In this article, I will tell whether you should specialize in Blockchain or AI as a software developer.
AI & Blockchain are boomint.
So on the one hand you gotta AI:
- Insane salaries. For a top AI guy, working for a top AI company, you. can make up to 1M per year
- Insane progress. Every week, there is a new crazy announcement. A new crazy product that is 10x better than the product released the week before.
And on the other hand you have Blockchain:
- Crazy bull market
- Bitcoin soon at 100k
- The euphoria is returning.
- Crypto bros are back: the VCs, the entrepreneurs, the developers, the Youtubers, even the scammers! Everybody is back, ready to crush it in this new bull market.
And we also have insane salaries. For the absolute top developerss in crypto, it’s possible to make 1M per year.
AI Opportunity 1
So let’s start with AI. What kind of opportunities you have
First, you have AI / ML researcher:
- As an AI researcher, your job is to create new algorithm, so that we can build better AI models
- You are very good at maths, probably have a PHD in Math
- When you hear about the crazy salaries in AI, this what we are talking about
AI Opportunity 2
Then, you have the models developers:
- People who train new models
- They use the models developed by the AI researchers
- Before you can use chatHGPT,
- model developers selected the data to use to train the model, put in
- They designed the architecture of how to train these models:
- How many computers are needed
- which software to use to do the training
- And they actually run the models
Model developers often have a background in both:
- data science
- and also know software development, especially backend development.
But they are first and foremost data science people
It’s less difficult than being an AI researcher, you quote and quote only need a master degree in data science, or a solid relevant work experience.
Salaries are high, you can make up to 200-300k per year, even though it’s not as high as for the top AI researchers.
AI Opportunity 3
Then, you have people who do model fine-tuning:
- They cannot train new models
- Instead they modify existing models for specific use-cases
- For example, in chatGPT, you can create a customGPT, which is a more specialized version of chatGPT
- There is a graphical user interface to create a customGPT, and everybody can do it
- But it’s also possible to do fine tuning from scratch, without any user interface
- You start from the dataset of an existing models, and you will modify it with some ML framework
Let’s say you use a models like Flux, for image generation. If you want to customize the style of the images that are produced, you can fine-tune it with some examples of images that you want.
You don’t really need a background a data science to do this. Usually there are some existing frameworks that you can use. You just need to have basic skills in development ,some basic understanding of AI, and that’s it.
For salaries, fine-tuning is a sort of extra skill that can be required at some jobs, but it’s not going to massively change your salary. For example, if you are a backend developer, and you know fine-tuning, maybe you can have a salary boost of 10, 20, 30pct. But that’s it.
AI Opportunity 4
Then, you have developers who just use already-trained model
- That’s developers who use AI as end-users
- For example, if you code with AI, with Cursor AI, with, you
- Or if you use the API of chatGPT or Claude, you would be in this category
The best AI opportunity
For the really crazy salaries, you have to be an AI researcher. Becoming an AI researcher is a long-term plan. You need to have a PHD. For most people, it’s not really an option.
The next big opportunity is model developer. It’s a bit easier than AI researcher, but it’s still quite involved. It’s not just programming, you really need some knowledge in data science, and data science is not as open to self-starter compared to programmer. If you are a developer and you want to get into this, you will need to make a lot of efforts. it’s not just like studying a new framework.
So where is the demand? What kind of AI coder companies need? Most developers completely misunderstand this, so pay attention.
When you think of the AI industry, the only pure AI companies that exist are hardcore companies that build AI infrastructure, like chatGPT. They build fundamental tools that lots of people will use, and they need some really hardcore AI people. So it’s just a few companies, that hire a few technical people, that they will pay very well.
But most of the demand for AI skills is not there. Most of the demand comes from companies in other industries, with specific use cases, like Finance, web development…or Blockchain. These companies don’t build AI infrastructure. They just use existing AI tools:
- They use AI tools either to build new features
- or to increase the productivity of their developers.
But they don’t operate at the same level as the deep AI infrastructure companies.
Combine AI & Blockchain
If you are like me and you are a developer who just want to keep up with AI and stay relevant, but you don’t want to go too deep in AI and Maths, just stay at the level of companies that just use AI tech, but don’t create new AI tech.
See AI as a tool that you can use to produce better software, but for a specific use-case.
So you don’t have to chose between AI and Blockchain. AI is something that will make you a Blockchain developer:
- Either in terms of the features you can build for your Blockchain application, by using some AI API
- Or for becoming a more productive Blockchain developer
Conclusion
So AI is great, but you still need to control its output. And to control its output, you need to understand what you do. And the best way to develop your skills is to practice and build lots of small projects.
I am building an entire platform for learning Blockchain by building small projects, the 30 Day academy. On the 30 day academy, each course has 30 challenges of increasing difficulty. One challenge per day over 30 days.
The platform just launched, and for the launch event you can get a specialize discount. To get the discount, follow this link.
Alright, that’s it for this today, bye!
Leave a Reply