You are talking past this person’s point for some reason. They’re just saying you should use “Pro Israel group, StopAntisemitism,…” and they’re not wrong? Why in the world are you so dead set on being clear in the title who exactly this scumbag organization is??

Zionism? The ideology that fundamentally is based on the belief that Jews cannot safely exist anywhere except in a global ghetto built on ethnonationalism and apartheid? Touching tips with antisemites? Naw, not possible!

I would have to look into the actual patent and file wrapper, but presumably it didn’t cover just any rounding of rectangular corners, but as you said, a defined range.

Where bad patents get through prosecution, they are problematic, be they design patents or utility patents, but design patents in general are not even a blip on the radar of what needs to be fixed in our IP system imo. As a general rule of thumb, they are in fact fairly narrow. Meanwhile, pharma patents very much need focused and thoughtful revisions, and IP around software needs to be reworked from the ground up basically, creating special rules for patents and dropping the legislatively declared copyright framework entirely. The problem is that reporting on IP is fucking awful so people say things like “ohmyglob this design patents doesn’t even have real claims” even though that’s literally how they are structured and enforcing the right requires a pretty intensive investigation of the drawings and line patterns therein.

But, sure, I’ll give that maybe Apple’s design patent in this case was overly broad. I’m not particularly interested in defending Apple’s IP.

Design patents effectively work like brand protection. They literally only protect new aesthetics and ornamentation. The reality is that the iPhone did start the trend of rounded corner rectangular touchscreen phones. When it first came out, it was a fairly novel form factor for a phone. It didn’t prevent other form factors from being released. Like, the fact that it is now so ubiquitous that we take for granted smartphones look this way is a testament to its success. And, actually, plenty of phones did right angle screen corners. Design Patents are extraordinarily narrow things and, among the many issues with the current USPTO and the US IP system in general, it is probably the absolutely least problematic piece.

What is described in the second point is literally how Design Patent claims work. They don’t work the same way as utility patents. Anyway, yea, people not knowing how patents actually work aside, leadership at the USPTO is currently fucked.

Israel, its many crimes aside, has a deeply integrated economy into a lot of our tech. Basically any advanced technology companies either or both have a research site in Israel or acquired an Israeli company. The BDS list is an ok starting point, but recognize that it’s kind of arbitrary and has a lot of gaps as well companies whose presence on it is a little questionable.

I thought that was weird, too, but that’s not what they’re arguing actually. Their argument is that these were pirated for personal use by various people on the company network over a course of years and that the IP address is not sufficient to identify the appropriate defendant (not Meta). Accordingly, they argue the case should be dropped because tje pleading does not, and cannot from what has been provided, identify a correct defendant. At first blush, it isn’t an unreasonable argument. It would be like suing a university for detecting porn torrents on its network over a number of years (and alleging that the relatively small number of torrents were for AI research/training data).

This is a distressingly unusually solid analysis for lemmy. I agree with one exception–writing to memory absolutely counts as a distribution. Accordingly, if a generative model output an infringing work, it for sure could create liability for infringement. I think this will ultimately work similarly to music copyright where conscious/explicitly intentional copying is not itself the threshold test, but rather degree of similarity. And if you have prompts that specifically target towards infringement, you’re going to get some sort of contributory infringement structure. I think there is also potentially useful case law to look at in terms of infringement arising out of work-for-hire situations, where the contractor may not have infringed intentionally but the supervisor knew and intended their instructions to produce an effectively infringing work. That is, if there is any case law on this pretty narrow fact pattern.

Effectively, this has been an ongoing initiative across DoTs for a long while now. The issue is that it’s a hodgepodge approach baked piecemeal into various grants and other programs. But, yeah, digital, vendor agnostic, secure transit infrastructure is always on a lot of DOT folks’ minds.

Complains about an issued patent but nowhere actually includes the claims of the issue patent in the text of the article. Jfc, what garbage. If you look up the issued claims, they are pretty narrow and easy to design around. This article is bait.

