Yeah, but it also duplicates the cat owners, so there’s one whose cat has died and a different one whose cat still lives.

Thanks, the paper linked from that Wikipedia entry is very interesting.

I haven’t heard of bitnet.

Then again my knowledge of Machine Learning is 3 decades old (so, even before Recurrent Neural Networks were invented, much less Attention) and then some more recent reading up on LLMs from an implementation point of view to understand at least a bit how they work (it’s funny how so much of the modern stuff is still anchored in 3 decades old concepts).

Right, if I understood it correctly, what you see as “IF” is the multi-headed attention stuff. I was under the impression that you can’t actually have non-contiguous functions there so even multi-headed attention stuff involves functions which don’t have +/- infinity in their first derivative - they can boost or suppress inputs but they don’t have the hard YES/NO transitions of logical IF.

However the Genetic Algorithms stuff is something completelly different from Neural Networks: it’s basically an Evolutionary method of finding the best “formula” to process inputs to generate the desired outputs by assessing different variants of the “formula” with the training data, picking the best ones and then generating a new generation of “formula” variants from the best ones and assessing those and keep doing it until the error rate is below a certain value - it’s basically a way of using “Natural” Selection for mathematical formulas.

As far as I can tell Genetic Algorithms can’t really scale to the size of something like an LLM (the training requirements would be even more insane) though I guess that technique could be used to train part of a Neural Network or to create functional blocks that worked together with NNs.

And yeah, MLPs trained via simple Backpropagation are exactly what I’m familiar with, having learned that stuff 3 decades ago as part of my degree when that was the pinnacle of NN technology and model architectures were still stupidly simple. That’s why I would be shocked if a so-called ML “expert” didn’t recognize that, as it’s the most basic form of Neural Network there is and it’s being doing the rounds for ages (that stuff was literally used to in automated postal code recognition in letters for automated mail sorting back in the 90s).

I would expect that for people doing ML a simple MLP is as recognizable as binary is for programmers - sure people don’t work at that level anymore, but at they should at least recognize it.

You mean that they’re actually competing multiple variants of a model against each other to see which ones get closer to generating the expected results, and picking the best ones to create the next generation?

Because that’s how Genetic Algorithms work and get trained, which is completelly different from how Neural Networks work and get trained.

Also the links in Neural Networks don’t at all use IF-functions: the output of a neuron is just a mathematical operation on the values of all its inputs (basically a sum of the results of a function applied to the input numbers, though nowadays there are also cyclic elements) - the whole thing is just floating values being passed down the network (or back up the network during training) whilst being transformed by some continuous function or other with no discontinuity like you would get with IF involved.

I was under the impression (from what I learned about how planes fly) that the spirals were actually a bad thing since energy was lost to turbulence via those (and hence why commercial jets have winglets on their wingtips) and the good part is the laminar flow of the air over and under the wings.

Is this not so in animal flight?

Yeah, if you’re supposedly in AI/ML and don’t recognize a (stupidly simplified) diagram for a Neural Network, you don’t really make stuff with it, you’re just another user (probably a “prompt engineer”).

Even people creating Machine Learning solutions with other techniques would recognize that as representing a Neural Network.

That should be as recognizable to a professional in that domain as a long string of 0s and 1s would be recognizable as binary to a programmer - even if you’re not working with it at that level, you recognized such building blocks of your trade.

I haven’t really done Neural Networks in 2 decades, and was under the impression that NNs pretty much dominate Machine Learning nowadays, whilst stuff like Genetic Algorithms were way less popular or not at all used anymore.

Is that the case?

There is literally no way in which a nation which is already well in its post Empire decay period can improve its lot by cutting ties with its natural partners, more so if it’s doing it to get closer ties with another nation now entering its own post Empire decay period.

Maybe if the UK was an “up and coming” young nation not hobbled by a thick web of cronyism, nepotism and socially and culturally decaying lazy elites more worried about maintaining their privileges and not being caught diddling children, than about making shit happen, and a population that in their most believes that “people should know their place”, this stupid, stupid idea might actually not hinder Britain’s progress too much, but for this walking corpse of a once mighty nation, cutting themselves from a “together we’re stronger” group like that isn’t going to end well.

Then again, if Britain wasn’t a decaying corpse I doubt the idea of Brexit would even come to be, much less gain enough votes to pass, as the required mix of societal calcification, desperation and the masses having long been manipulated into compliance and looking for the enemy outside rather than the enemy within, wouldn’t be there, so those many decades of “blame the EU” by politicians and the Press would not be there, a Tory politician would not believe he could boost himself in the polls by running a referendum to leave the EU and in such a referendum it would not be possible to convince enough people that they would be better of the country left the EU.

