Switching from Python to anything is always a performance enhancement on the web. Go is fantastic, it just works and it’s easy to use. It’s opinionated nature will turn people off, but companies will love it.
Edge cases, at best. I can give you one reason why you won't see this in most cases: Libraries. No language has a larger community group than Python. Never has been one, either.
GC cannot start its work until all goroutines are stopped(stop the world). GC has to wait until all running goroutines reach breakpoint, increasing GC time. I think Discord faced this issue as GC was taking so much time.
Interesting that putting in the time to rewrite Python applications to Go is worth it to these companies, I'm curious if Python can close the performance gap some more with the release of 3.11. Just curious if there is a Go equivalent of all the commonly used Python ML libraries.
Python can never surpass compiled languages but python will be still be used for data analysis and ML. The repl is perfect for any kind of scientific experimentations
Python ML libraries are already either written in C or in Fortran. If rewriting the whole code base is not an option, Python already has a plethora of other options to make it faster (Cython, numba, JIT compiled implementations, mypyc, etc.). But the biggest problem, I think, is the lack of talent pool for this.
Not agreed with your statement that go is best language. There is no best programming language. Every language pros and cons. And Some companies shifting go lang to rust. It doesn't mean rust is the best language.
Faster development/poc is more important to test business idea, and to capture market earlier. Then use C# or Java for better design and performance. There is no need for GO.
If you want to build a backend for your application just go with django first then if you need performance switch to go.why ? cause there is alot of python code out there if you get stuck on something.correct me if i'm wrong
What about Java to go? Or Node to go? I think python will survive due to data science but if someone starts building Pandas, NumPy, or PyTorch for Go it would be the nail in the coffin for python.
You are being silly. Pandas, NumPy etc. internally are built in C, C++, and in some cases, even FORTRAN. The reason Python was adopted to be a wrapper language/interface around this was due to its ease of use, quick prototyping/development, which made it especially easy for non-software developers (scientists, data analysts etc.). The only language around right now that could replace this stack in the future, is Julia.
@@burnere633 that would make things easier actually, all we need to wait for is first class support or binding for go from these popular libraries and things would shift greatly.
@@burnere633 yeah but that’s the catch isn’t it? Python is easy to work with but is dog slow. If it was a silver bullet we’d all be using nothing but python. I just don’t see data science and data engineering not being folded into the software engineer role instead of it remaining a specialty. So why not use a simple updated language that you can write and api for in one sprint then build a data lake the next?
@@crimsionCoder42 If you think that's happening you're highly underestimating what's happening to data. Because of AI, managing the ins and outs of the data network itself is becoming a very demanded specialty very comparable to the world 20 years ago where software was being built as an assistant, companion and general simplification tool to everything.
With all the respect to your work mate. The best programming language in the world is Python. And no one other programming language comes even close to it. Speed is not everything.
Yeah it is matter of use-case and cost, whether cost is "its more efficient and better tooling to write this in python" OR it is "it is more efficient by performance so i can reduce scaling issues and GCE/AWS cost of business"
I want your input, I'm a C# developer, I tried to recreate one of my apps in nodejs, but the amount of time it took me is massively lower compared to C#. I'm wondering if it's the same story with python vs go, I know python is definitely easier, however, by how much? is it really worth it to start with go? Because honestly I don't think it's that important to value performance over speed of development for startup companies and such.
Python is much easier to learn if you're coming from C#. If you haven't learned C, learning Go will be much more challenging in terms of understanding when and where to use pointers or not, Go generics have a strange syntax and the lack of dynamic typing is frustrating, and Go structs are much more tedious to use and write than making a simple class in Python. If you really need the performance, most people use Cython or gmpy, along with numerous other fast-math libraries.
@@Zero_Hour thank you for the reply, I actually learned some go, it has really clever shortcuts and its much easier than I initially thought, but I have not coded anything useful yet, however I coded a little bit of python. Your point hit home tbh, If performance is an issue, using cython or numpy most likely will be enough, and there is a abundance of tutorials online.
