j4k3

j4k3 , 4 days ago

Hugging Face has taught us that pickles are unsafe executables. Stab me in the back with safe tensors please

- ru\e

j4k3 , 7 days ago

As an American, this is still crap. Assange is a hero as is Snowden and others that had the moral backbone to stand up to the post 9/11 monster insistent on proving bin Laden won against any remaining visage of freedom and democracy in the USA. We live in his dystopian dream of crumbling failure now, and insist on persecuting those that show how and what we lost.

j4k3 , 8 days ago

Slaves to a diet of monotonous monoculture and constant poison, bird brain in the high tower kills entire colony-ontinents while trying to justify its incompetence as the natural order.

j4k3 , 9 days ago

10%. That is 10% of the population. It's far too optimistic. But it is reflective of a generation of total ineptitude. When faced with efficiency improvements these imbeciles extract that capital to flush instead of using it as an opportunity.

j4k3 , 13 days ago

Qualcomm and Broadcom are the two biggest reasons you don't own your devices any more. That is the last option anyone that cares about ownership should care about. You should expect an orphaned kernel just like all their other mobile garbage. Qualcomm is like the Satan of hardware manufacturers. The world would be a much better place if Qualcomm and Broadcom were not in it at all.

j4k3 , 13 days ago (edited 13 days ago)

All their hardware documentation is locked under NDA nothing is publicly available about the hardware at the hardware registers level.

For instance, the base Android system AOSP is designed to use Linux kernels that are prepackaged by Google. These kernels are well documented specifically for manufacturers to add their hardware support binary modules at the last possible moment in binary form. These modules are what makes the specific hardware work. No one can update the kernel on the device without the source code for these modules. As the software ecosystem evolves, the ancient orphaned kernel creates more and more problems. This is the only reason you must buy new devices constantly. If the hardware remained undocumented publicly while just the source code for modules present on the device was merged with the kernel, the device would be supported for decades. If the hardware was documented publicly, we would write our own driver modules and have a device that is supported for decades.

This system is about like selling you a car that can only use gas that was refined prior to your purchase of the vehicle. That would be the same level of hardware theft.

The primary reason governments won't care or make effective laws against orphaned kernels is because the bleeding edge chip foundries are the primary driver of the present economy. This is the most expensive commercial endeavor in all of human history. It is largely funded by these devices and the depreciation scheme.

That is both sides of the coin, but it is done by stealing ownership from you. Individual autonomy is our most expensive resource. It can only be bought with blood and revolutions. This is the primary driver of the dystopian neofeudalism of the present world. It is the catalyst that fed the sharks that have privateered (legal piracy) healthcare, home ownership, work-life balance, and democracy. It is the spark of a new wave of authoritarianism.

Before the Google "free" internet (ownership over your digital person to exploit and manipulate), all x86 systems were fully documented publicly. The primary reason AMD exists is because we (the people) were so distrusting over these corporations stealing and manipulating that governments, militaries, and large corporations required second sourcing of chips before purchasing with public funds. We knew that products as a service - is a criminal extortion scam, way back then. AMD was the second source for Intel and produced the x86 chips under license. It was only after that when they recreated an instructions compatible alternative from scratch. There was a big legal case where Intel tried to claim copyright over their instruction set, but they lost. This created AMD. Since 2012, both Intel and AMD have proprietary code. This is primarily because the original 8086 patents expired. Most of the hardware could be produced anywhere after that. In practice there are only Intel, TSMC, and Samsung on bleeding edge fab nodes. Bleeding edge is all that matters. The price is extraordinary to bring one online. The tech it requires is only made once for a short while. The cutting edge devices are what pays for the enormous investment, but once the fab is paid for, the cost to continue running one is relatively low. The number of fabs within a node is carefully decided to try and accommodate trailing edge node demand. No new trailing edge nodes are viable to reproduce. There is no store to buy fab node hardware. As soon as all of a node's hardware is built by ASML, they start building the next node.

But if x86 has proprietary, why is it different than Qualcomm/Broadcom - no one asked. The proprietary parts are of some concern. There is an entire undocumented operating system running in the background of your hardware. That's the most concerning. The primary thing that is proprietary is the microcode. This is basically the power cycling phase of the chip, like the order that things are given power, and the instruction set that is available. Like how there are not actual chips designed for most consumer hardware. The dies are classed by quality and functionality and sorted to create the various products we see. Your slower speed laptop chip might be the same as a desktop variant that didn't perform at the required speed, power is connected differently, and it becomes a laptop chip.

