Open link in next tab
Microsoft CEO of AI: Online content is 'freeware' for models
https://www.theregister.com/2024/06/28/microsoft_ceo_ai/
Unless you've got a lawyer, that is
Cool so we can just make up our own rules now. Well, all Microsoft products are freeware now because the same reason this guy
•
Windows XP code was leaked 2 years ago, so it's freeware according to this idi... stable genius .
Ok.. so from now on .. when I see a "repackaged" Microsoft product that for some reason.. which I don't care to know... doesn't ask for a payment.. I can use it without restrictions ?!! that's really nice of you Microsoft ... thank you.
Which is why I boycott as hard as I can every service this evil corporation provides (migrate your MS GitHub project away now so I can delete this account too)
Microsoft is in a death spiral.
Even my coworkers who are complete idiots with technology, who actively sabotage themselves every time they touch any piece of hardware and software, have soured entirely on nearly every Microsoft product across the board.
Its funny how quickly people change their minds when they dont understand the technology on a deeper level. Its just: "this is frustrating now I hate it" and no further thought.
Fair, then everything I can find on the Internet must be freeware too. Set the sails, matey!
No officer, this is not a pirated movie. It's generated by an AI model I created and trained with data from the internet and the fact that it's 99% identical to an existing movie is irrelevant.
my AI is so good, it generated one that’s 100% identical
plus my AI uses less than 99% of the electricity of Microsoft’s
•
Also, this ground breaking AI model I made to do this was umm accidentally erased and I also forgot how to do make it.
Jury: “seems reasonable”
As one person on Mastodon said, "AI is a toxic industry created by toxic people with toxic ideals".
I wouldn't go that far. As it turns out AI is a buzz word and buzz words have little meaning
If an LLM can save me 30 minutes writing nice emails and responses and help me brainstorm, debug, or elucidate my thoughts then it is very useful.
You really put 30 minutes of your own time above all of downsides this has for the rest of us who don't have a use for it (most of the world)?
All of the resources and energy spent to get you this product you like. You can't discount what it took to create something just because the final product is small and efficient. Take a look at the manufacturing footprint of nearly all complex hardware.
I'm not saying you created the AI but you are one of its supporters, without which there would be no AI.
If this was all just pitched as developing a new plain English coding language, I think the hype following it would be far more appropriate, but then the funding wouldn't follow to support the massive development costs of AI.
Its become a circle of hype chasing money chasing hype.
Its not you that is the problem so to speak though, its the collective "you's" who think the same way.
I'm not discounting it. Improving productivity for office workers by 1% across the world is a massive amount
The power used to train the AI is alot, but after that using the AI uses a lot less electricity, if an AI spikes my gpu by 10 seconds to type something that would have taken me 30 minutes, I've saved on electricity:
I don't care. They are really helpful for a many different tasks. It doesn't pull that much power to run locally on my machine.
"See I like AI because I'm selfish. Also those bad things are in the past, I'm using an ethical AI system now! But also, who gives a fuck because I only care about myself!"
Yeah you get it guy! Maybe you can be Trumps secretary of technology!
Mister/miss, LLMs that can run locally are fine. It's the infrastructure and the large scale of commercial cloud LLMs that create some issues. You have to read some researches on this topic.
It's freeware until someone else take m$ content without paying them, then it's copyright infringement.
•
From the article:
Also, in 2022, several unidentified developers sued OpenAI and GitHub based on claims that the organizations used publicly posted programming code to train generative models in violation of software licensing terms
They can argue about it not being a copy all they want. If there is a single GPL licenced line of code scraped then anything they produce is a derivative work & must be licenced GPL.
nice.
I’ll play the uniformed devils advocate here:
- Is the GPL license enforceable?
- And if so, I assume “derivative” will still subjective to some degree. Where do we draw the line between derivative and non-derivative?
I’m torn about my personal opinion about copyrights and software licensing in general. I think the main problem is the huge power imbalance between people and corporations, not so much the fact a company analyzed a bunch of available data to solve programming problems.
They don’t copy the data and sell it verbatim to others which would be a legal issue and in my mind also a moral issue, as they don’t add any additional value.
The only way I can see them weaseling out of this is by keeping the program running the model made in-house and proprietary while releasing the model in a format unusable without the base (proprietary) program. But maybe the GPL forbids such obfuscstion efforts (I don't know, I haven't studied it in detail)
GPL v2 don't, which lead to tivoization. But Linus himself didn't agree with that standing.
He seems to be confusing "freeware", which is basically a license for copyrighted work, with "public domain", which is the absence of a copyright.
I'm fine with that, but let's put some rules against this.
- Any AI models should be able to determine the source of their data to a defined level of accuracy.
- There should be a well-defined way to block data from being used by AI. If one of these ways (e.g. robots.txt) has been breached, the model has to be rebuilt without the data, and reparations made to the content owners.
What you're asking for is literally impossible.
A neural network is basically nothing more than a set of weights. If one word makes a weight go up by 0.0001 and then another word makes it go down by 0.0001, and you do that billions of times for billions of weights, how do you determine what in the data created those weights? Every single thing that's in the training data had some kind of effect on everything else.
It's like combining billions of buckets of water together in a pool and then taking out 1 cup from that and trying to figure out which buckets contributed to that cup. It doesn't make any sense.
Respectfully, I worked for Alexa AI on compositional ML, and we were largely able to do exactly this with customer utterances, so to say it is impossible is simply not true. Many companies have to have some degree of ability to remove troublesome data, and while tracing data inside a model is rather difficult (historically it would be done during the building of datasets or measured at evaluation time) it's definitely something that most big tech companies will do.
Sorry, I misinterpreted what you meant. You said "any AI models" so I thought you were talking about the model itself should somehow know where the data came from. Obviously the companies training the models can catalog their data sources.
But besides that, if you work on AI you should know better than anyone that removing training data is counter to the goal of fixing overfitting. You need more data to make the model more generalized. All you'd be doing is making it more likely to reproduce existing material because it has less to work off of. That's worse for everyone.
It’s not impossible lol. All a company would need to do is keep track of where they were getting content. If I use a script to download as much of the internet as possible and end up with a bunch of copyrighted content I could still get in trouble, hell there was even a guy arrested for downloading jstor without authorization.. Stop letting these guys get away with crimes just because you like the idea of the end product
Man it's crazy how these fuckers basically get to ignore copyright law whenever it's inconvenient to them but if you have one too many Windows machines provisioned they'll send the Spanish Inquisition after you.
Sure thing...now GPL/Creative Commons all your code involved in any way for your models, documentation, parameters, data sets, and allow full unlimited integration and modification by any parties to any portion of it.
The social contract? Tf. The social contract still required attribution in almost all cases for creative work unless explicitlf stated otherwise—especially in the case of comercial products like ChatGPT—so I don’t know where this joker is getting his ideas.
•
I went into a smidge more detail over on my Mastodon last night, but my response is summed up as “WTAF? No! Freeware is an explicit license, as anyone from the BBS days will recall.”