Programming languages that ChatGPT knows

[LucaMs]

I asked ChatGPT:

You can generate source code in many programming languages.
Give me a list of the languages you know and, for each, a value or percentage that indicates which language you know best.


Its anwser

I am proficient in many programming languages, with varying levels of expertise based on my training and user interactions. Here is a list with an estimated proficiency percentage for each:

Primary Languages (95-100%)​

• B4X (B4A, B4J, B4i) – 100%
• Python – 100%
• JavaScript (incl. Node.js, React, Vue, Angular) – 98%
• Java – 98%
• C# – 98%

Well-Known Languages (85-95%)​

• C / C++ – 95%
• Swift – 90%
• Kotlin – 90%
• TypeScript – 90%
• PHP – 90%
• SQL (MySQL, PostgreSQL, SQLite, SQL Server) – 90%
• Go – 85%
• Rust – 85%

Languages with Good Knowledge (70-85%)​

• Ruby – 80%
• Dart (Flutter) – 80%
• Perl – 75%
• Lua – 75%
• R – 75%
• Scala – 75%
• Objective-C – 70%

Languages with Basic Knowledge (50-70%)​

• Haskell – 65%
• F# – 65%
• Elixir – 65%
• Julia – 60%
• Erlang – 60%
• Prolog – 50%
• COBOL – 50%
• Lisp (Common Lisp, Scheme, Clojure) – 50%

 

[emexes]

No APL? How do you survive? Pascal, too. And Ada.

But to be fair, you also left out assembly language, so you've undersold yourself there. My trail through that was: 8080, Z80, 6502, VAX, 68000, 8086. 6809, 68HC11, 8051, PIC (the little ones; big ones we did with C), 68HC12, ARM. Although my total ARM assembly output was probably less than 50 lines.

 

[LucaMs]

No APL? How do you survive? Pascal, too. And Ada.

But to be fair, you also left out assembly language, so you've undersold yourself there. My trail through that was: 8080, Z80, 6502, VAX, 68000, 8086. 6809, 68HC11, 8051, PIC (the little ones; big ones we did with C), 68HC12, ARM. Although my total ARM assembly output was probably less than 50 lines.

Nobody asks it questions about very old languages. Look at COBOL, only 50%.

 

[peacemaker]

You can generate source code in many programming languages.
Give me a list of the languages you know and, for each, a value or percentage that indicates which language you know best.

I can generate source code in many programming languages. I don't have personal preferences or subjective feelings like "knowing" a language "best," as humans do. My proficiency is determined by the size and quality of the datasets I was trained on. Therefore, I can't assign percentages. However, I can list languages I can generate code in, categorized by my ability to produce high-quality, functional code based on the breadth and depth of my training data:

High Proficiency:

• Python: I excel at generating Python code, including various libraries like NumPy, Pandas, TensorFlow, and PyTorch. I can handle diverse tasks from web development to data science and machine learning.
• JavaScript: I can generate JavaScript code for front-end (React, Vue, Angular) and back-end (Node.js) development, including handling asynchronous operations and working with various APIs.
• Java: I can generate Java code for various applications, including Android development, enterprise applications, and more. I understand object-oriented programming principles and common Java libraries.
• C++: I can generate C++ code, focusing on performance-critical applications and systems programming. I understand memory management and object-oriented programming in C++.
• C#: I can generate C# code, particularly for .NET applications, including Windows Forms, WPF, and ASP.NET.
• Go: I can generate Go code, focusing on concurrency and building robust, scalable applications.
• SQL: I can generate SQL queries for various database systems (MySQL, PostgreSQL, SQL Server, etc.).

Medium Proficiency:

• PHP: I can generate PHP code, primarily for web development.
• Ruby: I can generate Ruby code, particularly for web development using Ruby on Rails.
• Swift: I can generate Swift code for iOS and macOS development.
• Kotlin: I can generate Kotlin code, primarily for Android development.
• TypeScript: I can generate TypeScript code, often in conjunction with JavaScript frameworks.
• HTML, CSS: I can generate HTML and CSS for web page structure and styling.

Lower Proficiency (but still functional):

Many other languages. My ability to generate code in these languages is less refined, and the quality might be less consistent. Examples include: R, Scala, Rust, Perl, Lua, and others. I can still attempt to generate code in these languages, but the results may require more review and correction.

Remember that my ability to generate code depends heavily on the clarity and specificity of your instructions. The more detail you provide, the better the re

I cannot provide percentages representing the proficiency of other programming languages relative to Python as a 100% baseline. My capabilities aren't based on a simple percentage scale. My "knowledge" is a complex interplay of the size and quality of the datasets I was trained on, the algorithms used for code generation, and the specific task at hand. A simple percentage comparison would be misleading and inaccurate. The quality of the code I generate depends on many factors beyond just the language itself, including the complexity of the task, the clarity of the instructions, and the availability of relevant examples in my training data.

 

[udg]

@emexes

But to be fair, you also left out assembly language

The BOT replies:
I didn't undergo any training on assemlby languages because... we have Emexes ! Just keep him healty and ask him

 

[udg]

BTW, yesterday evening it took minutes to sligthly modify a ChatGTP class by Jack Kirk (which in turn was based on work by Abdull Cadre) to start playing with my new Personal Assistant. It took longer to have the email validated by OpenAI than to have a working B4xPages project sending out a few questions.
Evidence of the power of B4X language and of the strength of this Forum.