Okay , many languages faster then python. In java you can create a single jar file for deployment too. The main point of all this what python is good for faster development but not an option in long term run.
📝Get your *FREE Golang Cheat Sheet* -
golangdojo.com/cheatsheet
I love things that confirm my cognitive bias
well that's social media for you, ain't it...
😀😀😀👌👌👌
Same
Everyone does
Isn't it??
Thank you for this video. I would like to see more Python to Go case studies.
UA-cam one of the best use cases from python to go
Switching from Python to anything is always a performance enhancement on the web. Go is fantastic, it just works and it’s easy to use. It’s opinionated nature will turn people off, but companies will love it.
This is cue for Let's Get Rusty to slip in and make a video about why some companies are migrating from Go to Rust.
Edge cases, at best. I can give you one reason why you won't see this in most cases: Libraries. No language has a larger community group than Python. Never has been one, either.
JavaScript is definitely bigger
@@williamross7552 Python is general purpose language so it has a library for almost everything whereas Javascript is mostly concentrated on web only.
"Not getting too creative" I think that's the BEST and MOST IMPORTANT thing with Go that's taken for granted.
One of Go's philosophy is being simple, readable and doing things in one way.
Any company migrate from Java/Spring to Go?
I really like Go, but would also like to see why companies are switching from Go to another language (like Discord due to the issue with GC)
GC cannot start its work until all goroutines are stopped(stop the world). GC has to wait until all running goroutines reach breakpoint, increasing GC time. I think Discord faced this issue as GC was taking so much time.
Discords case is unique. I don’t know if you actually read the blog you should actually give it a read bud.
Migration not equal to first time to market. After your customer grow switch to more reliable language. Python great for MVP.
Form for cheet sheet still broken, never receive the email. Not in spam either.
go is better than rust coz it's simple to learn and fast best of both world
@@UlrikBrun-ir7mh memory safety isnt a concern with GC languages. Thats a c/c++ thing
Interesting that putting in the time to rewrite Python applications to Go is worth it to these companies, I'm curious if Python can close the performance gap some more with the release of 3.11.
Just curious if there is a Go equivalent of all the commonly used Python ML libraries.
Python can never surpass compiled languages but python will be still be used for data analysis and ML. The repl is perfect for any kind of scientific experimentations
It may possible if python use strict type and compile.
But python is not designed like that.
But learning python is easy for beginner.
@@n000d13s for experimentations it's good definitely, but if performance counts wouldn't Go be able to handle predictions by a model a lot better?
Python ML libraries are already either written in C or in Fortran.
If rewriting the whole code base is not an option, Python already has a plethora of other options to make it faster (Cython, numba, JIT compiled implementations, mypyc, etc.). But the biggest problem, I think, is the lack of talent pool for this.
Python ML libraries are written in C or C++ so they are super performant
Man, you really made me like Go language, many thanks
Thanks for introducing me to Go!
You do realize that go is just braced python right?
Let's Go!
no, they're not
Not agreed with your statement that go is best language. There is no best programming language. Every language pros and cons.
And
Some companies shifting go lang to rust. It doesn't mean rust is the best language.
Exactly.
👍
Give me example where backend was moved from go to rust ? Rust development is costly and it's complicated. E.g. Go has GC where as Rust doesn't
@@sumer9999 I think Deno moved from go to Rust.
Faster development/poc is more important to test business idea, and to capture market earlier. Then use C# or Java for better design and performance. There is no need for GO.
except from Go performs better than java in serverless/microservices which are becoming more common
If you want to build a backend for your application just go with django first then if you need performance switch to go.why ? cause there is alot of python code out there if you get stuck on something.correct me if i'm wrong
What about Java to go? Or Node to go? I think python will survive due to data science but if someone starts building Pandas, NumPy, or PyTorch for Go it would be the nail in the coffin for python.