Whatabout … China!?

Won’t somebody think about … China???

Funny that when reading “covid-19 prevention” you forgot anti-maskers - which is actually a very visible “I win” statement - but instead went for not being vaccinated, which is not at all a visible thing hence nowhere as much a “I win” statement.

Finally, some good news!

A quick look through its documentation shows that it instructs the user how to go through a subset of the instructions the original user provided (or an alternative set of instructions if using Android 11+ as there it can use a different mechanism) plus a few more, in order to run a Shizuku service as user “adb”.

From then on, that Shizuku service can then be used by other apps to do everything the “adb” user can, including installing and updating applications.

So I guess it could be used by something like F-Droid to go around Google’s new mechanism to close down app installs.

For Android < 11 it’s is no more non-expert friendly than the instructions already provided by the original user, though it’s better in Android 11+ as there it’s all interacting with menus on the Android side (see here under Start Shizuku)

Most of that stuff is automatable - except the bit about activating Developer mode and USB Debugging on the device (steps 3 to 6) which only needs to be done once per device - so I expect we will soon see several nice GUI tools that automate the rest and eventually we might even just see stuff that talks directly to the phone over USB via libusb and using the same protocol as ADB, so installing the Android Platform Tools won’t at all be needed.

But yeah, at this point it requires people to at the very least be familiar with using the command line.

Well, I did do app development for Android for a couple of years, so I’ll be using ADB it install APKs in any device affected if needed.

I’ll also never do development work for Android ever again, beyond making utilities for myself if need something like that.

Beyond that, I’ll never buy an Android device that cannot be unlocked. Last one I got was a Xiaomi phone, which at the time could be unlocked (which I did and installed an alternative ROM on it before I started using it), but they stopped that so Xiaomi isn’t going to be getting any more money from me.

Mid to long-term, I expect Linux devices are the solution. I’m especially interested in getting a Linux tablet (7" or 8") to replace the tablet I currently use mostly for book reading and internet browsing when I’m out and about (hence the size needs to be small enough to fit a back or jacket pocket).

When I started looking into it, my expectation was that Linux tablets would make even more sense as devices than phones since they’re closer to notebooks in terms of how they’re used, but I haven’t really found all that many out there - there are more Linux phones than tablets - and all of them were 10" or more (so, too large for my use case).

(PS: suggestions welcome, even just stuff I can root and install something like Ubuntu Touch on it)

Am I so unusal in wanting an portable computing device with a big enough screen to read stuff, for the purpose of consuming media rather than working on (so no keyboard need), which is not so big that I need to haul it in a backpack, not a full-blown smartphone with all the bells as whistles (I already have a smarphone on my pocket with mobile data, camera and GPS, so why would I need that shit AGAIN on a tablet???) and not a locked-down system like iOS or Android?

Answer: Poland is running forwards, the UK is running backwards.

Curiously, this very newspaper is did a lot of work in making the UK run backwards even faster, most notably during the Leave Referendum.

This is really just a driver which sends a bunch of bytes via I2C to a microcontroller.

I2C is a very standard way of communicating with digital integrated circuits at low speed so this is not specific to the microcontroller used on Synology NAS devices (which is actually a pretty old and simple one) much less specific to drive leds.

So whilst technically this specific Linux Driver ends up controlling LEDs on a very specific device, the technique used in it is way more generic than that, and can be used to control just about any functionality sitting behind a digital integrated circuit that exposes an interface to control it via I2C, be it one that hardcodes it or one which, like this one, is a microcontroller which itself implements it in code.

All this to say that this is a bit bigger than just “LED driver”.

The same guys relentlessly defending the child mass murders in Gaza, here cosplay as being against child sexual abuse.

The irony is so massive that I’m surprised a black hole didn’t form at Politico’s main office in Germany.

Also, train lines and pipelines too can be targetted by autonomous explosive delivery mechanisms, though they’re probably a lot easier to repair than refineries.

That would be a correct reading if it wasn’t for the small (tiny, miniscule, tinsy winsy) detail that when it comes to what the IDF or the Israeli government allege this very newspaper actually uses words like “stated” instead of “alleged”.

(Similarly to how they write for Israeli soldiers that they “are killed” whilst for Palestinians the form used is the passive one: they “die”)

It’s the double standard that makes it propaganda rather than journalism - Journalists would go to great lengths to treat both sides exactly the same.