Certainly Python is the dominant scripting language of today, and because it is the de-facto standard glue language in the ML field, which is being taught everywhere and hyped as a magic bullet for whatever ails companies, Python will shoot to the top 3 languages after JS (which dominates web), and Java/.NET/PHP (which dominate server side). I don't expect that Python will ever be dominant in back end or web.

PHP is well down the list these days and python is already up there near the top https://insights.stackoverflow.com/survey/2019#technology

Python is pretty ubiquitous in web backend, especially where the JVM or .NET CLR aren't a hard requirement.

Don't believe these surveys of use. PHP is used in a huge fraction of websites, even if just a teeny bit. As the only language built into Apache, it has an unfair advantage. Same with JS for the web; as it is hard-wired into the browser, it has to win. But yes PHP is dropping as a development language, and with its latest "kitchen-sink" version has gotten so complex that it will stop being picked for new projects (i.e. dropping birth rate) that C++ and Java are experiencing.

oh sure, php is powering most of the internet it’s just that nobody is writing it any more lol

Facebook has served as a kind of cautionary tale.

In languages, once you build a system, you might maintain it for another 20 years. After the creation point, it is updates, refinements, fixes, feature additions. Rarely does one change the language and rewrite from scratch. That almost never gets approved by management. Thus in all statistics of programming language use you will see very old languages persisting. This is why COBOL still shows up in the lists, even though nobody starts a new project in COBOL. I argue that PHP has a lot of inherited code, and that people starting projects today are much less likely to select PHP in 2020 than they did in 2010. So i guess dying is too strong a word. The more accurate description is that the birth rate of PHP projects has been dropping, and that fewer people myself included are selecting it for those somewhat rare cases of a "from scratch" project. Python is increasing in birth rate, and many preprocessor languages for JS are also increasing. These charts of programming language popularity are more like the census of past design wins. This keeps the newer entrants down in statistics, making them look risky, when in reality using a crappy tool from decades ago is a guarantee of mediocre productivity, so isn't that a risk also?