You are being silly. Pandas, NumPy etc. internally are built in C, C++, and in some cases, even FORTRAN. The reason Python was adopted to be a wrapper language/interface around this was due to its ease of use, quick prototyping/development, which made it especially easy for non-software developers (scientists, data analysts etc.). The only language around right now that could replace this stack in the future, is Julia.
@@burnere633 that would make things easier actually, all we need to wait for is first class support or binding for go from these popular libraries and things would shift greatly.
@@burnere633 yeah but that’s the catch isn’t it? Python is easy to work with but is dog slow. If it was a silver bullet we’d all be using nothing but python. I just don’t see data science and data engineering not being folded into the software engineer role instead of it remaining a specialty. So why not use a simple updated language that you can write and api for in one sprint then build a data lake the next?
@@burnere633 Go is easier than Python.
@@crimsionCoder42 If you think that's happening you're highly underestimating what's happening to data. Because of AI, managing the ins and outs of the data network itself is becoming a very demanded specialty very comparable to the world 20 years ago where software was being built as an assistant, companion and general simplification tool to everything.
next check Java vs Go
?! From 100 ms to 10ms is a 1000% performance increase. Not 90% xD
kids who are saying rust is better than go they should know it comes with a cost though lmao :)
What cost?
i need comparison between Django rest api security and go rest api security
Now do "Why are Companies Migrating from Go to Rust!?"
Ive only seen discord do it tbh
only discord did that because of the go garbage collector, other garbage collected languages would have the same problem discord had with go
@@tjgdddfcn Can't a guy just say something over-the-top in the Internet and not be nitpicked at?!
Wrong channel.
Rust is only used for planet scale application with billions of users, application development is costly since it's more time consuming,
Where to find this companies that hire jr and internship?
With all the respect to your work mate. The best programming language in the world is Python. And no one other programming language comes even close to it. Speed is not everything.
That's a naive statement.
Nothing like " Best programming language "
Use the tool that fits your use case
Speed is very important, if my company spends 100K $ on compute cost, then with Go it will reduce to less than 50k$
Yeah it is matter of use-case and cost, whether cost is "its more efficient and better tooling to write this in python" OR it is "it is more efficient by performance so i can reduce scaling issues and GCE/AWS cost of business"
@@sumer9999i think its just the problem of your company that a couple extra milliseconds affects your cost by double.
So, to avoid confirmation bias, let's look at people that already decided to go towards go, okay... :D
Thanks 🙂
are you still working with java?
Golang seems pretty damn fast.
I want your input, I'm a C# developer, I tried to recreate one of my apps in nodejs, but the amount of time it took me is massively lower compared to C#. I'm wondering if it's the same story with python vs go, I know python is definitely easier, however, by how much? is it really worth it to start with go?
Because honestly I don't think it's that important to value performance over speed of development for startup companies and such.
Python is much easier to learn if you're coming from C#. If you haven't learned C, learning Go will be much more challenging in terms of understanding when and where to use pointers or not, Go generics have a strange syntax and the lack of dynamic typing is frustrating, and Go structs are much more tedious to use and write than making a simple class in Python. If you really need the performance, most people use Cython or gmpy, along with numerous other fast-math libraries.
@@Zero_Hour thank you for the reply, I actually learned some go, it has really clever shortcuts and its much easier than I initially thought, but I have not coded anything useful yet, however I coded a little bit of python. Your point hit home tbh, If performance is an issue, using cython or numpy most likely will be enough, and there is a abundance of tutorials online.
The SaaS company I work for is moving to go
This video is pure propaganda
Hell yes !! there is no languege better than other ! choosing a language it depends what you are trying to do !
Google tried to buy Python 15 years ago, buy the founder didn't sell it, so Google created Go and now it's better than Python.
That is not right
@@tommilutinovic1188😂😂😂 All these wild theories
Why would they buy a programming language.
Concurency
don not say something lke that. just stop
Okay , many languages faster then python. In java you can create a single jar file for deployment too. The main point of all this what python is good for faster development but not an option in long term run.
why companies are migrating to rust lol
Yoooooooo 😂
Why do these metrics matter. All these apps are slow lol