However, decompilation and recompilation are areas that AI could theoretically be extremely effective (due to the fact that much of it is based on educated guesswork in the first place, whether done by humans or not), and being able to turn proprietary software into plausibly “open source” software will be an effective tool for fighting back against big tech corporations (Microsoft, Google, Apple, Adobe, etc.) as well as for preserving classic, beloved software and games.
Listen, I’m as big of a generative AI hater as the next guy, if not bigger. I see myself as much as an artist and musician as a programmer, and I hate big tech for what I can only call the most egregious act of IP theft and cultural strip-mining in human history. It’s plagiarism on an industrial scale. It’s ethically wrong, economically a house of cards, and generally fucking sucks.
But at the same time, machine learning is an effective method of data-driven guessing and pattern recognition. Arguably it is the most effective method, which is why it’s much easier to use ML to make a program for things like image recognition than traditional algorithmic coding. Having recently read Chris Lewis’ blog on the decomp/recomp effort for Snowboard Kids 2 on the N64, it’s clear that LLMs are absolutely not a silver bullet of decompilation (at least, not yet), but it is almost an ideal problem space for what ML really does–guess shit based on a big pool of data.
So, my feeling is, if big tech companies are going to leverage these tools to steal from the people, then I think turnabout is fair play and I look forward to seeing the people leverage these tools for relative good. If people who sell art are going to be pushed out of the market due to slop generators, then it’s only fair for people to reverse engineer any and all of the software that matters in return.
No, sadly it wont.
However, decompilation and recompilation are areas that AI could theoretically be extremely effective (due to the fact that much of it is based on educated guesswork in the first place, whether done by humans or not), and being able to turn proprietary software into plausibly “open source” software will be an effective tool for fighting back against big tech corporations (Microsoft, Google, Apple, Adobe, etc.) as well as for preserving classic, beloved software and games.
Listen, I’m as big of a generative AI hater as the next guy, if not bigger. I see myself as much as an artist and musician as a programmer, and I hate big tech for what I can only call the most egregious act of IP theft and cultural strip-mining in human history. It’s plagiarism on an industrial scale. It’s ethically wrong, economically a house of cards, and generally fucking sucks.
But at the same time, machine learning is an effective method of data-driven guessing and pattern recognition. Arguably it is the most effective method, which is why it’s much easier to use ML to make a program for things like image recognition than traditional algorithmic coding. Having recently read Chris Lewis’ blog on the decomp/recomp effort for Snowboard Kids 2 on the N64, it’s clear that LLMs are absolutely not a silver bullet of decompilation (at least, not yet), but it is almost an ideal problem space for what ML really does–guess shit based on a big pool of data.
So, my feeling is, if big tech companies are going to leverage these tools to steal from the people, then I think turnabout is fair play and I look forward to seeing the people leverage these tools for relative good. If people who sell art are going to be pushed out of the market due to slop generators, then it’s only fair for people to reverse engineer any and all of the software that matters in return.