No, that’s not even remotely true. The person doesn’t know what they’re talking about whatsoever. Over 75% of patent prosecution (bringing am application to issued claims) revolves around arguing whether a piece of prior art preempts the instant application. Just a buck wild utter opposite understanding of how patent examination works.

Most of these figures are guesses along a spectrum of “educated” since many models, like ChatGPT, are effectively opaque to everyone and we have no idea what the current iteration architecture actually looks like. But MIT did do a very solid study not too long ago that looked at the energy cost for various queries for various architectures. Text queries for very large GPT models actually had a higher energy cost than image gen using a normal number of iterations for Stable Diffusion models actually, which is pretty crazy. Anyhow, you’re looking at per-query energy usage of like 15 seconds microwaving at full power to riding a bike a few blocks. When tallied over the immense number of queries being serviced, it does add up.

That all said, I think energy consumption is a silly thing to attack AI over. Modernize, modularize, and decentralize the grids and convert to non-GHG sources and it doesn’t matter–there are other concerns with AI that are far more pressing (like deskilling effects and inability to control mis- and disinformation).

That is not what judges have said. They’ve said that merely training on text is not a copyright infringement. However, companies that downloaded enormous amounts of pirated texts (i.e., stuff they did not have license to download in the first place) still infringed copyright just like anybody else. Effectively the courts have been holding that if you study material you have license to access, you aren’t infringing, but if you pirate that material, even if it is merely to study it, it’s still infringing. For better or worse this is basically basically how it’s always been.

I have no idea what Trump is proposing. Like most republicans, but especially him, he is incapable of even approaching understanding of nuanced and technical areas of law and/or technology.

I agree. I’m generally pretty indifferent to this new generation of consumer models–the worst thing about it is the incredible amount of idiots flooding social media witch hunting it or evangelizing it without any understanding of either the tech or the law they’re talking about–but the people who use it so frequently for so many fundamental things that it’s observably diminishing their basic competencies and health is really unsettling.

Diffusion models iteratively convert noise across a space into forms and that’s what they are trained to do. In contrast to, say, a GPT that basically performs a recursive token prediction in sequence. They’re just totally different models, both in structure and mode of operation. Diffusion models are actually pretty incredible imo and I think we’re just beginning to scratch the surface of their power. A very fundamental part of most modes of cognition is converting the noise of unstructured multimodal signal data into something with form and intention, so being able to do this with a model, even if only in very very narrow domains right now, is a pretty massive leap forward.

A quick search turns up that alpha fold 3, what they are using for this, is a diffusion architecture, not a transformer. It works more the image generators than the GPT text generators. It isn’t really the same as “the LLMs”.

It won’t tell us what to do, it’ll do the very complex thing we ask it to. The biggest issues facing our species and planet atm all boil down to highly complex logistics. We produce enough food to make everyone in the world fat. There is sufficient shelter and housing to make everyone safe and secure from the elements. We know how to generate electricity and even distribute it securely without destroying the global climate systems. What we seem unable to do is allocate, transport, and prioritize resources to effectively execute on these things. Because they present very challenging logistical problems. The various disciplines underpinning AI dev, however, from ML to network sciences to resource allocation algorithms making your computer work, all are very well suited to solving logistics problems/building systems that do so. I really don’t see a sustainable future where “AI” is not fundamental to the logistics operations supporting it.

I imagine not, though I haven’t looked into it.

There are many open sourced locally executable free generative models available.

You are agreeing with the post you responded to. This ruling is only about training a model on legally obtained training data. It does not say it is ok to pirate works–if you pirate a work, no matter what you do with the infringing copy you’ve made, you’ve committed copyright infringement. It does not talk about model outputs, which is a very nuanced issue and likely to fall along similar analyses as music copyright imo. It only talks about whether training a model is intrinsically an infringement of copyright. And it isn’t because anything else is insane and be functionally impossible to differentiate from learning a writing technique by reading a book you bought from an author. Even a model that has overfit training data, it is in no way recognizable to any particular training datum. It’s hyperdimensioned matrix of numbers defining relationships between features and relationships between relationships.