When someone tells you who they are, believe them. Microsoft’s AI chief Mustafa Suleyman:
With respect to content that is already on the open web, the social contract of that content since the ’90s has been that it is fair use. Anyone can copy it, recreate with it, reproduce with it. That has been freeware, if you like. That’s been the understanding.
↫ Mustafa Suleyman
This is absolute bullshit from the first word to the very last. None of this is true – not even in the slightest. Content on the web is not free for the taking by anyone, especially not to be chewed up and regurgitated verbatim by spicy autocomplete tools. There is no “social contract” to that effect. In fact, when I go to any of Microsoft’s website, documents, videos, or any other content they publish online, on the open web, and scroll to the very bottom of the page, it’s all got the little copyright symbol or similar messaging.
Once again, this underlines how entitled Silicon Valley techbros really are. If we violate even a gram of Microsoft’s copyrights, we’d have their lawyers on our ass in weeks – but when Microsoft itself needs to violate copyright and licensing on an automated, industrial scale, for massive profits, everything is suddenly peace, love, and fair use. Men in Silicon Valley just do not understand consent. At all. And they show this time and time again.
Meanwhile, the Internet Archive has to deal with crap like this:
The lawsuit is about the longstanding and widespread library practice of controlled digital lending, which is how we lend the books we own to our patrons. As a result of the publishers’ lawsuit, more than 500,000 books have been removed from our lending library.
↫ Chris Freeland at the Internet Archive Blogs
Controlled lending without a profit motive is deemed illegal, but violating copyright and licensing on an automated, industrial scale is fair use. Make it make sense.
Make it make sense.
This should go both ways. If he insists all content on the web is fair use, so is warez, very much including Microsoft products and associated cracks, keygens etc.
MS kinda turns a blind eye to that stuff as they know it keeps people trapped in using their products.
…and so should be various MS source code leaks. Fair for copyright infringement purposes. They are on the internet, after all…
People joke about this, but I actually like this stand.
Web used to be open, and anything you could access was fair use. Now we have dark patterns, “single page progressive web applications”, and uncrawlable websites like Amazon which hide their content behind dynamic frontends.
Even the web archive is incomplete, and a huge chunk of human knowledge erodes every day.
So, even if it is Microsoft, someone says “let’s keep a record of it all, and learn from it, at least in machine learning sense” it is better than having all those go to obsolescence.
What about those copies of the leaked Microsoft Windows source code that people pushed to GitHub?
If i put content publicly available in hope for intelligences to stumble upon it and maybe learn something, what difference does it makes if said intelligence is supported by a biological substrate or electrological substrate?
gagol2,
Yeah, learning should not be illegal. Copyright infringement means reproducing/redistributing works without permission, but both AI and humans can learn in generalizations that do not reproducing original content. I generally agree with the EFF’s take…
https://www.eff.org/deeplinks/2023/12/no-robotstxt-how-ask-chatgpt-and-google-bard-not-use-your-website-training
Websites can block robots, but in practice few owners do it.
This is osnews.com/robots.txt…
I had a feeling you would appreciate my renewed opinion, but i am much more curious regarding Thom’s standing on the idea. Of course accepting that intelligence is a topology and not anything sacred is difficult to accept, but also raise much more serious ethical issues as it would require humans to treat robots as citizens and not the disposable slaves we see them for.
https://advertongue.com/
Forbes put out a nice cover of the incident.
https://www.forbes.com/sites/barrycollins/2024/07/01/has-microsofts-ai-chief-just-made-windows-free/