OP not a publicly established personality, and the website is new too. So, they have a lot to prove, credibility chief among them, especially since the topic is mental health. Users discovering that some of the things are AI generated leads to a thought that maybe more, subtler things are AI generated as well - like the texts on the website. AI output is currently distrusted, and the companies' treatment of the authors of their training material is frequently questioned. So, this way, discovering AI content erodes trust, of which the author didn't have any to begin with - because they are just starting out!
So that is why AI is a deal. Looks-wise they are good, and they fit their context very well. But they do communicate much more than that, and that makes them iffy for many.
I think I big part of copyright law is whether the thing created from copyrighted material is a competitor with the original work, in addition to whether it's transformative.
LLMs are OBVIOUSLY not a replacement for the books and works that they're trained on, just like Google books isn't.
Authors are getting busted[1] on a regular basis publishing LLM-generated or augmented novels that plausibly compete commercially with human-written books. If the LLM was trained on any books in the same genre, it seems like a clear violation.
I think what the GP meant was that it doesn't compete with that specific work it was trained on.
That is, if you want to read Harry Potter, you'd rather buy it (or get it from Anne) than try to wrangle it out of an LLM. Therefore, it doesn't compete with the original work. IANAL, though.
Keep in mind the 4 factors of Fair Use (US-specific):
1. the purpose and character of the use, including whether such use is of a commercial nature or is for nonprofit educational purposes;
2. the nature of the copyrighted work;
3. the amount and substantiality of the portion used in relation to the copyrighted work as a whole; and
4. the effect of the use upon the potential market for or value of the copyrighted work.
For 1, maybe OpenAI could've been safe if they'd actually stayed "open", but nowadays every AI company clearly fails, as do many (but not all) of the LLM users. Contrast this with most traditional fanfiction and personal projects where there were scary letters and occasional bullying, but few actual law-based problems.
"Transformative" is also part of 1 and is often cited as letting LLMs get away with everything, but everybody argues that and doesn't always win. Also, it's quite linked with 4.
2 mostly isn't a problem but gets into nasty details.
For 3 (emphasis on "used"), this link once again proves that the point does fail.
For 4 we are indisputably seeing mass disruption in several fields, so that point fails.
To be clear - there are ways to use LLMs that balance much closer to the side of fair use, but that isn't how LLMs are advertised.
Why not? Imagine a story teller app that is instructed in narrating a story the follows Harry Potter 1 - I would expect that there are already a ton of these apps out there.
That's not the same as the LLM itself though. That's an LLM plus specific instructions which would likely need to include a fair number of details from the books
IANAL, but I don't think whether LLMs are successful as a replacement is very relevant.
LLMs are advertised for, and attempt to, replace the works that they're trained on. For HN users, this is most often for code-generation, but people in other fields use it for similar replacement in their own field.
No one advertises LLMs as a replacement for literature or news and those are some of the highest profile legal cases I'm aware of. Code generation is another relatively high profile case, but in those cases I've not really seen any copyrightable code reproduced by an LLM without the prompter specifically trying to make it happen.
I haven't tried running SD 3.5 specifically, but it's built on hugging face libraries which I personally always find to be a mess of dependencies that make it really hard to setup without the exact configuration the original developers used (which is often not provided in enough detail to actually work). This makes it pretty hard to run certain models especially if it's a few months/years after the original release.
For example this appears to be the requirements for the stability AI reference implementation for SD3.5 and there are no versions specified and it includes "transformers" which is just an enormous library.
Ah, tyvm, that maps well onto my knowledge set, I have a ONNX inference wrapper written in Dart. However, I have never been able to leverage transformers.js ONNX demo code, i.e. have a reference to port to Dart.
IIRC it is written in an abstraction layer that supports a transformers-like API surface. This also makes it opaque to figure out what you're actually passing to the model, adding a Python dep mess on top of that...woo boy.
Government employees profit from work they do too. There is no enterprise that people should engage in that does not provide profit to them personally.
That’s an odd take, I never said people should not be paid a living wage for their contributions to the country. However, when was the last time the government authorized a multibillion dollar pay package for a single worker? The government doesn’t have stock buy backs, they don’t have dividends to pay out. That is where private business greed comes in and has no place in public government work. That kind bloat is what I am highlighting in private business.
Afaik Tesla is not a government contractor, so your compensation comment is moot.
Where we see waste in government is in big, entrenched, dinosaurs where competition is very difficult and/or limited. What we need is a little more free market capitalism in those areas. Greed is not the driver of waste imo, incompetence is a far bigger contributor.
If you'd like to make tangible arguments instead of vaguely ranting against capitalism then I'm all ears.
Edit: also the multi-billion dollar compensation package was in stock, so it didn't actually waste any money
Edit 2: by the way you're still dodging my original legitimate question about what the government has ever actually built without private contractors. There are real answers to that question and if you are such a big proponent of the government building things themselves maybe you should know something about them.
It may be important to note that SpaceX and Tesla finances are likely intertwined. This is a common dynamic among Musk’s companies. So while Tesla is not a contractor, its finances do affect government contractor risk. Tesla stock has been the main collateral for Musk at the same time he engages in start-ups and I don’t think that is coincidence.
