Ok, so music education specifically… Sadly I can see that.

How to just play instruments and general basic features of genres… I could see generative AI able to do things on that front. Most notably ability to analyze your performance and identify gaps. Basic rote stuff without need for creativity or soul. The same stuff that countless musicians have already done. I don’t think there is a need to genAI up pre-recorded, non interactive lessons, but I could buy interactive AI experience as being feasible and competitive.

At least in theory, though I see that video of someone asking for feedback on their music and it’s just literal random farting and the AI was still quite complimentary. But Iif someone told me someone curated a more specialized GenAI experienced that worked better in the context… I would at least believe in the feasibility.

I have, it’s not an improvement over musicians. I see no way it could help a musician. I can see how it could help a non musician generate something that can pass for music, but I don’t see a musician getting use out of AI.

For writing music, if you could write music before, AI can’t help you. If you know how but just have a creative block, random AI crap won’t be any better in a musicians hands than a random person.

When it comes to performing, largely virtual instruments without AI could already handle those credibly. Stuff like violin might have taken longer, but it got there. The one exception is vocals, where vocaloid was about as far as non AI could get and while interesting, no one was going to mistake vocaloid for real singer. AI vocals can pass for a phoned in vocal performance. But I’d just as soon have a real singer, there’s no shortage thanks to decades of music and the Internet.

Bad news, if AI causes musicians to be denied a livelihood, it doesn’t matter if the musician adopts AI or not.

The mechanism for this would be that random slop generator is valued as highly as human written/performed music, and no amount of using AI will keep the musician “ahead” of the curve.

If the good musician uses AI with the same care taken for their own music, it probably won’t speed up the creative process.

Depressingly, I’ve seen enough slop released as “music” that I can believe the music economy has been terribly diluted by AI, but adopting AI won’t fix the problem for the musician that already has the skills to not need AI.

There is some ai company that routinely spams my work email. Various typical vague, “AI will fix everything with your work” promises.

Then some report came out that 95% of business AI adoption failed. I figured either it would be ignored, or some response explaining how the study was mistaken or didn’t say what folks said it did.

Instead, that company said “Study shows that 95% of business use of AI fails, you don’t want to be in that 95%, so use our services to make sure you are in that 5%”. Had to have it to them, they really found a way to steer into the skid.

I agree it is a crap comparison, just warning that even as the slop straight from the LLM is not copyrightable, blend in a little human touch up and it is “blessed” into being copyrighted. So it’s of little consequence unless the operator just passes it through untouched or makes it somehow easy to identify and remove the human tweaks.

In terms of slopping up a rip off under the assumption that it would obviously be the same as the original slop operators starting point, AI companies in their defense have already stated that operators can violate copyright implicitly as they legally declared they aren’t on the hook for any resultant infringement…

Quick, divert ICE to dealing with that invasion instead.

So funny you should mention that, I worked at a company that dealt with Linux and third-party proprietary code. They kept the software developers highly segregated because they feared accidental copyright infringement. They thought at the time that even a human learning and accidentally reproducing something they remembered from working on proprietary code too risky.

The practical difference is that the third-parties were likely to sue, but the corpus of trained data is pretty much open source projects that may have a right to sue but in practice everyone knows they aren’t going to be able to chase down violations due to lack of resource.

Well, it has been litigated and it continues to be litigated so the legal community doesn’t consider it completely settled, but so far court behavior seems inclined to agree with the AI companies.

That said, the AI company defense has broadly been “training is fair use, and even if an operator manages to get the model to produce an infringing work, the user is liable, not us”. So that “user is on the hook part” should be worrying people, but no one has yet tried to test this sort of oblivious infringement. I would expect it should at least be plausible that the operator was oblivious to the copyright violation they induced the model to commit, but still demand the offending material be removed.

Currently the industry is betting that it is “too big to fail” at this point and no one would dare pull that trigger on an unsuspecting violator. Further, copyright doesn’t cover “trivial” lifts from a source and the likelihood favors an accidental copy being a trivial part of a whole.

This is one scenario where the GenAI has been pretty useful, to a point.

It can help someone review code. It has a lot of false positives, but it does occasionally catch something, and I can forgive the false positives, even if they are super annoying and time consuming, so long as it is catching something on occasion.