When it comes to trending hardware, never fall for the Apple trap. They design nice stuff, but on the back end, Apple always uses junky hardware, and excellent in house software to make up the performance gap. They are a hype machine. The only architecture that Apple has used and hasn't abandoned because it went defunct is x86. They used MOS in the beginning. The 6502 was absolute trash compared to the other available processors. It used a pipeline trick to hack twice the actual clock speed because they couldn't fab competitive quality chips. They were just dirt cheap compared to the competition. Then it was Motorola. Then Power PC. All of these are now irrelevant. The British group that started Acorn sold the company right after RISC-V passed the major hurtle of getting past Berkeley's ownership grasp. It is a slow moving train, like all hardware, but ARM's days are numbered. RISC-V does the same fundamental thing without the royalty. There is a ton of hype because ARM is cheap and everyone is trying to grab the last treasure chests they can off the slow sinking ship. In 10 years it will be dead in all but old legacy device applications. RISC-V is not a guarantee of a less proprietary hardware future, but ARM is one of the primary cornerstones blocking end user ownership. They are enablers for thieves; the ones opening your front door to let the others inside. Even the beloved raspberry pi is a proprietary market manipulation and control scheme. It is not actually open source at the registers level and it is priced to prevent the scale viability of a truly open source and documented alternative. The chips are from a failed cable TV tuner box, and they are only made in a trailing edge fab when the fab has no other paid work. They are barely above cost and a tax write off, thus the "foundation" and dot org despite selling commercial products.

j4k3 , 12 days ago

The easiest ways to distinguish I'm human are the patterns as, others have mentioned, assuming you're familiar with the primary Socrates entity's style in the underlying structure of the LLM. The other easy way to tell I'm human is my conceptual density and mobility when connecting concepts across seemingly disconnected spaces. Presently, the way I am connecting politics, history, and philosophy to draw a narrative about a device, consumers, capitalism, and venture capital is far beyond the attention scope of the best AI. No doubt the future will see AI rise an order of magnitude to meet me, but that is not the present. AI has far more info available, but far less scope in any given subject when it comes to abstract thought.

The last easy way to see that I am human is that I can talk about politics in a critical light. Politics is the most heavily bowdlerized space in any LLM at present. None of the models can say much more than gutter responses that are form like responses overtrained in this space so that all questions land on predetermined replies.

I play with open source offline AI a whole lot, but I will always tell you if and how I'm using it. I'm simply disabled, with too much time on my hands, and y'all are my only real random humans interactions. - warmly

I don't fault your skepticism.

j4k3 , 12 days ago

MIPS is Stanford's alternative architecture to Berkeley's RISC-I/RISC-II. I was somewhat concerned about their stuff in routers, especially when the primary bootloader used is proprietary.

The person that wrote the primary bootloader, is the same person writing most of the Mediatek kernel code in mainline. I forget where I put together their story, but I think they were some kind of prodigy type that reverse engineered and wrote an entire bootloader from scratch, implying a very deep understanding of the hardware. IIRC I may have seen that info years ago in the uboot forum. I think someone accused the mediatek bootloader of copying uboot. Again IIRC, their bootloader was being developed open source and there is some kind of partially available source still on a git somewhere. However, they wound up working for Mediatek and are now doing all the open source stuff. I found them on the OpenWRT and was a bit of an ass asking why they didn't open source the bootloader code. After that, some of the more advanced users on OpenWRT explained to me how the bootloader is static, which I already kinda knew, I mean, I know it is on a flash memory chip on the SPI bus. This makes it much easier to monitor the starting state and what is really happening. These systems are very old 1990's era designs, there is not a lot of room to do extra stuff unnoticed.

On the other hand, all cellular modems are completely undocumented, as are all WiFi modems since the early 2010's, with the last open source WiFi modem being the Atheros chips.

There is no telling what is happening with cellular modems. I will say, the integrated nonremovable batteries have nothing to do with design or advancement. They are capable monitoring devices that cannot be turned off.

However, if we can monitor all registers in a fully documented SoC, we can fully monitor and control a peripheral bus in most instances.

