OpenAI is good at unminifying code • Hacker News Board

Re: OpenAI is good at unminifying code

#11

Post by j_maffe » Thu, Aug 29, 2024, 11:13 AM UTC

croes wrote: post #6

api wrote: post #4

Anyone working on decompiler LLMs? Seems like we could render all code open source. Training data would be easy to make in this case. Build tons of free GitHub code with various compilers and train on inverting compilation. This is a case where synthetic training data is appropriate and quite easy to generate. You could train the decompiler to just invert compilation and the use existing larger code LLMs to do things…

Unminifying isn't decompiling. It's just renaming variable and functions and inserting line breaks.

Minifying includes way more tricks than shorter variable names and removing white-space

Re: OpenAI is good at unminifying code

#12

Post by j_maffe » Thu, Aug 29, 2024, 11:14 AM UTC

LLMs are excellent at text transformation. It's their core strength and I don't see it being used enough.

Re: OpenAI is good at unminifying code

#13

Post by runiq » Thu, Aug 29, 2024, 11:15 AM UTC

Please let this one have knock-on effects on reverse engineering.

Re: OpenAI is good at unminifying code

#14

Post by tgv » Thu, Aug 29, 2024, 11:15 AM UTC

firtoz wrote: post #5

nwoli wrote: post #2

Hopefully it can help do this on emscripten files too and help adblockers dechipher obfuscated code for that purpose

Likewise for css class names

I can imagine that finetuning a model for this task could be very successful. Time for another AI startup.

Re: OpenAI is good at unminifying code

#15

Post by johndough » Thu, Aug 29, 2024, 11:15 AM UTC

api wrote: post #4

Anyone working on decompiler LLMs? Seems like we could render all code open source. Training data would be easy to make in this case. Build tons of free GitHub code with various compilers and train on inverting compilation. This is a case where synthetic training data is appropriate and quite easy to generate. You could train the decompiler to just invert compilation and the use existing larger code LLMs to do things…

> Anyone working on decompiler LLMs?

Here is an LLM for x86 to C decompilation: https://github.com/albertan017/LLM4Decompile

Re: OpenAI is good at unminifying code

#16

Post by bravetraveler » Thu, Aug 29, 2024, 11:16 AM UTC

LLMs are good at modeling and transforming text, news at 11. AI proponent hypes AI. I could go on, but I shouldn't have been this sarcastic to start with

Re: OpenAI is good at unminifying code

#17

Post by man4 » Thu, Aug 29, 2024, 11:16 AM UTC

[dead]

Re: OpenAI is good at unminifying code

#18

Post by roflmaostc » Thu, Aug 29, 2024, 11:17 AM UTC

BluSyn wrote: post #8

api wrote: post #4

Anyone working on decompiler LLMs? Seems like we could render all code open source. Training data would be easy to make in this case. Build tons of free GitHub code with various compilers and train on inverting compilation. This is a case where synthetic training data is appropriate and quite easy to generate. You could train the decompiler to just invert compilation and the use existing larger code LLMs to do things…

The potential implications of this are huge. Not just open sourcing, but imagine easily decompiling and modifying proprietary apps to fix bugs or add features. This could be a huge unlock, especially for long dead programs. For legal reasons I bet this will become blocked behavior in major models.

I've never seen a law forbidding decompiling programs. But, some programs forbid to decompile applications by the license agreement. Further, you still don't have any right on this source code. It depends on the license...

Re: OpenAI is good at unminifying code

#19

Post by albert_e » Thu, Aug 29, 2024, 11:17 AM UTC

Should the title say ChatGPT or gpt-4 (the model) instead of OpenAI (the company)?

Re: OpenAI is good at unminifying code

#20

Post by jraph » Thu, Aug 29, 2024, 11:17 AM UTC

api wrote: post #4

Anyone working on decompiler LLMs? Seems like we could render all code open source. Training data would be easy to make in this case. Build tons of free GitHub code with various compilers and train on inverting compilation. This is a case where synthetic training data is appropriate and quite easy to generate. You could train the decompiler to just invert compilation and the use existing larger code LLMs to do things…

> Seems like we could render all code open source

Unfortunately not really. Having the source is a first step, but you also need the rights to use it (read, modify, execute, redistribute the modifications), and only the authors of the code can grant these rights.