In open source, a problem is so many people in your community now want to ‘help’ by using their AI subscriptions to review the existing code and then spam the project with their ‘findings’ because they don’t understand well enough to double check the AI. This greatly amplifies the false positive problem with mind numbing volume of dupes.

This has also come up for generative AI with consistent result.

Problem in those cases is that they hinged upon the human passing through the non-human product without even claiming to have modified it. If human modification has been done, then that modification is subject to copyright. In the case of a composite of human and non-human origination, good luck extracting the non-human portion of the work to take advantage of that copyright opening.

The hiccup is that the statement is that pure AI output is not subject to copyright, but once a human does something ‘transformative’ to it, then that is. So if a human can work back to the untouched AI output, they can go to town but they can’t reproduce any modification that was done after the AI generation. Since in practice, the GenAI row output is not provided it’s impossible for the human to know what the human touches were to undo, so that scenario does not present a whole lot of realistic implications.

TLDR: In short, like a lot of AI use the biggest problem is more how other people are using it that I have little control over.

The thing is that CodeGen utility varies greatly on what is being advocated and what situation it is being applied to. CodeGen getting things going when the operator isn’t strongly opinionated about the details for fairly common patterns is fairly strongly in the wheelhouse. It is somewhat worse at amending a project in the same broad ways it was able to do when starting from scratch, even a project the model itself generated. So as you progress, it is more and more likely that a human will need to understand what is going on to be effective in modification. This presents a problem as people work themselves into a mess and can’t get out of it and this causes a fair number of projects to just get abandoned because they can’t go anywhere.

Even as it has demonstrable utility, the real world implications can be a mess. For example, the other week someone used agentic AI to open up 70 ‘security findings’ on a project I work on. To it’s credit, it found one actual issue, and while another issue was incorrect, looking into it I did find a separate issue it didn’t notice, so I got two fixes out of it. However I had to deal with 68 completely stupid things that weren’t anything. The operator at least up front sent an email that they didn’t understand any of it and how much was real, but wanted to share in case any of it is useful. But I guess that’s the price I pay for the two valid issues that might have otherwise not been caught. Then a few days after settling those, another user opened up over 60 and they were all dupes. Then the next week another user opened up a bunch that were all dupes too. Now it’s ridiculous. One could argue that I could fight fire with fire and put an LLM on triaging the issues and closing out the dupes, but that means random github users can now make me spend my money on LLM services just by opening issues. Further, after being on the receiving end of AI chatbot triage in customer support, I hate inflicting that on the humans opening issues.

Then of course there’s the code submissions and expectations around how I should handle them. There was a longstanding understanding that folks try their best, but implementing feature requests takes effort and folks are broadly understanding at delays or being a bit down the priority list. Now someone will Claude up a merge request instead of an issue and be impatient because “Claude already did the hard work, all you have to do is accept it”, and it’s a mess of code to review. It’s not like the code at the hands of the operator is exactly good, for example someone sent a merge request because some feature raised an exception for them and blocked it from working and Claude “fixed” it. Problem is Claude changed the code to catch the exception, do nothing, and just say that it worked. In their test scenario, the feature was trying to make things the way they already were, so a no-op was no problem, but when I made their test case actually start from another state, it failed but still said it worked.

Another project I used to be a part of recently got handed over from the maintainers that long maintained it and lost issue to a super Slop-happy AI user. They got it rigged so it’s all agentic and “addresses all issues and resolves all pull requests automatically from anyone”. The project had some poor issue hygeine, so this meant the agentic code was doing things like fixing an issue from 2017 where the user complained it didn’t work with Ubuntu 16.04. No idea if the fix actually worked, but it did make code changes that realistically no one would care about if they only did what they want. Except now they are getting bunch of new issues because they say “hey, the latest update broke a bunch of stuff I relied on”, and another said “hey, it was great to see my issue get fixed last week, but this week the issue is back?” and messages saying “the update pace is ridiculous, why are there 3-4 releases a day, the former cadence was once a quarter”. It’s become a broiling mess of chaos because some AI enthusiast decided to make a nearly abandoned project that people were using a playground for AI usage. Dude even said point blank it’s fine, just open an issue and AI will fix it for you and if it breaks, then someone else just open another issue and it’ll get fixed and he is confident that eventually the users will shape the project into some equilibrium that way, so he has to do very little except pay for the tokens.

Actually that is how copyright law could work.

