archontes

A lot of creative thinking goes into the curation of the training set.

I've got disappointing news for you. The copyright protecting your open source software never applied to its function, only its creative expression, and only insofar as the expression was separable from its function.

Only if a human wrote it.

It's almost certainly true that the collection of "all the good music from the 70s" is copyrightable as a collection, but that doesn't make the weights the result of a creative process.

I wonder what basis they think they have to claim copyright on AI weights at all, as the weights are not the result of a creative process.

Nothing other than breaking my graphics drivers. Who needs a monitor?

Try doing it ten times, though.

If a universe were entropically uniform, wouldn't the arrow of time definitely be negative? Wouldn't any random evolution result in a lessening of entropy by definition?

Will try.

If I type 'goog', and the history result is google.com, then pressing tab should result in calling the search function in the tab with google as the search engine, and with the cursor in the box for the query.

It should not cycle through the drop down menu.

I want to begin typing a website's name, have firefox recognize it and indicate it to me, and press 'tab' to begin searching.

The lack of that single feature makes firefox unpleasant to use for me.

Hopefully you can take a multipart approach. Transit needs to improve by an equal measure. The North Shore still doesn't have a train. Most of the trains that do run take too long, and the highways just don't flow like modern highways. That reduces the utility of land. It doesn't matter if you build a house in Helensville if nobody's going to utilize it because it's a 50 minute drive to the city. I get that people do it, but people try not to. And it's not bad bad, but it's bad enough that it disincentivizes utilization.

Machiavellian... evil... semantics.

Interesting. It's likely the point of constraining the graduate course to grad students had nothing to do with their maturity. But you see what you think is a sign of immaturity, and turn the constraint into a maturity filter.

I'm quite grown, and I wonder about the ownership/control of the cluster and why he didn't simply lock the professor out entirely, contingent on the approval of his waiver.

If anything, doing something as small as lowering the priority of his jobs instead of brazenly stonewalling him might be the sign of immaturity.

What projects do you find useful?

Can you think of a more efficient method than encoding pi onto the spin state of every electron?

I think in the context of this conversation, that failure mode is a reason for preferring the primitive option.

Doesn't the tor browser disable javascript entirely?

Just a Russian individual. They were selling Russian foil magic cards, which are rare enough to warrant a premium.

To be fair, crypto was the easiest option when I ordered some stuff from Russia, so score one to it.

I am not confusing that issue. I am stating outright that it is not sufficiently clear that a human is responsible for the neural net weights. The gradient descent may be considered automatic.

Section 313.2 of the U.S Copyright compendium states, "The (copyright) office will not register works produced by a machine or mere mechanical process that operates randomly or automatically without any creative input or intervention from a human author".

Moreover, the Copyright Act's definition of a computer program is “a set of statements or instructions to be used directly or indirectly in a computer to bring about a certain result”. Neural net weights are clearly neither statements nor instructions, so it's unlikely the weights would gain copyright as a computer program.

It's indeterminate whether the SSO doctrine[1] is operable on neural net weights, as it can be difficult to determine in which order weights are used. If those weights were to gain copyright protection, they may not be enforceable on similar sets of weights.

Regarding whether resembling the original work is the legal standard for being considered an infringing work: the test applied by a court is called Substantial Similarity[2], and it's extremely unlikely that neural network weights would be substantially similar to training data when considering the idea-expression divide.

Copying into the model... are you suggesting that if I have an image that's represented by an array in memory, and I copy that array from one memory location to another in my pc, that act of copying is the infringing act?

Fair use may even be a bit non-sequitur in this case. If I look at copyrighted art, I don't have a fair use case, because I haven't incorporated the art into anything, though the act of viewing has changed my brain. Training a neural net on training data may not even be considered incorporation.

On your last broad point... okay? I suppose.

[1] https://en.wikipedia.org/wiki/Structure,_sequence_and_organi... [2] https://en.wikipedia.org/wiki/Substantial_similarity

Two issues with that. Training an AI has thus far been regarded in courts as non-creative, and thus the AI model/weights are not copyrightable, nor does the model sufficiently represent the original work, so it's not a derivative work.

It's literally the example, so...

https://en.wikipedia.org/wiki/Derivative_work#Examples_of_de...

Usually the copyright on a data set is held by the person who accumulated the data, but mere compilation is not enough to bestow copyright. The arrangement and selection itself needs to be sufficiently creative or original.

Putting all your photos in a folder does create a dataset, but doesn't meet the threshold of creative input for that dataset to be copyrightable as a compilation.

Now, you do hold the copyright on all of your photos. This means that you can restrict their transmission to others. You can even transmit them to someone as part of a contract that they not engage in a particular activity, like training an AI.

But if they proceed to train an AI using them, they've committed a breach of contract, not copyright. And I'll tell you, I'm not a lawyer myself, but instead a SME who works with a lawyer, and in order to receive relief for breach of contract, a judge is going to look to said contract in order to calculate damages. If you have a contract that says that something is a breach, but no language that assists a judge in calculating damages, you shouldn't assume you're going to get any relief in excess of, perhaps, an injunction.

And I'm afraid my knowledge is very US-specific.

It's worth noting that copyright doesn't protect code in the instance that the function cannot be divorced from the creative expression.

Comments and all would almost certainly be an infringement, as they are completely expressive and non-functional.

But you'll find that if you strip out those comments, it's possible that it would fail the test of the idea being separable from the expression [1].

[1] https://casetext.com/case/kern-river-gas-transmission-v-coas...

This is almost certainly false. Derivative work is a category of copyrightable material where an existing work was further creatively transformed in order to create said derivative work. For example painting a moustache on the Mona Lisa.

That creates a derivative work, which has a separate copyright on the creative addition (the moustache).

Training an AI has been found to be not creative work. The AI weights, as a result, are not a derivative work, and likely cannot be copyrighted.

Your GPL example is an unexpectedly poignant one, as well, though it will be to your chagrin.

If you publish your GPL code, and I learn from it and produce code that does the same thing, that is not restricted by copyright. Furthermore, if it's found that the expression of the code itself cannot be separated from the function, the expression is not copyrightable in the first place, and the GPL is redundant.

It's only in the instance that I reproduce your code exactly and the function is separable from the creative expression that your copyright is valid.

Regarding your JPEG compression, it's possible that a ML model will produce outputs that, of themselves, are copyright infringement. A recognizable picture of Ariel, for example. That doesn't mean that the training, the model, or anything else it produces is infringing.

It isn't copyright infringement to train an AI on information that you obtained legally.

Example: If I put my book on mybook.com, and you download it (legally), you can read it, learn from it, and produce works in a similar style, all without my consent, and copyright offers no tools to restrict that.

The only tool copyright offers to protect against that is distribution.

It has nothing to do with training. The only part that copyright would protect against is transmitting your photos' information to Adobe.

There's probably a clause in there about Adobe being able to do exactly that for reasons.

This continues to be a confusion in copyright. You can't put a photo online and tell me how I can consume it. You can tell me I can't distribute it to anyone else.

It's reasonable they could do a chargeback on the basis of violation of implied warranty of fitness.

Edit: it's probably a violation of implied warranty of merchantability.