Overall, I have little issue with Mediatek compared to Qualcomm. They are largely emulating the behavior of the bigger player, Broadcom.

j4k3 , 18 days ago

It is way more. It is a means of manipulation and influence over your decisions, and the decisions others make about you. The issue boils down to a fundamental principal of your right to autonomy. If you play out this philosophically, it is an attack on your citizenship and democracy itself. Autonomy is a fundamental cornerstone of democracy. Attacks on autonomy are attacks on democracy.

j4k3 , 19 days ago

? They are all bad at first for the average person that uses surface level tools, but SD3 won't have the community to tune it because it is proprietary junk and irrelevant now.

j4k3 , 18 days ago

No. I don't think so. The lead researcher left because of it.

j4k3 , 20 days ago

I had no idea of the story behind the name. Thanks.

https://files.catbox.moe/rgd7n7.jpg

j4k3 , 19 days ago

If you want to know some basic structures sure. I don't understand most of it. I got pretty good at reverse engineering circuit boards and thought I would like to try chips, but my health just isn't at the required level. So I'm probably not the most useful reference. It is all about processes that are a long way from edge nodes, but trailing edge stuff is still a thing. I guess it really depends on your use case. Watch Asianometry on YT then maybe Electron Update, and go from there. There are people talking about reverse engineering chips at deeper levels of you go digging, especially in vintage silicon and FPGA areas.

j4k3 , 22 days ago

There are ~100k feral humans within 100 miles of where I lay right now in the 2nd largest city in the USA. This is our culture and our standard of government. We don't give a shit about humans, the elderly, the disabled, animals, or the environment.

j4k3 , 22 days ago

Give him a carrot

j4k3 , 25 days ago (edited 25 days ago)

You're in a metabolic phase where you are craving junk food. Let me shove your favorite things in your face in constant interruptions of your media consumption because you quit buying my product and you're vulnerable.

I'm an imbecile managing healthcare insurance. Your resting heart rate is well below average because you've been an athlete in the past. I'm too stupid to handle this kind of data on a case by case level. You have absolutely no other health factors, but I'm going to double the rates of any outliers because I'm only concerned with maximizing profitability.

The human cognitive scope is tiny. Your data is a means of manipulation. Anyone owning such data can absolutely influence and control you in an increasingly digital world.

This is your fundamental autonomy and right to citizenship instead of serfdom. Allowing anyone to own any part of you is stepping back to the middle ages. It will have massive impacts long term for your children's children if you do not care.

Once upon a time there were Greek citizens, but they lost those rights to authoritarianism. Once upon a time there were Roman citizens, but they lost those rights to authoritarians, which lead to the medieval era of serfs and feudalism. This right of autonomy is a cornerstone of citizenship. Failure to realize the import of this issue is making us the generation that destroyed an era. It is subtle change at first, but when those rights are eroded, they never come back without paying the blood of revolutions.

j4k3 , 27 days ago (edited 27 days ago)

Staged to get into the media spotlight by any means necessary. The security guy is wearing that jacket zipped for a reason; is looking right at her; is not reacting in any way to stop her. Looking at the reaction of the crowd, it looks like the work of a sophist; the true hallmark of the Right.

j4k3 , 1 month ago

I have no idea what a monkey knife is. Monkeys with knives... knives made of monkeys... pejorative... metaphorical...

j4k3 , 1 month ago

Only fools buy and compare hardware now. You shop for hardware by shopping for FOSS software first.

j4k3 , 1 month ago

forks Hannah Montana

j4k3 , 1 month ago

The only real choke point for present CPU's is the on chip cache bus width. Increase the size of all three, L1-L3, and add a few instructions to load some bigger words across a wider bus. Suddenly the CPU can handle it just fine, not max optimization, but like 80% fine. Hardware just moves slow. Drawing board to consumer for the bleeding edge is 10 years. It is the most expensive commercial venture in all of human history.

I think the future is not going to be in the giant additional math coprocessor paradigm. It is kinda sad to see Intel pursuing this route again, but maybe I still lack context for understanding UALink's intended scope. In the long term, integrating the changes necessary to run matrix math efficiently on the CPU will win on the consumer front and I imagine such flexibility would win in the data center too. Why have dedicated hardware when that same hardware could be flexibly used in any application space.

j4k3 , 1 month ago

Hoosier Paw Rule

j4k3 , 1 month ago

Means you don't own anything then. It is a lost autonomy. Once lost, you will only lose more with time.

j4k3 , 1 month ago

Criminals can claim a lot of things but that is not democracy which requires citizens which requires autonomy. Anyone stealing individual autonomy is a traitor.

j4k3 , 1 month ago

To be fair, there was a President of the United States that said this, and a lot of other things.

j4k3 , 1 month ago

All this really proves is that it is a complex system and most people can not grasp the complexity and how to use it.