On the ingest/training side I think the courts have been mostly consistent with your view and sided with the AI companies, though some lawyers are still opening new cases so at least some of them aren’t so sure.

The problem is the defense the AI companies have given is still a potential problem for software. They state that even as the ingest of works allow the model to generate material that would violate the copyright of trained data, that it’s the fault of the operator, not the AI company. So even if a user oblivious to the original material manages to get a model to create a knock-off of a copyrighted work, it is still the user’s fault, not the AI company.

This implication to my knowledge has not been tested in court. I haven’t seen cases where someone found and sued an AI user for plausibly oblivious copyright violation. Now I cannot imagine that the consequences can be very punitive if they can show the AI models will produce such a violation without the user indicating any awareness of the source material, but they may still demand it be ripped out, which depending on age of code could be a huge mess and a great loss of functionality of other things that depend on the offending code. Even if the particular case could be won on arguments about whether it meets the criteria of an infringement in that case, fighting the case is hard enough.

The question is whether the risk of accidental copyright violation and the risk of such a violation even being noticed is significant enough. Currently in the GenAI hype world where tons of money is seen to be had from fueling the phenomenon, that risk appears low. Question is if the hype dies down and the money suckers want more money, will they at least start making copyright claims against folks focusing on how the AI defense was to throw the users under the bus. Again, open source projects themselves are probably low on the list of targets, as financial penalties seem unlikely, but it can’t be ruled out.

I guess the point would be what does the local area get in exchange for losing their forest?

A lot of other developments at least provide the local populace some service or jobs. These are just “give me your land, your power, and water and we will use it to advance our business interests and do nothing for the locals”, oh and btw, give us a tax break for the privilege of having us build a datacenter in your area.

Now I’m being bombarded by ads saying ‘um, actually datacenters give you tons of permanent jobs and makes power cheaper’, so they are just flat out lying instead of actually doing anything to make it worth the local community. Well, except key people on local town councils who somehow end up with NDAs forbidding them from disclosing what conversations they are having with these developers…

Because solar on the roof won’t even be noticable against their power footprint. Given current practices, each square foot of a datacenter would need about 1,000 times more square footage worth of solar.

It’s just that somehow if they get to go to space, they decide they suddenly just need like two servers in a ‘datacenter’ instead of thousands. And they need to go to space because Elon says (checks notes) that it’s too hard to make more natural gas turbine parts…

Fun fact, over half of the S&P 500 have market caps over 10x revenue.

This has not ended well for companies in the past.

That is a careful what you wish for.

Not so smart as to challenge competitively, but not so dumb as to think they can go bombing whoever they feel like to get their way.

Maybe, but history is full of populations that were pretty docile… Until they weren’t.

I’ve personally seen how patience and complacency in a supremely calm individual can evaporate when things are tight. People who are always careful and taking care of the long term abandoning all of that if they feel desperate in a situation.

If the ballot box doesn’t work as promised by the existing social contract, that will be the last straw for some people. Particularly if asshats like Stephen Miller keep up the violent rhetoric about exactly how terrible they want to be to any dissenters. That’s a lot of fear, rage, and impotence for folks to get hit with, and it has a much higher risk of bubbling over than you would imagine.

I’ll confess to thinking I’m personally unlikely to risk things, but other people have surprised me

Why would that be making excuses for the USA? It’s not that one must be ok and one must be bad, both can be absolutely doing things wrong.

Don’t have to view China positively, just view USA as worse in the moment. Which is fair. USA is the one violently disrupting the global economy directly with military operation in Iran and supporting whole heartedly the Israeli offensive against essentially unarmed populations.

China is not significantly violently intervening beyond it’s borders in this time in history, preferring the much more broadly acceptable economic dominance. It is providing disheartening support to Russia in its offensive into Ukraine, but grading on a curve Ukraine at least is armed and supported by others while Gaza and Lebanon are just being victimized in a one-sided conflict. Both are unacceptable offensives, but one is more horrific than the other. It’s horrible to compare in any way that might be seen as diminishing the horrors of Ukraine, but it’s the unfortunate reality when we are deciding which ‘bad guys’ we think are the ‘worse guys’.

The biggest China issues that are ongoing and not past or potential future are largely within their borders. As far as potential future misbehavior that may be indicated, the hypothetical is outweighed by the USA’s actual behavior.