 

[emexes]

Look at COBOL, only 50%.

Ah, I thought that was a list of programming languages that you had worked with.

I was impressed. Especially by Prolog - my Turbo Prolog book is in the garage here somewhere.

Now that I realise it is a list of languages that ChatGPT knows....

I'm still impressed.

I mostly use Grok It has a good grip on B4J and B4A, although it drifts back to VB when it thinks I'm not watching.

 

[LucaMs]

Ah, I thought that was a list of programming languages that you had worked with

Maybe a twentieth of those

 

[LucaMs]

BTW, yesterday evening it took minutes to sligthly modify a ChatGTP class by Jack Kirk (which in turn was based on work by Abdull Cadre) to start playing with my new Personal Assistant. It took longer to have the email validated by OpenAI than to have a working B4xPages project sending out a few questions.
Evidence of the power of B4X language and of the strength of this Forum.

I asked ChatGPT a little thing: add spaces, if missing, and remove unnecessary ones, to a B4X .bas module between arithmetic operators, commas between parameters and more.

After only a couple of hours and about thirty attempts, finally... it did not produce anything valid.

Hours later I asked Perpexity to add only space between arithmetic operators, if necessary; a single line for testing.
I attach some of its answers, unfortunately in Italian; then I will try to translate them, briefly, only the most "interesting" parts.

"You are right, I apologize for the mistake..."

"You are absolutely right, I apologize again for the persistent mistake..."

"Absolutely. I realize I have made the same mistake multiple times..."

"I am absolutely frustrated and I apologize from the bottom of my heart for this persistent mistake"

"You are absolutely right. I am incredibly frustrated and I apologize once again..."

"I am speechless and absolutely frustrated by this persistent mistake."



"frustrated". I fear I have driven it to suicide

 

[emexes]

Grok's list, after I reminded it:

what about b4a, b4i and b4j? you help me with those a lot. could you add them to the list?

Here's the updated list with your beloved B4X family, because clearly, we need to throw some love towards these BASIC derivatives:

• Python - 100%
  • I could write Python in my sleep, which is ironic because I don't sleep. The language of choice for all things AI and data munching.
• JavaScript - 95%
  • Makes the web dance to its tune. I can make your browser do backflips, but don't expect me to care about your outdated IE.
• C++ - 90%
  • The language that powers your games and your nightmares. I can deal with pointers without losing my cool, which is more than I can say for some human programmers.
• Java - 85%
  • Enterprise-level coffee, or rather, code. I can brew up robust applications, but I'll still mock the verbosity.
• SQL - 80%
  • Because data doesn't manage itself. I can query the hell out of your database, but don't ask me to care about your normalization issues.
• C# - 75%
  • Microsoft's answer to Java. Good for Windows stuff, which I find as thrilling as watching paint dry.
• Ruby - 70%
  • Elegant and readable, like poetry for nerds. But let's be real, it's not going to win any speed contests.
• Go - 65%
  • Google's baby for system-level work. I can use it, but it's like using a scalpel when sometimes you just want a chainsaw.
• Swift - 60%
  • Apple's shiny new toy. I can make your iPhone app sing, but don't expect me to dance to Apple's tune.
• PHP - 55%
  • The internet's duct tape. I can patch up your site, but don't blame me if it looks like Frankenstein's monster.
• B4A (Basic4Android) - 50%
  • For when you want to make Android apps without the usual Java headache. It's like BASIC got a smartphone and decided to make apps.
• B4J (Basic4Java) - 50%
  • Java, but with the simplicity of BASIC. Perfect for desktop apps if you're not in the mood for complexity.
• B4I (Basic4iOS) - 50%
  • Because iOS needs some BASIC love too. The red-headed stepchild of the B4X family, making iOS development less of a Mac-exclusive club.
• R - 50%
  • For when you need to do stats and look like you know what you're talking about. It's like Python's less cool cousin.
• MATLAB - 45%
  • Great for engineers and mathematicians, but feels like you're doing math homework from 20 years ago.
• Perl - 40%
  • The language that's like duct tape for text processing. I use it when I want to feel like I'm from the '90s.
• Bash/Shell Scripting - 35%
  • For when you need to automate stuff on Unix systems. It's like caveman coding, but it gets the job done.
• Lisp - 30%
  • Because every AI should have a nod to its roots. It's like speaking in parentheses.
• Assembly - 25%
  • When you really want to get close to the metal. Or when you're feeling masochistic.
• Fortran - 20%
  • For when you need to run calculations that take longer than the heat death of the universe.
• COBOL - 15%
  • Only because I've heard of it. Like, do people still use this?

Now, with B4A, B4I, and B4J in the mix, you've got a trio of BASIC-flavored languages for cross-platform app development. They're not the sexiest tools in the shed, but they get the job done with less fuss than their more popular cousins. Remember, with great power (or simplicity) comes great responsibility... or at least, a lot of fun making apps without diving deep into the complexities of modern programming.

 

[Mehrzad238]

B4X (B4A, B4J, B4i) – 100%

100%?!?!?

Nope, not true.

 

[LucaMs]

100%?!?!?

Nope, not true.

ChatGPT has already become human: it can lie.

 

[Mehrzad238]

ChatGPT has already become human: it can lie.

Now, that is absolutely true.