It's not AGPL 3.0. The binaries are MIT, the codebase (from where the MIT binaries are built from) is AGPL 3.0, except for the bits of the codebase that are Apache 2.0, and there's some kind of a promise about not enforcing a part of AGPL if you don't link to their platform directly and exclusively use the bits of the code that are Apache 2.0, and also don't make a modified version of the software. And also you can just license it commercially too.
It says you "may be licensed" to use the source code under AGPL v3.0, but never actually makes an unambiguous statement that suchandsuch code is licensed under AGPL v3.0.
The concept of MIT licensing a compiled software artifact, but not the code used to generate the artifact, is also extremely strange.
Right, the correct way here is to simply grant _everyone_ a license to _everything_ under the terms of the AGPL (or whatever). You can then separately license portions under other terms.
You don't need to note the commercial licensing option in the license itself; it's irrelevant to that grant. You just state that elsewhere.
Why is the government doing this, this seems like a ridiculous waste.
Here in Germany private corporations provide APIs for this. Google maps straight up tells you the price at nearby stations.
Maybe the UK government should focus on things such as their crumbling infrastructure, their almost non existent GDP growth or getting rid of their knife murderer and rapist population?
No it wasn't. A hardware defect so disastrous that it affects floating point computation on the neural engine, yet so minor that it does not affect any of the software on the device utilizing that hardware is exceedingly improbable.
The conclusion, that it was not the fault of the developer was correct, but assuming anything other than a problem at some point in the software stack is unreasonable.
The hardware itself is utilized by many pieces of software on any Apple device. Face ID uses it, Siri uses it, the camera uses it, there are also other Apple on device LLM features, where you could easily test whether the basic capabilities are there.
I highly doubt that you could have a usable iPhone with a broken neural engine, at the very least it would be obvious to the user that there is something very wrong going on.
All neural accelerator hardware models and all neural accelerator software stacks output slightly different results. That is a truth of the world.
The same is true for GPUs and 3d rendering stacks too.
We don't usually notice that, because the tasks themselves tolerate those minor errors. You can't easily tell the difference between an LLM that had 0.00001% of its least significant bits perturbed one way and one that had them perturbed the other.
But you could absolutely construct a degenerate edge case that causes those tiny perturbances to fuck with everything fiercely. And very rarely, this kind of thing might happen naturally.
You are correct that implementations of numerical functions in hardware differ, but I do not think you correctly understand the implications of this.
>And very rarely, this kind of thing might happen naturally.
It is not a question of rarity, it is a question of the stability of the numerical problem. Luckily most of the computation in an LLM is matrix multiplication, which is s extremely well understood numerical problem and which can be checked for good condition.
Two different numerical implementations on a well conditioned problem and which requires much computation, differing significantly would indicate a disastrous fault in the design or condition of the hardware, which would be noticed by most computations done on that hardware.
If you weigh the likelihood of OP running into a hardware bug, causing significant numerical error on one specific computational model against the alternative explanation of a problem in the software stack it is clear that the later explanation is orders of magnitude more likely. Finding a single floating point arithmetic hardware bug is exceedingly rare (although Intel had one), but stacking them up in a way in which one particular neural network does not function, while other functions on the hardware run perfectly fine, is astronomically unlikely.
I have seen meaningful instability happen naturally on production NNs. Not to a truly catastrophic degree, but, when you deal in 1024-bit vectors and the results vary by a couple bits from one platform to another, you tend to notice it. And if I've seen it get this bad, then, surely someone has seen worse.
> The conclusion, that it was not the fault of the developer was correct, but assuming anything other than a problem at some point in the software stack is unreasonable.
What do you mean? The developer is perfectly justified in being upset over a basic example not functioning correctly, due to bug on behalf of Apple's developers. It just wasn't reasonable to assume that the bug was due to malfunctioning hardware.
In my experience of University much of student activism was done by students who struggled the most academically and often made it part of their agenda to ease academic standards.
You had an outlier experience, unfortunately. I went to a state school, and most of even my student government class wound up prominently in the Congress or White House.
Why do you think that it is a given that politicians perform better in school? Here in Germany many prominent politicians were academic failures or dropouts.
We had student council elections, where the candidates were listed by degree and the semester, the trend was overwhelmingly that the candidates were in semesters after the usual graduation date, often having studied for more then 5 years.
I said that student activists struggled academically, your counter point was that at your school student activists made it into high political offices.
What other interpretation of your words could there be except you believing that people in high political offices are above average students.
You are posting this on something which can only be described as "social media".
There is enormous value in letting humans communicate freely through the internet and any harm which comes from the misuse needs to be very carefully weighed against the benefits.