Two edits and you hand waved away the point about CEO compensation? oof. Governments and private businesses have the same problems since they are both made up of imperfect humans. The government has the great oversight mechanism that private businesses don’t which is called voting, which is why I think it takes the cake here.
I never said that the government should never use private contractors (always need janitorial services), but when tax dollars are being used to fund technology and research, the government should never cede ownership of those outputs to private business.
I think the OP was alluding to unreliability in the CEOs mental state, but I’ll add another aspect: they are sometimes skirting well-established norms. An example is not performing material quality checks on critical parts. This is standard practice in the domain, yet they choose not to and it resulted in a loss of a rocket and its payload. They later added those quality checks to their process. SpaceX is good, but there’s no need for repeating well worn industry mistakes just because you fancy yourself as “different”
It would potentially mitigate the risk identified by the GGP regarding “unreliable” contractors who force risks on you that you may not want. Same reason I often choose to do house maintenance myself. Not to say it’s also not without costs/risks, it just comes down to which balance you prefer.
NASA still makes these competitive contracts though and picks among several contractors. Afterwards NASA is still involved in design through reviews and other lines of communication.
Using your analogy, if I do hire a contractor I'll talk with them a lot about what they're going to do and make sure it's generally in line with what I want, but they're generating most of the ideas and just incorporating what I say.
Eh, not so much. They have reviews, but it is a much more hands off approach. *
There were instances where NASA engineers brought up issues with designs and were told it wasn’t their role to drive the design. The concept of CCP was they were buying a ride, not a rocket. Just like you don’t tell Airbus what engine they should use when you buy a plane ticket.
* IMO the goal of CCP was to find a mechanism to informally circumvent many requirements. NASA could always waive requirements but I don’t think many people were willing to sign on the dotted line even if they disagreed with the requirements. CCP unburdens them from the same requirements while also allowing them to avoid full responsibility for the decision. (More charitably, it also allowed them to avoid some political costs, like having to spread projects across multiple political areas to avoid funding cuts.)
Right, reviews, where important design concerns can be raised. IDK what specific design concerns you're referring to, but just because an issue is raised doesn't mean it's a real issue.
Again, you don't want two different organizations trying to design one thing.
You missed the part where NASA engineers were told to pipe down about concerns because it wasn’t their place to drive the design. There were numerous, the ones I’m familiar with involved touch screens in cockpits and the amount of reliability needed in safety critical hardware.
NASA still sets requirements and invites several companies to compete for contracts with different solutions. See the lunar lander contracts from a few years ago for example.
You are ascribing beliefs to be based on others in this thread I think.
What I think is that if a company is going to build and provide the solution then they should own the design. NASA should of course get to be involved in reviews and discussions, which they absolutely fucking are, but I do not think that it makes sense for one organization to design something and the other one to build it as if there's like a hard line between these two activities.
I'm not convinced that is how it worked in the days of Apollo either as you've just asserted that without citation.
How are you measuring effectiveness? It seems like you might have a pretty shallow perspective on what NASA does and what their goals are. For example, do you know how many mission directorates they have and how they differ?
Not the person you're responding to, but JWT, SLS, and several other projects have suffered extreme bloat in both cost and timeline. Mega projects like that are some of the most public -facing things NASA does, so they unfortunately tend to drive public perception.
I will never argue that NASA doesn't accomplish amazing things, but large parts of the organization are ineffective. IDK if I'd go so far as to say the entire organization is ineffective, but large parts are.
I also don't think we should cut NASA's budget at all. We should cut the bloat and redirect it to more projects.
It does read to me that many people view this with the same lens we apply to private companies.
As an example, some of the bloat is intentional because it buys down political risk. It would be more efficient to have a program like SLS done within a single NASA center. But that also makes it easier to cut funding because there are a limited number of constituents affected by those cuts. This is exacerbated by long timeline projects that don’t align with shorter political timelines. By spreading the project to many centers, it adds inefficiency but also ensures the survivability of the project. To an extent, there has a a good chance there never would be audacious projects if they were run with maximum financial efficiency. So you’re stuck with the choice of an efficient project that never gets completed or inefficient one that does. As a taxpayer, I’m not thrilled with that dynamic but I understand why it exists
JPL is also not really “NASA” in the same sense. There are only a handful of civil servants and the CalTech as a contractor. It’s similar to the “quasi-government” operations of national lab.
I generally agree with this sentiment, but they did reach orbit with their sole launch of New Glenn! An admirable thing, even if it took like a quarter of a century.....
I was in a full arm cast (including fingers) for 9 months in college while taking many CS classes and I just used my one hand to type on the full qwerty keyboard. It definitely slowed me down, but I got up to ~60 wpm with the one hand. I think it's easier to just stick with the layout you know vs trying to learn to type with one hand in a new layout.
The nice part is that I can still type pretty quickly one-handed (maybe 50 wpm? Haven't measured in a while) and it's convenient sometimes.
IDK, I feel like good kids movies tend to stay good into adulthood. Presumably you wouldn't feel like it's torture to sit through something like the Lion King (the original of course) as an adult? Like I'm not saying they're amazing movies, but I feel like good kids movies aren't painful for adults to watch. You can still have good writing, it just has to be something a kid can follow.
reply