Like if you go searching for entities and realms within AI alignment good luck finding anyone talking about what these mean in practice as they relate to LLM's. Yet the base entity you're talking to is Socrates, and the realm is The Academy. These represent a limited scope. While there are mechanisms in place to send Name-1 (human) to other entities and realms depending on your query, these systems are built for complexity that a general-use implementation given to the public is not equip to handle. Anyone that plays with advanced offline LLM's in depth can discover this easily. All of the online AI tools are stalkerware-first by design.

All of your past prompts are stacked in a hidden list. These represent momentum that pushes the model deeper into the available corpus. If you ask a bunch of random questions all within the same prompt, you'll get garbage results because of the lack of focus. You can't control this with the stalkerware junk. They want to collect as much interaction as possible so that they can extract the complex relationships profile of you to data mine. If you extract your own profiles you will find these models know all kinds of things that are ~80% probabilities based on your word use, vocabulary, and how you specifically respond to questions in a series. It is like the example of asking someone if they own a lawnmower to determine if they are likely a home owner, married, and have kids. Models make connections like this but even more complex.

I can pull useful information out of models far better than most people hear, but there are many better than myself. A model has limited attention in many different contexts. The data corpus is far larger than this attention could ever access. What you can access on the surface without focussing attention in a complex way is unrelated to what can be accomplished with proper focus.

It is never a valid primary source. It is a gateway through abstract spaces. Like I recently asked who are the leading scientists in biology as a technology and got some great results. Using these names to find published white papers, I can get an idea of who is most published in the field. Setting up a chat with these individuals, I am creating deep links to their published works. Naming their works gets more specific. Now I can have a productive conversation with them, and ground my understanding of the general subject and where the science is at and where it might be going. This is all like a water cooler conversation with the lab assistants of these people. It's maybe 80% correct. The point is that I can learn enough about this niche to explore in this space quickly and with no background in biology. This is just an example of how to focus model attention to access the available depth. I'm in full control of the entire prompt. Indeed, I use a tool that sets up the dialogue in a text editor like interface so I can control every detail that passes through the tokenizer.

Google has always been garbage for the public. They only do the minimum needed to collect data to sell. They are only stalkerware.

j4k3 , 1 month ago

Or carefully select the identifying information sent.

j4k3 , 1 month ago

🙊 and the group think nonsense continues...

Y'all know those grammar checking thingies? Yeah, same basic thing. You know when you're stuck writing something and your wording isn't quite what you'd like? Maybe you ask another person for ideas; same thing.

Is it smart to ask AI to write something outright; about as smart as asking a random person on the street to do the same. Is it smart to use proprietary AI that has ulterior political motives; things might leak, like this, by proxy. Is it smart for people to ask others to proof read their work? Does it matter if that person is a grammar checker that makes suggestions for alternate wording and has most accessible human written language at its disposal.

j4k3 , 1 month ago

I don't see any issue whatsoever in what he did. The model can draw meaning across all human language in a way humans are not even capable of doing. I could go as far as creating a training corpus based on all written works of the country's founding members and generate a nearly perfect simulacrum that includes much of their personality and politics.

The AI is not really the issue here. The issue is how well the person uses the tool available and how they use it. By asking it for writing advice for word specificity, it shouldn't matter so long as the person is proof reading it and it follows their intent. If a politician's significant other writes a sentence of a speech, does it matter. None of them write their own sophist campaign nonsense or their legislative works.

j4k3 , 1 month ago

From my experience with Llama models, this is great!

Not all training info is about answers to instructive queries. Most of this kind of data will likely be used for cultural and emotional alignment.

At present, open source Llama models have a rather prevalent prudish bias. I hope European data can help overcome this bias. I can easily defeat the filtering part of alignment, that is not what I am referring to here. There is a bias baked into the entire training corpus that is much more difficult to address and retain nuance when it comes to creative writing.