By the way, fossile fuels are partly responsible for the greatest uplift in human quality of life. Pesticides are essential in creating a sustainable food output.
Collectively, "social media" is basically everything that allows interactive expression and evolution of different opinions online. The manipulation of Big Social is simply a deterrent to that process, in effect it's a tool against good social media.
> any harm which comes from the misuse needs to be very carefully weighed against the benefits.
Also, misuse can be addressed in different ways, blanket bans and surveillance-enabling authentication are misuse themselves.
> By the way, fossile fuels are partly responsible for the greatest uplift in human quality of life. Pesticides are essential in creating a sustainable food output.
Yeah, and they are not sentient beings we owe something in return to.
When they were useful we used them. Now that they’re more harmful than useful we stop using them.
>Autonomous cars, drones cheerfully obey prompt injection by road sign
This is just a lie and has nothing to do with the research, which in a baffling decision does not appear to be linked in the article. https://arxiv.org/pdf/2510.00181
This is of course a trend in many western countries. With some, like the UK and Australia, leading the way.
At this point I do not think it is reasonable to deny the harm that certain modes of social interactions over the internet have caused. At the same time these bans should not be considered reasonable options. They exist to cover for the decade of inaction of politicians in addressing youth dissatisfaction and dysfunction.
A reasonable approach should not assume that the root cause of this dysfunction is youth interacting with social media, but should consider what lead to this in the first place. Apparently most adults seem to be capable of dealing with this situation, if they are not why would this ban, or at least some regulation, not extend to social media for adults.
In general I believe that dysfunction in the youth has multiple causes and that overuse of social media is just on part of the puzzle and that unhealthy use of social media is often caused by other problem and used as a coping mechanism.
These bans will not be effective and they will be assaults on the free internet, as the bureaucrats establishing the laws are also seeking to control the internet for themselves and will use this as a backdoor.
> At this point I do not think it is reasonable to deny the harm that certain modes of social interactions over the internet have caused
Yes, it is reasonable to doubt the purported harms are real, because
1) I've yet to see evidence that the medium is the problem,
2) people keep telling me that they don't need evidence because the harms are obvious, and
3) I have an strong prior, as an American, that anyone preventing people sharing ideas with each other is a villain of history.
The furor over youth social media has all the hallmarks of a moral panic, including over-reliance of weak evidence, personal attacks against skeptics, and socially disruptive remedies of dubious efficiency, the collateral damage of which people justify by pointing to harms to children they say, falsely, are obvious and ongoing.
I'm not convinced that these social media bans are solving a real problem. The more people breathlessly tell me I'm a bad person for asking for evidence of the alleged harms, the more I think it's a public mania, not a civilizational problem.
It really doesn't help that it'd be suspiciously convenient for the worst actors in power if sharing ideas on the internet required ID.
For the reasons outlined in my post I believe that it is hard to show specific causal claims which relate overuse of mobile devices and especially social media to specific problems. Although I think for some specific cases this could still be reasonably inferred.
Just to be clear, the evidence seems overwhelming. This is not some novel research field, but this questions has been researched for long enough to have been pretty conclusively answered.
>1) I've yet to see evidence that the medium is the problem,
This is not relevant to the claim. The claim is that the specific usage pattern of young adults is harmful to their development.
>I'm not convinced that these social media bans are solving a real problem.
People do not realize how dire the current economic situation is. Many of the large traditional businesses are on the verge of becoming unprofitable.
The monumental task of ripping out the IT systems they have built up over the last few decades, to move away from the US will actively threaten the existence of some of these companies.
People are living in a fantasy land where e.g. Germany has an enormous automotive industry which can be arbitrarily regulated and still be profitable enough to keep the German economy afloat. This is non longer the case and many EU companies are currently struggling for their existence.
Many European companies are already struggling financially, especially the large traditional businesses which form the backbone of the European economy at large.
Now you demand that these companies should rip out decades of the IT systems they have built up, which form the backbone of their day to day operations and replace them with third rate alternatives, nowhere near in capability, support and coverage?
Yes, I love open source and I wish to see it succeed, but this proposal is suicidal. Even if a superior and less costly alternative did exist (and it does not, just to be clear), just the effort of switching over would ruin these companies.
Expecting the crash of the most important economy in the world based on two graphs, where you do chart astrology, is such an insanely stupid argument it is hard to fathom.
With all these charlatans predicting imminent collapse it is always imperative to consider how strongly they believe in their revealed preferences, based on how much they have invested in their position. That said, how much money does OP have invested e.g. in shorting the S&P 500? Or any equivalent. Let me guess, zero dollars.
reply