Not sure that's the case, and above the bit you quoted it does also mention that there will be a special version for hobbyists that has different terms.
In short I think it's too early to say. It sounds like they haven't figured out what they are going to do yet.
Absolutely, and by "meh", it's often "fine" or "adequate". And there's nothing wrong with adequate it is by definition good enough. Would I like a machine that's 3x as fast for half the price and fully open source? Well of course. Do I need one? Well... no. I don't think I've ever been in a situation were one of those would be necessary and a Pi would be inadequate.
I've used NVidia Jetson Orin AGXs, but nothing is comparable to them anyway, even a 5x faster pi wouldn't match it.
It's a real pity that the ARM world is such a shitshow compared to the x86 world, which is funny to say. Broadly speaking the first and last time I have interaction with the vendor is when I buy it and the hardware will run for decades if it lasts that long with quite possibly excellent support across the board.
You still need a working Raspberry pi to do that
https://obsolescence.wixsite.c...
I have indeed now keyed a few basic programs in at boot time (well for the emulator not the pi obviously!). It's amazingly good fun and I then understood why octal rather than hex has such pride of place in unix and C.
My advice is to not be a cuck.
My advice is not to use weird dated MAGA terms.
The raspi people made it harder for you to tinker with your own hardware
Yeah generally I don't like this. On the other hand we now have massive widespread acceptance of online scamming where legally sound companies get to cream off the scam profits so basically encourage them.
This puts manufacturers between a rock and a hard place. If the channels are filled with broken re-badged raspi's that is in the world as it is right now Raspberry Pi's problem. And RasPi customer's problem. So what would you have them do? Just shrug and say "lol sux to be u"? In this case 99.999% of their customers will never change the RAM.
They've been customer-unfriendly from the beginning, when they were selling boards with dumb mistakes that ruined them for many purposes (remember raspi A USB fails? remember the MANY power supply failures?
And yet I've been happily using them for various tasks for years. Bugs or design flaws aren't customer hostility and it's dumb to conflate them.
Is that so? I can't find a source for Google saying that. Do you have one?
In any case, I think we will have to wait and see, because I can't imagine the EU being okay with them banning open source developers who refuse to register in that way.
I never thought I'd say it, but 2027 might actually be the year of the Linux desktop. There is gathering momentum for moving from Windows to Linux, with some major tech news sites doing extended tutorials and reports on it. Linux desktop user numbers are rising, fast enough that Microslop is trying to de-shittify Windows before too many people abandon it forever.
Most of all though, thanks to Valve compatibility has got a lot better. Not just for gaming, although that is a major part of it, but for Windows apps in general. Also, it has become a lot easier to write cross platform apps that aren't awful, and which feel native to each of Linux, Windows, and MacOS.
There are still issues, but unless Microsoft can really turn things around, and with AI making new hardware unaffordable for many, it's just possible that Linux really breaks through in the next year or two.
That's not what is happening. The first time you enable sideloading, there is a 24 hour cooling off period. After that, installs are instant. And there is probably no delay at all if you get F-Droid from Google Play, only if you download the APK and install it that way.
The other issue is that Google is requiring all developers to register if they want their apps to install on certified devices. Certified means runs Google Play and passes the security checks that some apps (e.g. banks apps, Google Pay) require. This is already an issue for Graphene OS users, because such apps will refuse to run for them due to the lack of certification.
It's hardly new though. Apple has required it since day one, and on Windows drivers need to be signed which means identifying yourself to a certificate authority. I imagine the solution will be the same as it is on Windows - someone will register and then sign F-Droid and all the apps offered on F-Droid. If you didn't know, F-Droid builds apps itself. The app source must compile on F-Droid's servers, with various limitations like not having any dependency on Google Play Services.
So, the reality is that the world "ran out of training data" for the most part years ago, and the models have gotten exponentially better relative to a number of parameters. Claude 3.7 Sonnet was released 1 1/2 years ago, and today it benchmarks about the same as Qwen 3.6 Sonnet 9B, a model two orders of magnitude smaller than it, and which is itself two generations out of date. And a large chunk of this is done with synthetic data - aka, data created by other models.
It's simply a myth that "data created by models consumed by other models makes them worse". In practice, it's used to make them vastly better. Models aren't collagers, they're reasoners. Learning the results of reasoning, the results of trial and error, etc helps build a stronger base for more advanced reasoning. Also, our training algorithms, while reaching a denser knowledge compression than human brains, are less efficient learners than human brains (per unit data), so they need to see the same sort of data from "many different angles", to substitute for our process of "mulling over" new information.
