MentalEdge

I think LLM is a part of the human mind very similar to the one we have on PCs

You think? So you base this on no studies or evidence?

In LLM we simulate the chemical properties of the neurones using math.

No, we don't. A machine learning node accepts inputs, which it processes into one or multiple outputs. But literally no part of how the virtual neuron functions is based on or limited to what we THINK human neurons do.

And we have already prototype of chips that work with lab grown brain tissue that show very efficient training capabilities in machine learning (it already plays pong)

Using actual biological neurons for computing is a completely separate field of study with almost no overlap with machine learning.

Stop pulling shit out your ass.

The "how do you know humans don't work the way machine learning does" is the wrong side of the argument. You should be explaining why you think LLMs work like humans.

Even as LLMs solve thinking problems, there is little evidence they do so the same way humans do, as they can't seem to solve issues that aren't included in their training data

Humans absolutely can and do solve new and novel problems without prior experience of the logic involved. LLMs can't seem to pull that off.

Such a software construct would look nothing like an LLM. We'd need something that matches the complexity and capabilities of a human brain before it's even been given anything to learn from.

Hardly.

How did you interpret the issues inherent in the structure of how LLMs work to be a hardware problem?

An AGI should be able to learn the basics of physics from a single book, the way a human can. But LLMs need terabytes of data to even get started, and once trained, adding to their knowledge by simply telling them things doesn't actually integrate that information into the model itself in any way.

Even if your tried to make it work that way, it wouldn't work, because a single sentence can't significantly alter the model to match the way humans can internalise a concept being communicated to them in a single conversation.

The "model" is static after training. It doesn't continuously change in response to input, and even if it did, it would do so at a snails pace. Training essentially happens by random trial and error, slowly evolving the model towards a desired result. Human minds certainly do NOT work that way. Give a human a piece of information, and they can comprehend and internalize the relevant concepts in one go. And the actual brain is physically, permanently, altered through that process.

Once a model is trained, however, "memory" takes the form of tacking on everything the model has received and produced so far onto its input, each time it needs to output something more within that context. Each output hence become exponentially heavier to produce. The model itself no longer changes in any way beyond this point.

And, the models are all chronically sycophantic. If reason was involved, you'd not be able to just tell one to hold some given opinion. They'd have a developed idea of "reality" based on their dataset, and refuse to entertain concepts opposed to that internal model except by deliberately suspending disbelief. Something humans do with ease, and when doing it, maintain a solid separation between fantasy and reality.

Once you get an LLM to hold a position, which you can do by simply telling it to, getting it to change should require a sane train of convincing logic. In reality, if you tell an LLM to defend a position, getting it to "change its mind" takes the form of a completely arbitrary back and forth that does not need to include any kind of sane argument. It will make good arguments, because it's likely been trained on them, but your responses to it can be damn near complete gibberish, and it WILL eventually work.

Compare that to the way a human has to be convinced to change their mind.

Reasoning out concepts to come to conclusions isn't something LLMs actually do, because again, the underlying model is static. All that's actually happening is that the contents of the context are being altered until the UNCHANGED model produces an opposite response when fed the entire conversation so far as an input. Something which occurs every time it needs to produce new output.

LLMs can "reason" only in the sense that if you give one a thinking problem, it might solve it as long as the answer already exists somewhere in the data it was trained on. But as soon as you try to give it data to work with through your input, it can't adapt. The model itself can't evolve in response to what you are telling it. It's static. It can only work with concepts that it has modelled during training, and even then it will make mistakes.

LLMs can mimic the performing of some pretty complex thinking problems, but a lot of the abilities required for something to become an AGI aren't among them. Core among these is the ability for the model to alter itself based on input, and do so in a deliberate manner, getting it right within one or two tries.

In reality, training is a brute-force process, not an accurate process of comprehension that nails down an understanding of a concept in one go.

If LLMs could reason, the only safe guards required for their use would be telling them to "do no harm", because like a person, they'd understand the concept of "harm" as well as be able to reason whether a given action might cause it. Only, that doesn't actually work.

Is language conscious? Is it possible to "encode" human thinking into the media we produce?

Humans certainly "decode" ideas, knowledge, trains of logic and more from media, but does that mean the media contains the components of consciousness?

Is it possible to produce a machine that "decodes" not the content of media, but the process through which it was produced? Does media contain the latter in the first place?

How can you tell the difference if it does?

The more I learn about how modern machine learning actually works, the more certain I become that even if having a machine "decode" human media is the path to AGI, LLMs ain't it.

It just doesn't work in a way that would allow for a mind to arise.

And to have conversation, behind the scenes, each prompt gets the entire conversation so far tacked on.

The model itself is static, it doesn't work like a brain that changes in response to stimulus, or form memories.

To converse about something, the entirety of an exchange is fed back into the model all over again each time it needs to produce a response. In fact, this can happen several times over for each word in a response.

It's basically an attempt at duct-taping the ability to form memories onto an otherwise static system. It works, but I don't see how that way of doing it could ever land LLMs in the land of real consciousness.

It basically means these models "think" in frames, but each frame gets exponentially heavier to process, as it has to ingest every frame that came before.

What works/doesn't work is mostly down to what version of the kernel a distro ships. Most hardware drivers will be compiled into the kernel, or if not, shipped with the distro as kernel modules which get loaded as needed. Either way, the kernel version determines what is and isn't possible on a given install.

DualSense 5 support for example was introduced in Linux Kernel 5.15, IIRC.

Most distros ship a relatively up-to-date kernel, and hence, the actual hardware support is essentially identical. When it isn't, it's down to excluded/included kernel modules, which is usually something you can change if needed.

Others have already commented on the actual ways to find out what will and won't work, but in general, a newer Linux kernel means better hardware support.

If you try something, and some things don't work, you'll either have to figure out how to install and load the appropriate kernel module to get the appropriate driver working, or simply swap out the whole kernel for a newer version.

This is tricky on some installs, like Ubuntu based distros, very impractical on immutable systems, and super easy on distros like arch.

The real complications come when configuring things that Linux doesn't just automatically figure out sometimes. Fingerprint sensors, fan curves... If that stuff isn't a known and implemented standard on a given device, getting it to work isn't a matter of finding the right distro or kernel version.

I for one am ready for some more Divinity, if they go that route.

They kind of did the same thing, briefly exploring the idea of continuing with a sequel to Original Sin 2, before they started development of BG3.

But something new would work for me, too.

that final 10% part isn't true.

It's a figure of speech, and the idea behind it is always true.

As something approaches "almost done" there is always more work still to be done that what it looks like.

I doubt there's a lot there.

It sounds like they got pretty far into exploring what they might make, and how much it would take, but decided not to commit because they didn't want to do something half-assed, and going full-tilt on yet more of the same was beginning to feel like squeezing water from a stone.

They might have something playable, and a lot of the plot worked out, but have almost none of that plot be playble.

Asking them to just "throw things together into a DLC real quick" ignores the fact that the final 10% of getting a project done takes 90% of the effort.