Comment by tntxtnt
2 months ago
I get why Microsoflt loves AI so much - it basically devour and destroy open source software. Copyleft/copyright/any license is basically trash now. No one will ever want to open source their code ever again.
2 months ago
I get why Microsoflt loves AI so much - it basically devour and destroy open source software. Copyleft/copyright/any license is basically trash now. No one will ever want to open source their code ever again.
It fits perfectly with Microsoft's business strategy. Steal other people's ideas, implement it poorly, bundle it with other services so companies force their employees to use it.
I'm so mad Teams exists
I really think that if Microsoft would be forced to improve user experience of Teams, it would lead to measurable impact when it comes to happiness of humankind.
Not just code. You can plagiarize pretty much any content. Just prompt the model to make it look unique, and that’s it, in 30s you have a whole copy of someone’s else work in a way that cannot easily be identified as plagiarism.
I struggle to find this argument compelling, as it sounds more of a straw man argument than a legitimate complain.
If I write a hash table implementation in C, am I plagiarizing? I did not come up with the algortithm nor the language used for implementation; I "borrowed" ideas from existing knowledge.
Lets say I implemented it after learning the algorithm from GPL code; is ky implementation a new one, or is it derivative?
What if it is from a book?
What about the asm upcodes generated? In some architectures, they are copyrighted, or at least the documentation is considered " intellectual property"; is my C compiler stealing?
Is a hammer or a mallot an obvious creation, or is it stealing from someone else? What about a wheel?
> I struggle to find this argument compelling, as it sounds more of a straw man argument than a legitimate complain.
Dude, there are entire websites dedicated to using diffusion models to rip off the styles of specific artists so that people can have their "work" without paying them for it.
You can debate the ethics of this all you want, but if you're going to speak on plagiarism using generative AI, you should at least know as much as the average teenager does about it.
6 replies →
There is still value in quality and craftsmanship. You might not be of that opinion, and you might not know anyone who is, but I do.
When I get an obviously AI-generated response from someone I'm trying to do business with, it makes me think less of them. I do value genuine responses, far more than the saccharine responses AI comes up with.
1 reply →
There will always be a market for niche, high quality electron tweaking. Thing is, it will be a highly competitive market, way outside of reach for >90% of today's professionals, thats why people are worried.
People that don't know that "computer" used to be a profession back in the day.
Maybe it's going the other direction. It lets Microsoft essentially launder open source code. They can train an AI on open source code that they can't legally use because of the license, then let the AI generate code that they, Microsoft, use in their commercial software.
Maybe someone should vibe code the entire MS Office Suite and see how much they like that. Maybe add AD while they are at it. I'm for it if that frees European companies from the MS lock in.
Good idea. My country spends over billion dollars on Microsoft licenses annually, which is more than 200 euros per capita. I think billion dollars a year spent on dev salaries and Claude Code subscription to build MS office replacement would pay itself back quickly enough.
Even better - train a model on MS source code leaks and use it to work on Wine fork or as you said - vibe coded MS office. This would be hilarious.
There is Libre Office https://www.libreoffice.org/
Actually the opposite is happening, more and more vibe coded source code is making it to github.
You could argue about quality but not "No one will ever want to open source their code ever again".
They always did what they wanted with open source code, not sure why people think this is different