Pretty sure Stallman would have some choice words to say about his idea of "software freedom" being twisted to include "relying on a corporation's subscription service for all your software needs and being unable to do anything yourself".
replyGive it a few months and open-weight self-hosted models will be able to do similar reverse engineering work. It is an extraordinary future we live in.
replyThat doesn't really solve "being unable to do anything yourself", though.
reply(It's strange to me how often the "Hacker News" commentariat seems not to value knowing things.)
This just plain won't happen. They'll either drive hardware prices through the roof, or heavily regulate open models. Radical changes are coming. It's a bleak, bleak future we live in.
replyPlease explain how this is more liberating than access to source code? Wouldn't this be reliant on continued and cheap access to the models? Even the open weight ones are heavily funded and not out of the kindness of their hearts. Just a few comments down is a thread about being blocked by the usual guardrails from the main providers.
replyBut this is access to source code. All source code is now essentially open. If not now, it will be in a few months. You aren't any more dependent on the models then before , you just have the source code of everything that compiles on your computer. What happens in the future is anyone's guess, but for all software published until this point in history, closed source no longer exists.
replyIt's "access" to "source code" as long as you can pay Anthropic/OpenAI and they don't raise their prices or lock you out.
replyNo I mean sure, if you want to decompile it yourself from scratch, but once anyone does, then the source code is available to anyone. So from that point onwards, no one is dependent on LLMs to decompile it again.
replyGLM 5.3, an open model, is also quite good at reverse engineering.
replyreally creeped out by the ai bro takes, specifically the ones "at least in a few months", almost as if there's no engineering principles anymore and just hype
replyEh, decompiling is happening now, at current LLM capabilities. The resulting codebase compiles into something that is functionally identical to the original. What engineering principles exactly do you need beyond "works exactly the same as the original product", in this particular context?
replyMy guess is we will be streaming everything in the future with whatever crap hardware they let us buy.
reply