- Open source allows for the inspection and modification of the source code, unlike compiled and closed software.
- Licenses (such as MIT or Apache) regulate the legal use, distribution, and commercial application of these programs.
- Open-source AI democratizes technology by allowing efficient models to compete with proprietary systems.
- Verifying authenticity involves checking licenses and using vulnerability analysis tools.
You've probably experienced wanting to install a new tool on your computer and coming across the term " open source ," but not being entirely sure what it really means or how to verify the safety of what you're downloading. In the world of computing, this is crucial to avoid problems with your operating system or compromising the privacy of your data.
To understand this properly, we need to look back. In the nineties, it was common to buy discs with pre-compiled software —that is, instructions that the computer understands perfectly (like bytecode or opcodes) but which are a complete enigma to a human. This opacity ensured that no one could steal the code or analyze how the program worked internally.
The difference between closed-source software and open source
Unlike proprietary software, open-source communities operate on a different principle: they voluntarily share their source code. The logic is that collaborating and improving software collectively is far more cost-effective than trying to hide it. In fact, anyone can examine the code to learn, fix bugs, or even fork it — which is simply creating their own version by adding new features.
Note that just because it's open source doesn't mean anything goes. The authors define specific licenses (like MIT or Apache 2.0) that dictate what you can do with the code. Some are very permissive, while others allow students to use it for free but prohibit commercial use in companies.
How to verify the legitimacy of a program
If you come across an extension, like Vim-type keyboard shortcuts in Firefox sometimes do, and you see that it has an MIT license but retains the copyright, don't worry. The MIT license is one of the most common in the open source world and confirms that the code is accessible. To be absolutely sure that a program is genuine, it's best to inspect the repository where the code is hosted and verify that the license allows you to study and modify it.
The leap towards Open Artificial Intelligence
This philosophy has taken hold in AI. According to the Open Source Initiative, open-source AI is that which allows anyone to study its workings, modify its components, and share the results. This is key to preventing a few companies from controlling everything and to ensuring that technological progress is democratic and transparent.
Today we have very powerful proprietary models like GPT-4 or LaMDA, but there are open-source alternatives that are making waves. For example, IBM's Granite or Meta's Llama models allow developers to customize the tool to their needs. A stunning example is DeepSeek with its R1 model , which has demonstrated that competitive AI can be trained with a fraction of the budget of the commercial giants, using more affordable hardware.
Security and analysis tools
For developers, it's not enough for the code to be open source; they need to ensure it's free of vulnerabilities. Tools like Snyk Code analyze code before each commit, detecting critical vulnerabilities and errors that could be exploited by malicious actors, thus guaranteeing that code transparency goes hand in hand with security.
Having access to the source code, whether in a simple browser extension or a complex language model, allows us to audit security and foster collective innovation, breaking down the barrier of closed executables and allowing technology to evolve more openly and efficiently for all users.