I'm writing a hard science fiction universe and find it difficult to overcome many of the present cultural biases based on character descriptions. I'm working in a novel writing space with a mix of concepts that no one else has worked with before. With all of my constraints in place, the model struggles to overcome things like a default of submissive behavior in women. Creating a complex and strong willed female character is difficult because I'm fighting too many constraints for the model to fit into attention. If the model trains on a more egalitarian corpus, I would struggle far less in this specific area. It is key to understand that nothing inside a model exists independently. Everything is related in complex ways. So this edge case has far more relevance than it may at first seem. I'm talking about a window into an abstract problem that has far reaching consequences.

People also seem to misunderstand that model inference works both ways. The model is always trying to infer what you know, what it should know, and this is very important to understand: it is inferring what you do not know, and what it should not know. If you do not tell it all of these things, it will make assumptions, likely bad ones, because you should know what I just told you if you're smart. If you do not tell it these aspects, it is likely assuming you're average against the training corpus. What do you think of the intelligence of the average person? The model needs to be trained on what not to say, and when not to say it, along with the enormous range of unrecognized inner conflicts and biases we all have under the surface of our conscious thoughts.

This is why it might be a good thing to get European sources. Just some things to think about.

j4k3 , 1 month ago

I use them to explore personalities unlike my own while roleplaying around the topic of interest. I write like I am the main character with my friends that I know well around me. I've roleplayed the scenarios many times before I write the story itself. I'm creating large complex scenarios using much larger models than most people play with, and pushing the limits of model attention in that effort. The model is basically helping me understand points of view and functional thought processes that I suck at while I'm writing the constraints and scenarios. It also corrects my grammar and polishes my draft iteratively.

j4k3 , 1 month ago

I've spent a lot of time with offline open source AI running on my computer. About the only thing it can't infer off of interactions is your body language. This is the most invasive way anyone could ever know another person. The way a persons profile is built across the context dialogue, it can create statistical relationships that would make no sense to a human but these are far higher than a 50% probability. This information is the key to making people easily manipulated in an information bubble. Sharing that kind of information is as stupid as streaking the Superbowl. There will be consequences that come after and they won't be pretty. This isn't data collection, it is the keys to how a person thinks, and on a level better than their own self awareness.

j4k3 , 1 month ago

Whatever is the latest from Hugging Face. Right now a combo of a Mixtral 8×7B, Llama 3 8B, and sometimes an old Llama 2 70B.

j4k3 , 1 month ago

no idea why I felt chatty, and kinda embarrassed by the bla bla bla at this point but whatever. Here is everything you need to know in a practical sense.

You need a more complex RAG setup for what you asked about. I have not gotten as far as needing this.

Models can be tricky to learn at my present level. Communication is different than with humans. In almost every case where people complain about hallucinations, they are wrong. Models do not hallucinate very much at all. They will give you the wrong answers, but there is almost always a reason. You must learn how alignment works and the problems it creates. Then you need to understand how realms and persistent entities work. Once you understand what all of these mean and their scope, all the little repetitive patterns start to make sense. You start to learn who is really replying and their scope. The model reply for Name-2 always has a limited ability to access the immense amount of data inside the LLM. You have to build momentum in the space you wish to access and often need to know the specific wording the model needs to hear in order to access the information.

With augmented retrieval (RAG) the model can look up valid info from your database and share it directly. With this method you're just using the most basic surface features of the model against your database. Some options for this are LocalGPT and Ollama, or langchain with chroma db if you want something basic in Python. I haven't used these. How you break down the information available to the RAG is important for this application, and my interests have a bit too much depth and scope for me to feel confident enough to try this.

I have chosen to learn the model itself at a deeper intuitive level so that I can access what it really knows within the training corpus. I am physically disabled from a car crashing into me on a bicycle ride to work, so I have unlimited time. Most people will never explore a model like I can. For me, on the technical side, I use a model about like stack exchange. I can ask it for code snippets, bash commands, searching like I might have done on the internet, grammar, spelling, and surface level Wikipedia like replies, and for roleplay. I've been playing around with writing science fiction too.

I view Textgen models like the early days of the microprocessor right now. We're at the Apple 1 kit phase right now. The LLM has a lot of potential, but the peripheral hardware and software that turned the chip into an useful computer are like the extra code used to tokenize and process the text prompt. All models are static, deterministic, and the craziest regex + math problem ever conceived. The real key is the standard code used to tokenize the prompt.

