> Systemically, I'm concerned that there is a lack of professional liability, rigorous industry best practices, and validation in the software industry which contributes to why we see Boeings flying themselves into the ground, financial firms losing everyone's data day in and out, and stories floating around our industry publications about people being concerned about the possibility of a remotely exploitable lunar lander on Mars.
> There's a heap of [comical?] tropes in the software industry that are illogical/counterproductive to the advancement of our profession and contribute to why other professions think software developers are a bunch of immature spoiled children that require constant supervision.
3 weeks ago you posted something titled "The future belongs to people who can just do things".
Today you post this:
> Because cli.mjs is close to 5mb - which is way bigger than any LLM context window out here. You're going to need baby sit it for a while and feed it reward tokens of kind words ("your doing good, please continue") and encourage it to keep on going on - even if it gives up. It will time out, lots...
I don't think you are someone who can just "do things" if you think a good way to de-obfuscate 5MB of minified javascript is to pass it to a massive LLM.
Why do you feel the need to be so rude about an interesting little blog post?
Obviously you don’t need an LLM to prettify obfuscated JavaScript. But take a look at the repo. It didn’t just add the whitespace back — it restored the original file structure, inferred function and variable names, wrote TypeScript type definitions based on usage, and added (actually decent) comments throughout the source code. That simply isn’t possible without an LLM.
Please link to a tool that can infer function/variable names and TypeScript type definitions from minified JS without using LLMs or requiring significant user input.
You didn't say it's impossible for a tool to do it. You said it's impossible.
Also, it's guessing at the names, guessing at type definitions, and it makes further guesses based on its previous ones, correct or no. If you don't already know what you're doing, you're in trouble.
For someone to claim that they're sharing the "decompiled" source of Claude Code for the public good is self-important back-patting nonsense, let alone misleading.
> I get the sense that you’re frustrated that something you’ve invested a lot of time and energy into learning is being automated. Maybe you’re scared because the technology has moved so quickly and your understanding of how to use it hasn’t kept up.
This is profiling/projection. You're incapable of responding to the GP's points so you're instead emotionally lashing out and attacking them. This is not really suitable for HN.
> But making someone’s day worse on Hacker News isn’t a good way to deal with that.
This suggests that you're incapable of distinguishing criticism of someone's work with personal attacks on them (furthered by the profiling that you tried to conduct above). Those things are not the same. If your day is ruined by someone posting reasonable criticism of an article that you personally submitted to HN, a place explicitly designed for intellectual curiosity, your expectations need to be adjusted.
OP is an adult who published his writing and posted it here himself for feedback. It's not all going to be positive.
> I get the sense that you’re frustrated that something you’ve invested a lot of time and energy into learning is being automated. Maybe you’re scared because the technology has moved so quickly and your understanding of how to use it hasn’t kept up
Wrong on all counts. I use LLMs to write code all the time, and I know how they work, which is why I find processing 5MB of JS through one to be an obscene waste of energy.
I do not use LLMs to publicly claim abilities I don't already have myself. Reading this article does not worry me one bit about my job security.
I get the feeling that you think they are not advancing their profession and that the quotes you have made of their work are such obvious examples of some problem with them that your parting question is some sort of stinging rebuke - is that correct? If so I have to admit I'm not following the connections.
> Systemically, I'm concerned that there is a lack of professional liability, rigorous industry best practices, and validation in the software industry which contributes to why we see Boeings flying themselves into the ground, financial firms losing everyone's data day in and out, and stories floating around our industry publications about people being concerned about the possibility of a remotely exploitable lunar lander on Mars.
> There's a heap of [comical?] tropes in the software industry that are illogical/counterproductive to the advancement of our profession and contribute to why other professions think software developers are a bunch of immature spoiled children that require constant supervision.
3 weeks ago you posted something titled "The future belongs to people who can just do things".
Today you post this:
> Because cli.mjs is close to 5mb - which is way bigger than any LLM context window out here. You're going to need baby sit it for a while and feed it reward tokens of kind words ("your doing good, please continue") and encourage it to keep on going on - even if it gives up. It will time out, lots...
I don't think you are someone who can just "do things" if you think a good way to de-obfuscate 5MB of minified javascript is to pass it to a massive LLM.
Do you think you are advancing your profession?