If we go back to the time five years ago, who was the first to develop language AI and big data says that is not clear. At that time Matlab, Scala, R, Java will be someone to their stations, but the time came now, it is no longer suspense, especially from the Pytorch after Facebook Open Source, locking the first card trends are becoming evident. If we want to debate who is just second question.
However, before the market was some noise that scientific data or java to start from, because some of Hadoop and other big data infrastructure is written in java, developers who want to join AI and big data industry, the eggs in Python this basket is not only safe, but necessary. Or another way of saying, if you want to mix in the future in this industry, what do not want, to close their eyes to learn Python.
Of course, Python is not without its problems and weaknesses.
It grammatically separate faction, was not used to so many veterans; "naked" Python is slow on different tasks slower than the C language about several times to several thousand times vary; since the global interpreter lock (GIL ) limit, a single Python programs can not execute concurrently on multicore; Python 2 and Python 3 versions of two long parallel, many modules need to maintain two different versions, developers choose to bring a lot of unnecessary confusion and trouble ; the absence of any control of any company, nor has always been a technology giant is willing to die quite Python. So, with respect to the application of a wide range of Python, its core infrastructure to get input and support it is actually very weak.
Until today, 26-year-old Python is not yet an official standard JIT compiler, by contrast, Java language in the first three years after its release he received a standard JIT.
Python is such a variety of problems with the first phalanx of ××× rushed to hand, but even to a few years ago, not many people believe it has a chance for the title, many people think Java's position can not be shaken, but also Some people say that all programs will be rewritten in Java.
But today we look at, Python is already a first language and data analysis of AI, network *** *** first language, is becoming the first introductory programming language teaching, cloud computing management system first language. Python has long become one of the Web development, game script, computer vision, networking management and smart phone robot development, and constitute a CRM system and solve the problem of telephone robot really easy to use mainstream language, along with Python users can expect growth, it also has the opportunity to climb to the top in multiple areas.
Take AI, we must first ask, AI's main crowd where? If we talk about the topic of static today, you might think that AI is the main research institutions of AI scientists, has a PhD in machine learning algorithm specialists and experts.
Like Kai-fu Lee's "AI dividend syllogism" clearly tells us not to be limited to the front of the AI market, appropriate long-term view point, looking back three to five years, you will see the whole AI industry, from a smart phone robot, eco-smart city, smart voice, the working population will gradually form a huge pyramid structure, the above is just the tip of AI scientists a little bit, 95% or more of the AI technician, will be AI engineers, application engineers and AI tools users.
I believe these people will be almost a clean sweep Python, Python has become a huge reserve army camp.
These potential Python users still outside technology circles, but with the development of AI applications. Millions of people are teachers, staff, engineers, translators, editors, doctors, sales managers and civil servants will coerced their industry knowledge and data resources in the field, the influx of AI in Python and the tide of profound change across IT, or the overall structure and appearance DT (data Technology) industry.
Python also help AI and Data Science, climb to the top position of the ecological chain programming language. AI tied together with Python, for them, whether it is e-commerce, search engines, social networks or smart hardware, data cows are just the future, electronic nerve and ecological chain downstream implementation tools, we will take orders on their own.