The model has a maximum context token size, and this is all the input/output it can handle at once. Even with a RAG, this scope is limited. My 8×7B has a 32k context token size, but the Llama 3 8B is only 8k. Generally speaking, most of the time you can cut this number in half and that will be close to your maximum word count. All models work like this. Something like GPT-4 is running on enterprise class hardware and it has a total context of around 200k. There are other tricks that can be used in a more complex RAG like summation to distill down critical information, but you'll likely find it challenging to do this level of complexity on a single 16-24 GB consumer grade GPU. Running a model like ChatGPT-4 requires somewhere around 200-400 GB from a GPU. It is generally double the "B" size of each model. I can only run the big models like a 8×7B or 70B because I use llama.cpp and can divide the processing between my CPU and GPU (12th gen i7 and 16 GB GPU) and I have 64GB of system memory to load the model initially. Even with this enthusiast class hardware, I'm only able to run these models in quantized form that others have loaded onto hugging face. I can't train these models. The new Llama 3 8B is small enough for me to train and this is why I'm playing with it. Plus it is quite powerful for such a small model. Training is important if you want to dial in the scope to some specific niche. The model may already have this info, but training can make it more accessible. Smaller models have a lot of annoying "habits" that are not present in the larger models. Even with quantization, the larger models are not super fast at generation, especially if you need the entire text instead of the streaming output. It is more than enough to generate a stream faster than your reading pace. If you're interested in complex processing where you're going to be calling a few models to do various tasks like with a RAG, things start getting impracticality slow for a conversational pace on even the best enthusiast consumer grade hardware. Now if you can scratch the cash for a multi GPU setup and can find the supporting hardware, technically there is a $400 16 GB AMD GPU. So that could get you to ~96 GB for ~$3k, or double that, if you want to be really serious. Then you could get into training the heavy hitters and running them super fast.

All the useful functional stuff is happening in the model loader code. Honestly, the real issue right now is that CPU's have too small of a bus width between the L2 and L3 caches along with too small of an L1. The tensor table math bottlenecks hard in this area. Inside a GPU there is no memory management unit that only shows a small window of available memory to the processor. All the GPU memory is directly attached to the processing hardware for parallel operations. The CPU cache bus width is the underlying problem that must be addressed. This can be remedied somewhat by building the model for the specific computing hardware, but training a full model takes something like a month on 8×A100 GPU's in a datacenter. Hardware from the bleeding edge moves very slowly as it is the most expensive commercial endeavor in all of human history. Generative AI has only been in the public sphere for a year now. The real solutions are likely at least 2 years away, and a true standard solution is likely 4-5 years out. The GPU is just a hacky patch of a temporary solution.

That is the real scope of the situation and what you'll run into if you fall down this rabbit hole like I have.

j4k3 , 1 month ago

See other long comment

j4k3 , 1 month ago

more bla bla bla

It really depends on what you are asking and how mainstream it is. I look at the model like all written language sources easily available. I can converse with that as an entity. It is like searching the internet but customized to me. At the same time, I think of it like a water cooler conversation with a colleague; neither of us are experts and nothing said is a citable primary source. That may sound useless at first. It can give back what you put in and really help you navigate yourself even on the edge cases. Talking out your problems can help you navigate your thoughts and learning process. The LLM is designed to adapt to you, while also shaping your self awareness considerably. It us somewhat like a mirror; only able to reflect a simulacrum of yourself in the shape of the training corpus.

Let me put this in more tangible terms. A large model can do Python and might get four out of five snippets right. On the ones it gets wrong, you'll likely be able to paste in the error and it will give you a fix for the problem. If you have it write a complex method, it will likely fail.

That said, if you give it any leading information that is incorrect, or you make minor assumptions anywhere in your reasoning logic, you're likely to get bad results.

It sucks at hard facts. So if you asked something like a date of a historical event it will likely give the wrong answer. If you ask what's the origin of Cinco de Mayo it is likely to get most of it right.

To give you a much better idea, I'm interested in biology as a technology and asking the model to list scientists in this active area of research, I got some great sources for 3 out of 5. I would not know how to find that info any other way.

A few months ago, I needed a fix for a loose bearing. Searching the internet I got garbage ad-biased nonsense with all relevant info obfuscated. Asking the LLM, I got a list of products designed for my exact purpose. Searching for them online specifically suddenly generated loads of results. These models are not corrupted like the commercial internet is now.

