Your brand reputation and position yourself as a thought leader in the space, but be warned. Writing a book is not easy. I was spending eight hours every day writing when I was staying in Tokyo for a few months. Write a book. If you want to convey a way of thinking rather than just demonstrating facts or techniques. What books uniquely offer the reader as Long-form content is the ability to learn how to think like the author at a certain level of detail. You shouldn’t write a book solely for the money, but instead, if you feel like none of the educational content that you’ve seen offers your perspective on which technology stack works best and envisions where we’re headed in the way that you do, publishers do take a royalty fee, though, and this varies per publisher. If you don’t want to deal with that, you can self publish your book. Through a service like create space books are a good source of passive income. They’re a recurring revenue stream, and you can also use the book as leverage to later Get speaking gigs like what Caillou has been doing after publishing his latest book. Ai superpowers highly recommend it By the way I loved it. The fifth way is to make educational content online blogs, videos. There are a lot of examples of this for blogs. Medium pays its bloggers for publishing on the platform and it’s become a very popular tool in the AI community. If you want to make a paid course, Udemy is probably your best bet to get started as fast as possible without having to build a learning management system yourself. But if you’re willing to play the long game then. Youtube is a great option. You won’t make money from ad revenue at first, but as you build an audience, it will come. You can get creative with revenue stream here as well. Once you’ve established an audience things like partnering with affiliate referral programs, patreon and brand sponsorships. The sixth way is to create an automated trading bot. I’ve got several videos on machine learning for trading and I’ll link to them In the description, Creating a high frequency trading bot focused on crypto currencies, like Bitcoin is a popular project idea in the AI community. The advantages of focusing on crypto currencies is that you’re not necessarily starting with a handicap against the big trading firms because the technology is relatively new and the apis are really good. People have definitely made profitable. Bots, but no one will share their source code because that would be giving away their secrets. The best thing to do in this case is to go on Github and search for relevant crypto trading bot repositories. Until you find some well documented examples and learn from them and the last way to make money with machine. Learning is to compete in machine learning competitions. Kaggle is a great choice. The winners of their competitions receive hefty cash prizes new. Mirai is another good example data. Scientists compete to see whose model can predict asset prices the best and earn money if they win. Hackathon’s are also always a fun way to earn prize money and there are several listings like Haiti. Nuggets that help you find them. Find a local hackathon and compete. You could even make a living attending and winning hackathons around the world anyway. I hope you found my tips useful. Wizards links to everything will be in the video description. Please subscribe for more programming videos and for now. I’m gonna make some paper so thanks for watching.