> It's also a mistake that implies a complete lack of familiarity with scientific publishing, unfortunately, which makes it a bit difficult to take your judgements regarding plausibility very seriously.

It's still peer reviewed, and as the sibling comment said, more applicable to this type of research. Also you now went from raising understandable objections to refusing the argument because it comes from a specific journal, which doesn't sound very scientific to me

You're right it isn't fair to reject someone's scientific argument just because they seem unfamiliar with how professional science works.

We shouldn't have believed the study more if it actually had been in Nature.

I don't think that's what I was saying, though.

The issue in this thread was about taking a step back and looking at the overall plausibility of the conclusions, taking together multiple studies.

I agree with the GP that the argument doesn't really pass the smell test.

That's still the main issue, and it is something that people who don't understand scientific publishing struggle understanding/doing, because they lack the intuition for how certain results came about.

They are just predicting the next token. In human text it's more common to talk to other people than a computer, so they end up talking to the computers like they were people.

Because java is garbage-collected and doesn't have any of the problems of C++ exceptions, so checked exceptions just become a nuisance of having to try/catch everything.

You don't have to try/catch everything.

You do have to catch checked exceptions and that is the issue under debate.

No, you can just declare that the exception in the throws clause.

Earth (and the solar system) is 4.3B years old, a bit more than a third of that, so it's not really that much time in comparison no?

> There's no point maintaining the illusion that we're soliciting feedback or discussion on the issues tracker when we are not.

You could have just said this (maybe you did when linking the code of conduct) instead of writing a paragraph of confrontational arguments and it would have looked way better imho.

> You may think it hyperbolic but drive-by negativity by non-code-contributor users is the biggest existential risk to projects like Homebrew.

If this was true every oss project would either be dead or be entirely comprised of dicks, neither of which are the case.

> You could have just said this

Yup, you're right, I should have. We will adjust the CONTRIBUTING.md accordingly.

> If this was true every oss project would either be dead or be entirely comprised of dicks, neither of which are the case.

I didn't say every OSS project, I said projects like Homebrew. I know that Homebrew would be dead without many of my personal interventions. You can believe me or not but, unless you're a Homebrew maintainer, it's unlikely your opinion about what happens behind the scenes is informed.

A human leg isn't designed to be repaired from the outside, it doesn't need replacement parts, it doesn't need external maintenance. It's designed around the maintenance and repair procedures of the human body which are able to fix it in situ and produce new parts on the fly. That allows it to have more freedom of design than a mechanical part that needs to be made from standardized parts in order to be serviced and manufactured.

The limitation is not cognitive, it's a matter of efficiency. We could, if we wanted to, make a lamp with all custom parts with the maximum efficiency possible in the design, but those efficiency gains would be at the cost of extremely difficult manufacturing. It's a tradeoff that the human body can make because it needs to produce only one (or two) legs, while us humans mass manufacture the stuff we make.

On one hand it's true, on the other people want stability in their jobs. You need a compromise between flexibility for the employer and stability for the employee, and one such compromise (in this area) is having a consulting company whose contract with the client can be easily terminated, while keeping the actual people doing the work with a job and stable income.

LLMs are not humans, nowhere near.

We don't know what LLMs encode because we don't know what the model weights represent.

On the second point it depends how the models were made to reporduce text verbatim. If i copy-paste someone's article in MS word i technically made word reproduce the text verbatim., obviously that's not Word's fault. If i asked an LLM explicitly to list the entire Bee Movie script it would probably do it, which means it was trained on it, but that's through a direct and clear request to copy the original verbatim.

> If i copy-paste someone's article in MS word i technically made word reproduce the text verbatim., obviously that's not Word's fault. If i asked an LLM explicitly to list the entire Bee Movie script it would probably do it, which means it was trained on it, but that's through a direct and clear request to copy the original verbatim.

But that clearly means that the LLM already has the Bee Movie script inside it (somehow), which would be a copyright violation. If MS word came with an "open movie script" button that let you pick a movie and get the script for it, that would clearly be a copyright violation. Of course if the user inputs something then that's different - that's not the software shipping whatever it is.

That's not a fair comparison. The user in the word example already had access to the infringing content to copy it, and then paste it into word.

But it has to have that copy, verbatim, to produce it, as you acknowledge.

If dropbox was hosting and serving IP from paramount, paramount would be able to submit a DCMA request to get that data removed.

Not only can you not submit a DMCA request to chatGPT, they can't actually obey one.

> If i asked an LLM explicitly to list the entire Bee Movie script it would probably do it, which means it was trained on it, but that's through a direct and clear request to copy the original verbatim.

Huh? The "request" part doesn't matter. What you describe is exactly like if someone ships me a hard drive with a file containing "the entire Bee Movie script" that they were not authorized to copy: it's copyright infringement before and after I request the disk to read out the blocks with the file.

It wouldn't amplify them, there's no energy gain. A gravitational lenses just bends the trajectory of the waves. If you do it right it may be possible to use that to focus the signal on a directional trajectory(as in, the same energy is redirected in a single direction instead of being spread out), if I'm not mistaken.

True, it doesn't amplify, it just focuses. So that's different to the three body problem, in which it wasn't directional as far as I remember.