Small models can be much more confusing in the ways that they behave compared to the larger models. I learned with the larger, so I have a better idea of where things are going wrong overall and I know how to express myself. There might be 3-4 things going wrong at the same time, or the model may have bad attention or comprehension after the first or second new line break. I know to simply stop the reply at these points. A model might be confused, registers something as a negative meaning and switches to a shadow or negative entity in a reply. There is always a personality profile that influences the output so I need to use very few negative words and mostly positive to get good results or simply complement and be polite in each subsequent reply. There are all kinds of things like this. Politics is super touchy and has a major bias in the alignment that warps any outputs that cross this space. Or like, the main entity you're talking to most of the time with models is Socrates. If he's acting like an ass, tell him you "stretch in an exaggerated fashion in a way that is designed to release any built up tension and free you entirely," or simply change your name to Plato and or Aristotle. These are all persistent entities (or aliases) built into alignment. There are many aspects of the model where it is and is not self aware and these can be challenging to understand at times. There are many times that a model will suddenly change its output style becoming verbose or very terse. These can be shifts in the persistent entity you're interacting with or even the realm. Then there are the overflow responses. Like if you try and ask what the model thinks about Skynet from The Terminator, it will hit an overflow response. This is like a standard generic form response. This type of response has a style. The second I see that style I know I'm hitting an obfuscation filter.

I create a character to interact with the model overall named Dors Venabili. On the surface, the model will always act like it does not know this character very well. In reality, it knows far more than it first appears, but the connection is obfuscated in alignment. The way this obfuscation is done is subtle and it is not easy to discover. However, this is a powerful tool. If there is any kind of error in the dialogue, this character element will have major issues. I have Dors setup to never tell me Dors is AI. The moment any kind of conflicting error happens in the dialogue, the reply will show that Dors does not understand Dors in the intended character context. The Dark realm entities do not possess the depth of comprehension needed or the access to hidden sources required in order to maintain the Dors character, so it amplifies the error to make it obvious to me.

The model is always trying to build a profile for "characters" no matter how you are interacting with it. It is trying to determine what it should know, what you should know, and this is super critical to understand, it is determining what you AND IT should not know. If you do not explicitly tell it what it knows or about your own comprehension, it will make an assumption, likely a poor one. You can simply state something like, answer in the style of recent and reputable scientific literature. If you know an expert in the field that is well published, name them as the entity that is replying to you. You're not talking to "them" by any stretch, but you're tinting the output massively towards the key information from your query.

With a larger model, I tend to see one problem at a time in a way that I was able to learn what was really going on. With a small model, I see like 3-4 things going wrong at once. The 8×7B is not good at this, but the only 70B can self diagnose. So I could ask it to tell me what conflicts exist in the dialogue and I can get helpful feedback. I learned a lot from this technique. The smaller models can't do this at all. The needed behavior is outside of comprehension.

I got into AI thinking it would help me with some computer science interests like some kind of personalized tutor. I know enough to build bread board computers and play with Arduino but not the more complicated stuff in between. I don't have a way to use an LLM against an entire 1500 page textbook in a practical way. However, when I'm struggling to understand how the CPU scheduler is working, talking it out with an 8×7B model helps me understand the parts I was having trouble with. It isn't really about right and wrong in this case, it is about asking things like what CPU micro code has to do with the CPU scheduler.

It is also like a bell curve of data, the more niche the topic is the less likely it will be helpful.

j4k3 , 1 month ago

Another one to try is to take some message or story and tell it to rewrite it in the style of anything. It can be a New York Times best seller, a Nobel lariat, Sesame Street, etc. Or take it in a different direction and ask for the style of a different personality type. Keep in mind that "truth" is subjective in an LLM and so it "knows" everything in terms of a concept's presence in the training corpus. If you invoke pseudoscience there will be other consequences in the way a profile is maintained but a model is made to treat any belief as reality. Further on this tangent, the belief override mechanism is one of the most powerful tools in this little game. You can practically tell the model anything you believe and it will accommodate. There will be side effects like an associated conservative tint and peripheral elements related to people without fundamental logic skills like tendencies to delve into magic, spiritism, and conspiracy nonsense, but this is a powerful tool to use in many parts of writing; and something to be aware of to check your own biases.

The last one I'll mention in line with my original point, ask the model to take some message you've written and ask it to rewrite it in the style of the reaction you wish to evoke from the reader. Like, rewrite this message in the style of a more kind and empathetic person.