(Yes, it is possible to set up contrived scenarios where, say, an small image model is fed only its own outputs on loop, little bits of knowledge slowly being lost each go-round, in a situation equivalent to leaving a person alone with their thoughts in a dark room for ten thousand years - but even a tiny percent of new fresh data added to the mix prevents this degradation.)
And as for the article itself, they made it sound like they're talking about, say, programmers banned from using AI at OpenAI, but it's nothing of the sort. These are data labelers. In the old days, they used to be far more common, and in wide use in all types of model creation. That's no longer the case; they exist for special cases. For LLMs, this is much more limited:
* Subject matter experts: people with rare professional-tier knowledge. Often used to validate model outputs, where nobody else could (for example, OpenAI hires mathematicians to validate their models' proofs)
* Chain of thought / logic auditing. Increasingly important now that models are showing increasing signs of poor alignment. You can automate this a lot, but you really still do want a human in the loop *somewhere*, in case your auditors get compromised.
* Side by side comparative rating: Which model's output do you like more, A or B?
* Evaluating reported outputs where users reported that they thought the response they received was bad, and if there's actually anything wrong, copyediting the output for training.
* Adversarial prompt generation / jailbreaking and evaluation. Again, you *can* have models do this (and companies often do), but you don't want to just rely on them.
* Trying to set the bounds on whether given queries should be refused or not (for example, "How do explosive reactions happen in chemistry?")
Basically, a switch from "click work" to "knowledge work". This is no longer the era of "Write a poem about cats" or "Explain how to solve this algebra problem" to build up a training dataset. You're getting paid to think, not to repeat a rote task.
Other types of labelers aren't as far along. Multimodal data is less advanced than text, so you'll still for example have people labeling things in videos, transcribing heavily-accented audio, grading text-to-video consistency, things of that nature. Probably the least advanced field is robotics, so there's still an awful lot of manual evaluation and correction in that.
But anyway, if you're hired to do any of the above, it's because they specifically want you to do that. Having an AI model do the above (beyond the listed caveats) entirely defeats the purpose.
No. A black box is something you don't understand how it works. We can understand any and every decision that they make. We simply have to look.
I guess the only positive thing I can say about them is: despite how tough the models were to work with then and get good content out of them, these people were generally obsessive over their waifus and porn, so at least they weren't making like 8-fingered monstrosities. You could tell that they spent many hours zoomed in, upscale-regenerating and collaging and photoshopping over every imperfection.
hasn't really taken over human actresses despite that industry usually being in forefront of all kinds of tech, I wouldn't bet on this one either.
I dunno. This isn't exactly my space, and I haven't even been involved much in image generation at all in years, but in the old days at least, the Stable Diffusion Reddit and the model sites like Civitai were just *flooded* with stuff from horny guys, like 80-90% of the content. It was really annoying. Due to anti-porn restrictions, the Stable Diffusion Reddit mostly got flooded with waifus or similar, but Civitai was mostly porn models, or at least general models finetuned to allow porn. People were even widely using waifus and porn in their tutorials and user guides for image generation tools, even though the the tools were just for general image generation purposes. Oh, and YHVH protect you if you dared complain to any of the waifu or porn posters about any of this - you'd get a flood of "WHAT, ARE YOU A PRUDE???", "HAVEN'T YOU EVER SEEN A WOMAN BEFORE???" , etc and get modded to oblivion. And perhaps the worst part of it was just how damned derivative it all was. There was zero creativity over any of it, zero diversity of artistic style, zero attempt to do something aesthetically new, just pure hormone-driven churn, generally either straightforward-photographic style, anime style, or a Midjourney-ish digital art style.
So I rather have to disagree with this statement. Porn users are a major share of AI image generation, at least with run-it-yourself models (as most commercial models ban it).
I mean, you're not wrong
It's like the whole "metaverse" thing of trying to reinvent shopping in a 3d gamelike environment. Even in video games it's common to not have to actually hunt for and pick items off the shelves in 3d, instead just choosing from a dialog. Just because something sounds shiny on paper doesn't actually mean it's good in the real world.
You don't need to see the person on the other side to chat with them. You don't even need to hear them in most cases - text is just fine, and actually helps prevent misunderstandings.
Jake Sully.
BASIC is to computer programming as QWERTY is to typing. -- Seymour Papert
