0x00
Some time ago, another programming language dark horse suddenly rushed into the field of vision of developers and officially opened for download.
Its name is Mojo, and I believe many friends saw it at that time.
As stated on the official website, Mojo is a programming language for AI developers, with a syntax a bit like Python.
According to Mojo's official website, it combines the ease of use of Python with the high performance of the C language, unlocking the programmability of AI hardware and the scalability of AI models.
0x01
And just a few days ago, Mojo actually made a big move, officially announcing the official open-source of its standard library.
Mojo officials said that in order for Mojo to fully realize its potential, open source is the only way to go, and said that the core modules of the Mojo standard library have been open-sourced and released under the Apache 2 license.
At present, Mojo officially allows developers to submit external contributions through GitHub PRs, encouraging developers to participate in the development and improvement of Mojo, so as to promote the growth of the community.
In addition to the release of the open source code, the official release also released the complete commit history of the Mojo standard library, so that developers can trace the evolution of the code and better understand the background of the code.
That's not all, the official also gave a special guide to Mojo open source contributions.
It can be seen that Mojo officials really attach great importance to open source and community.
0x02
Mojo seems to be quite capable of playing, so which company pushed it out?
After taking a look, I realized that Mojo is a new programming language launched by the artificial intelligence company Modular.
Modular, on the other hand, is a very young nascent AI startup, founded in 2022 by Chris Lattner and Tim Davis.
When it comes to these two founders, I believe some students also know about them, and they are all top experts in the industry.
Among them, Chris Lattner is also known as the "father of LLVM" and "the father of Swift", and has led the construction of AI and core systems in many well-known technology giants such as Apple, Google, and Tesla.
Modular's vision is ambitious, with the goal of reinventing AI infrastructure from the bottom up.
Previously, Modular AI had raised hundreds of millions of dollars in funding, which is a remarkable achievement for a start-up company that has just been born.
In addition, in the list of investors on the company's official website, it can be seen that many well-known investment institutions in the field of AI have participated.
0x03
According to the official website, Mojo is a programming language with several obvious features.
1. First of all, the performance aspect.
Mojo takes full advantage of the features and functionality of the hardware, including multi-core, vector and accelerator units, as well as advanced compiler and heterogeneous runtime mechanisms, to achieve performance comparable to C++ and CUDA without adding complexity.
In the area of parallelization, Mojo leverages MLIR to enable Mojo developers to take full advantage of vectors, threads, and AI hardware units.
2. The second is interoperability.
As we all know, today, the Python ecosystem is extremely prosperous, with countless functions, libraries, frameworks, models, tools, and so on.
Mojo, on the other hand, has access to the entire Python ecosystem. With Mojo, for example, libraries like Numpy and Matplotlib can be seamlessly plugged and mixed into your code.
3. The second is scalability.
Scalability is also an advantage for Mojo. Mojo can upgrade existing operations in the user model so that developers can easily extend the user's model with pre-processing, post-processing, custom substitution, and other operations.
0x04
Mojo was originally released in early May last year, and within a few months of its launch, it had formed a basic scale and ecosystem.
Soon after, Modular's official website announced that Mojo was officially available for download, starting with Linux, and adding support for Mac in subsequent iterations, as well as Windows support in the pipeline.
This also means that developers can experiment and write their own Mojo code through the Mojo SDK.
Shortly after Mojo was officially announced to be available for download, an open-source author named Aydyn Tairov made a breakthrough attempt with Mojo.
The author had previously ported the popular C-only llama2.c project on GitHub to a Python-based llama2.py.
This time, Aydyn Tairov ported llama2.py to llama2.mojo, and the result was very unexpected, with a nearly 250-fold performance improvement after the port.
Even so, the author still thinks there is some room for improvement in it.
Seeing Mojo's performance, many netizens said that Python has encountered a strong opponent this time, and Mojo may even replace Python in the future?
In response, Chris Lattner, CEO of the company, directly responded:
Mojo is not a threat to Python, on the contrary, it helps Python developers become stronger.
At the end of the article, the relevant address is also attached, and interested partners can try it.
- Mojo Official Website: https://www.modular.com/mojo
- Mojo Docs: https://docs.modular.com/mojo
As for how this programming language will evolve in the next AI era, we can wait and see.