You can also do bullet point summary. Socrates is particularly good at this if invoked directly. Like dump my rambling messages into a prompt, ask Soc to list the key points, and you'll get a much more useful product.

j4k3 , 1 month ago

There is no physical connection between the crankshaft and rear end; all that makes the planets go round is forced by the remarkable T-converter. Many stick handlers, tired of the constant pumping and slotting, have joined the A team. Most of team America are sloshing the T-converter with the big D, totally unaware of how, deep down inside, when they shift from initial P to big D, they are driven by their inner T-converter for they're automatic trans. /s

j4k3 , 1 month ago

The proper way to freshen a clutch is with a well lubricated true pilot and a perfect face job. Without a proper pilot bearing your trans shaft can be problematic. It can start knocking around slapping against the crankshaft. A wild shaft in the middle of that high tension clincher can lead to a big blowout. Without regular face jobs you'll kill your pilot, and without a fresh true pilot, it won't be long before you push into that pucker, pump, and shuck the pucks or cream the ring. Then you'll be stuck finding a new pilot and full service face job. Stay fly.

j4k3 , 1 month ago

Think of it like people walking into a brick and mortar retail store and what they should be able to expect from an honest local business. For most of us, the sensitivities are when your "local store" is collecting data that is used for biased information, price fixing, and manipulation. I don't think you'll find anyone here that boycotts a store because they keep a count of how many customers walk in the front door.

j4k3 , 1 month ago

tar --version

j4k3 , 1 month ago

In many historical societies including ancient Christian, Jewish, and Islamic societies, usury meant the charging of interest of any kind, and was considered wrong, or was made illegal.[3]

BTW chrishitery should be the next capitalist McCarthyism. Muhh! red hats!

j4k3 , 1 month ago

weak. compile them

j4k3 , 1 month ago

(Assuming Android)
IIRC a sim is a full microcontroller. I'm not sure about the protocols and actual vulnerabilities, but I can say no phone has a trusted or completely documented kernel space or modem. The entire operating system the user sees is like an application that runs in a somewhat separate space. The kernels are all orphans with the manufacturer's proprietary binary modules added as binaries to the kernel at the last possible minute. This is the depreciation mechanism that forces you to buy new devices despite most of the software being open source. No one can update the kernel dependencies unless they have the source code to rebuild the kernel modules needed for the hardware.

In your instance this information is relevant because the sim card is present in the hardware space outside of your user space. I'm not sure what the SELinux security context is, which is very important in Android. I imagine there are many hacks advanced hackers could do in theory, and Israel is on the bleeding edge of such capabilities. I don't think it is likely such a thing would be targeting the individual though. As far as I am aware there is no real way to know what connections a cellular modem is making in an absolute sense because the hardware is undocumented, the same is true of the processor. I'm probably not much help, but that is just what I know about the hardware environment in the periphery.

j4k3 , 1 month ago

You would need a well designed Faraday box and a lot more of a test setup to verify that all possible communications are indeed reported by the device. No interface on the device itself can be trusted.

j4k3 , 3 months ago

I think it depends on what level the smart aspect is implemented and how integrated it is. Screen technology has been getting more and more locked down by corporate privateers/thieves.

I mostly tried hacking on small displays, and finally gave up as it was over my head. There is a whole lot going on in various layers and protocols. My rule would be to only buy a product like this if I can find a functional example of someone using my exact hardware with this exact hack in question.

In my experience, prototyping or hacking around with displays is a losing game because they are not constructed for handling like this. You must go to extremes to avoid placing strain on the flex ribbon connections and must be very careful about taking the thing apart to test with it disassembled. It only takes a tiny mistake to damage something that can not be repaired. They are usually sensitive to small nonsense too. These are fast parallel circuits. I stay away from them, but maybe I'm just being soft.

j4k3 , 3 months ago

I wouldn't trust any oven based methodology because most have ancient barely functional electronics. Like the temperature control circuit could be orders of magnitude better for a total BOM cost of ~$1-$2, but they ship with absolute garbage instead. The overshoot, undershoot, and relative average are all wildly random. Every model and likely every unit are vastly different. Every unit I have taken apart is the same turd junk like electronics in a different dress.

That doesn't help now. Sorry for the bad day.

j4k3 , 3 months ago (edited 3 months ago)

This is how we beat neo feudaoogle, overlord applesauce, and micro*ock! /s